Technologies for Managing Resource Allocation With a Hierarchical Model

Abstract

Technologies for allocating resources of a set of managed nodes to workloads with a hierarchical model include an orchestrator server to receive resource allocation objective data. The orchestrator server is further to determine an initial assignment of a set of workloads among the managed nodes, receive telemetry data from the managed nodes, generate, from the telemetry data, a hierarchical model indicative of the resource utilization of each managed node, determine, with the hierarchical model, differences in resource utilization for each workload as a function of the managed node that performed the workload, determine, as a function of the telemetry data and the determined differences, an adjustment to the assignment of the workloads to increase an achievement of at least one of the resource allocation objectives without decreasing the achievement of any of the other resource allocation objectives, and apply the adjustment to the assignments of the workloads among the managed nodes as the workloads are performed.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS

The present application claims the benefit of U.S. Provisional Patent Application No. 62/365,969, filed Jul. 22, 2016, U.S. Provisional Patent Application No. 62/376,859, filed Aug. 18, 2016, and U.S. Provisional Patent Application No. 62/427,268, filed Nov. 29, 2016.

BACKGROUND

In a typical cloud-based computing environment (e.g., a data center), multiple compute nodes may execute workloads (e.g., processes, applications, services, etc.) on behalf of customers. The workloads may exhibit different resource usages based on the particular operations they perform. Furthermore, different compute nodes in the data center may have different characteristics, either because the compute nodes include differing hardware and/or are located in difference places within the data center. For example, some compute nodes may be equipped with a more efficient processor, faster or more memory, or more efficient fans, than other compute nodes in the same data center. Furthermore, different physical areas in a data center may have a different ambient temperatures that affect the internal temperatures of the compute nodes. Even when all of the compute nodes and temperatures within a data center are initially homogenous, differences may arise over time, such as when components are upgraded, compute nodes are throttled to extend their life expectancy, or other compute nodes are overclocked to increase their performance. As such, a theoretically efficient distribution of workloads among the compute nodes may, in reality, be suboptimal if it is based on the assumption that the compute nodes themselves are identical in their hardware capabilities and are placed in locations having identical characteristics.

BRIEF DESCRIPTION OF THE DRAWINGS

The concepts described herein are illustrated by way of example and not by way of limitation in the accompanying figures. For simplicity and clarity of illustration, elements illustrated in the figures are not necessarily drawn to scale. Where considered appropriate, reference labels have been repeated among the figures to indicate corresponding or analogous elements.

FIG. 1 is a diagram of a conceptual overview of a data center in which one or more techniques described herein may be implemented according to various embodiments;

FIG. 2 is a diagram of an example embodiment of a logical configuration of a rack of the data center of FIG. 1;

FIG. 3 is a diagram of an example embodiment of another data center in which one or more techniques described herein may be implemented according to various embodiments;

FIG. 4 is a diagram of another example embodiment of a data center in which one or more techniques described herein may be implemented according to various embodiments;

FIG. 5 is a diagram of a connectivity scheme representative of link-layer connectivity that may be established among various sleds of the data centers of FIGS. 1, 3, and 4;

FIG. 6 is a diagram of a rack architecture that may be representative of an architecture of any particular one of the racks depicted in FIGS. 1-4 according to some embodiments;

FIG. 7 is a diagram of an example embodiment of a sled that may be used with the rack architecture of FIG. 6;

FIG. 8 is a diagram of an example embodiment of a rack architecture to provide support for sleds featuring expansion capabilities;

FIG. 9 is a diagram of an example embodiment of a rack implemented according to the rack architecture of FIG. 8;

FIG. 10 is a diagram of an example embodiment of a sled designed for use in conjunction with the rack of FIG. 9;

FIG. 11 is a diagram of an example embodiment of a data center in which one or more techniques described herein may be implemented according to various embodiments;

FIG. 12 is a simplified block diagram of at least one embodiment of a system for managing the assignment of workloads among a set of managed nodes based on a hierarchical model of the managed nodes;

FIG. 13 is a simplified block diagram of at least one embodiment of an orchestrator server of the system of FIG. 12;

FIG. 14 is a simplified block diagram of at least one embodiment of an environment that may be established by the orchestrator server of FIGS. 12 and 13;

FIGS. 15-17 are a simplified flow diagram of at least one embodiment of a method for assigning workloads among a set of managed nodes based on a hierarchical model of the managed nodes that may be performed by the orchestrator server of FIGS. 12-14; and

FIG. 18 is a simplified diagram of at least one embodiment of a hierarchical model that may be established by the orchestrator server of FIGS. 12 and 13.

DETAILED DESCRIPTION OF THE DRAWINGS

While the concepts of the present disclosure are susceptible to various modifications and alternative forms, specific embodiments thereof have been shown by way of example in the drawings and will be described herein in detail. It should be understood, however, that there is no intent to limit the concepts of the present disclosure to the particular forms disclosed, but on the contrary, the intention is to cover all modifications, equivalents, and alternatives consistent with the present disclosure and the appended claims.

References in the specification to “one embodiment,” “an embodiment,” “an illustrative embodiment,” etc., indicate that the embodiment described may include a particular feature, structure, or characteristic, but every embodiment may or may not necessarily include that particular feature, structure, or characteristic. Moreover, such phrases are not necessarily referring to the same embodiment. Further, when a particular feature, structure, or characteristic is described in connection with an embodiment, it is submitted that it is within the knowledge of one skilled in the art to effect such feature, structure, or characteristic in connection with other embodiments whether or not explicitly described. Additionally, it should be appreciated that items included in a list in the form of “at least one A, B, and C” can mean (A); (B); (C); (A and B); (A and C); (B and C); or (A, B, and C). Similarly, items listed in the form of “at least one of A, B, or C” can mean (A); (B); (C); (A and B); (A and C); (B and C); or (A, B, and C).

The disclosed embodiments may be implemented, in some cases, in hardware, firmware, software, or any combination thereof. The disclosed embodiments may also be implemented as instructions carried by or stored on a transitory or non-transitory machine-readable (e.g., computer-readable) storage medium, which may be read and executed by one or more processors. A machine-readable storage medium may be embodied as any storage device, mechanism, or other physical structure for storing or transmitting information in a form readable by a machine (e.g., a volatile or non-volatile memory, a media disc, or other media device).

In the drawings, some structural or method features may be shown in specific arrangements and/or orderings. However, it should be appreciated that such specific arrangements and/or orderings may not be required. Rather, in some embodiments, such features may be arranged in a different manner and/or order than shown in the illustrative figures. Additionally, the inclusion of a structural or method feature in a particular figure is not meant to imply that such feature is required in all embodiments and, in some embodiments, may not be included or may be combined with other features.

FIG. 1 illustrates a conceptual overview of a data center 100 that may generally be representative of a data center or other type of computing network in/for which one or more techniques described herein may be implemented according to various embodiments. As shown in FIG. 1, data center 100 may generally contain a plurality of racks, each of which may house computing equipment comprising a respective set of physical resources. In the particular non-limiting example depicted in FIG. 1, data center 100 contains four racks 102A to 102D, which house computing equipment comprising respective sets of physical resources (PCRs) 105A to 105D. According to this example, a collective set of physical resources 106 of data center 100 includes the various sets of physical resources 105A to 105D that are distributed among racks 102A to 102D. Physical resources 106 may include resources of multiple types, such as—for example—processors, co-processors, accelerators, field-programmable gate arrays (FPGAs), memory, and storage. The embodiments are not limited to these examples.

The illustrative data center 100 differs from typical data centers in many ways. For example, in the illustrative embodiment, the circuit boards (“sleds”) on which components such as CPUs, memory, and other components are placed are designed for increased thermal performance In particular, in the illustrative embodiment, the sleds are shallower than typical boards. In other words, the sleds are shorter from the front to the back, where cooling fans are located. This decreases the length of the path that air must to travel across the components on the board. Further, the components on the sled are spaced further apart than in typical circuit boards, and the components are arranged to reduce or eliminate shadowing (i.e., one component in the air flow path of another component). In the illustrative embodiment, processing components such as the processors are located on a top side of a sled while near memory, such as DIMMs, are located on a bottom side of the sled. As a result of the enhanced airflow provided by this design, the components may operate at higher frequencies and power levels than in typical systems, thereby increasing performance. Furthermore, the sleds are configured to blindly mate with power and data communication cables in each rack 102A, 102B, 102C, 102D, enhancing their ability to be quickly removed, upgraded, reinstalled, and/or replaced. Similarly, individual components located on the sleds, such as processors, accelerators, memory, and data storage drives, are configured to be easily upgraded due to their increased spacing from each other. In the illustrative embodiment, the components additionally include hardware attestation features to prove their authenticity.

Furthermore, in the illustrative embodiment, the data center 100 utilizes a single network architecture (“fabric”) that supports multiple other network architectures including Ethernet and Omni-Path. The sleds, in the illustrative embodiment, are coupled to switches via optical fibers, which provide higher bandwidth and lower latency than typical twisted pair cabling (e.g., Category 5, Category 5e, Category 6, etc.). Due to the high bandwidth, low latency interconnections and network architecture, the data center 100 may, in use, pool resources, such as memory, accelerators (e.g., graphics accelerators, FPGAs, ASICs, etc.), and data storage drives that are physically disaggregated, and provide them to compute resources (e.g., processors) on an as needed basis, enabling the compute resources to access the pooled resources as if they were local. The illustrative data center 100 additionally receives usage information for the various resources, predicts resource usage for different types of workloads based on past resource usage, and dynamically reallocates the resources based on this information.

The racks 102A, 102B, 102C, 102D of the data center 100 may include physical design features that facilitate the automation of a variety of types of maintenance tasks. For example, data center 100 may be implemented using racks that are designed to be robotically-accessed, and to accept and house robotically-manipulatable resource sleds. Furthermore, in the illustrative embodiment, the racks 102A, 102B, 102C, 102D include integrated power sources that receive a greater voltage than is typical for power sources. The increased voltage enables the power sources to provide additional power to the components on each sled, enabling the components to operate at higher than typical frequencies.

FIG. 2 illustrates an exemplary logical configuration of a rack 202 of the data center 100. As shown in FIG. 2, rack 202 may generally house a plurality of sleds, each of which may comprise a respective set of physical resources. In the particular non-limiting example depicted in FIG. 2, rack 202 houses sleds 204-1 to 204-4 comprising respective sets of physical resources 205-1 to 205-4, each of which constitutes a portion of the collective set of physical resources 206 comprised in rack 202. With respect to FIG. 1, if rack 202 is representative of—for example—rack 102A, then physical resources 206 may correspond to the physical resources 105A comprised in rack 102A. In the context of this example, physical resources 105A may thus be made up of the respective sets of physical resources, including physical storage resources 205-1, physical accelerator resources 205-2, physical memory resources 205-3, and physical compute resources 205-5 comprised in the sleds 204-1 to 204-4 of rack 202. The embodiments are not limited to this example. Each sled may contain a pool of each of the various types of physical resources (e.g., compute, memory, accelerator, storage). By having robotically accessible and robotically manipulatable sleds comprising disaggregated resources, each type of resource can be upgraded independently of each other and at their own optimized refresh rate.

FIG. 3 illustrates an example of a data center 300 that may generally be representative of one in/for which one or more techniques described herein may be implemented according to various embodiments. In the particular non-limiting example depicted in FIG. 3, data center 300 comprises racks 302-1 to 302-32. In various embodiments, the racks of data center 300 may be arranged in such fashion as to define and/or accommodate various access pathways. For example, as shown in FIG. 3, the racks of data center 300 may be arranged in such fashion as to define and/or accommodate access pathways 311A, 311B, 311C, and 311D. In some embodiments, the presence of such access pathways may generally enable automated maintenance equipment, such as robotic maintenance equipment, to physically access the computing equipment housed in the various racks of data center 300 and perform automated maintenance tasks (e.g., replace a failed sled, upgrade a sled). In various embodiments, the dimensions of access pathways 311A, 311B, 311C, and 311D, the dimensions of racks 302-1 to 302-32, and/or one or more other aspects of the physical layout of data center 300 may be selected to facilitate such automated operations. The embodiments are not limited in this context.

FIG. 4 illustrates an example of a data center 400 that may generally be representative of one in/for which one or more techniques described herein may be implemented according to various embodiments. As shown in FIG. 4, data center 400 may feature an optical fabric 412. Optical fabric 412 may generally comprise a combination of optical signaling media (such as optical cabling) and optical switching infrastructure via which any particular sled in data center 400 can send signals to (and receive signals from) each of the other sleds in data center 400. The signaling connectivity that optical fabric 412 provides to any given sled may include connectivity both to other sleds in a same rack and sleds in other racks. In the particular non-limiting example depicted in FIG. 4, data center 400 includes four racks 402A to 402D. Racks 402A to 402D house respective pairs of sleds 404A-1 and 404A-2, 404B-1 and 404B-2, 404C-1 and 404C-2, and 404D-1 and 404D-2. Thus, in this example, data center 400 comprises a total of eight sleds. Via optical fabric 412, each such sled may possess signaling connectivity with each of the seven other sleds in data center 400. For example, via optical fabric 412, sled 404A-1 in rack 402A may possess signaling connectivity with sled 404A-2 in rack 402A, as well as the six other sleds 404B-1, 404B-2, 404C-1, 404C-2, 404D-1, and 404D-2 that are distributed among the other racks 402B, 402C, and 402D of data center 400. The embodiments are not limited to this example.

FIG. 5 illustrates an overview of a connectivity scheme 500 that may generally be representative of link-layer connectivity that may be established in some embodiments among the various sleds of a data center, such as any of example data centers 100, 300, and 400 of FIGS. 1, 3, and 4. Connectivity scheme 500 may be implemented using an optical fabric that features a dual-mode optical switching infrastructure 514. Dual-mode optical switching infrastructure 514 may generally comprise a switching infrastructure that is capable of receiving communications according to multiple link-layer protocols via a same unified set of optical signaling media, and properly switching such communications. In various embodiments, dual-mode optical switching infrastructure 514 may be implemented using one or more dual-mode optical switches 515. In various embodiments, dual-mode optical switches 515 may generally comprise high-radix switches. In some embodiments, dual-mode optical switches 515 may comprise multi-ply switches, such as four-ply switches. In various embodiments, dual-mode optical switches 515 may feature integrated silicon photonics that enable them to switch communications with significantly reduced latency in comparison to conventional switching devices. In some embodiments, dual-mode optical switches 515 may constitute leaf switches 530 in a leaf-spine architecture additionally including one or more dual-mode optical spine switches 520.

In various embodiments, dual-mode optical switches may be capable of receiving both Ethernet protocol communications carrying Internet Protocol (IP packets) and communications according to a second, high-performance computing (HPC) link-layer protocol (e.g., Intel's Omni-Path Architecture's, Infiniband) via optical signaling media of an optical fabric. As reflected in FIG. 5, with respect to any particular pair of sleds 504A and 504B possessing optical signaling connectivity to the optical fabric, connectivity scheme 500 may thus provide support for link-layer connectivity via both Ethernet links and HPC links. Thus, both Ethernet and HPC communications can be supported by a single high-bandwidth, low-latency switch fabric. The embodiments are not limited to this example.

FIG. 6 illustrates a general overview of a rack architecture 600 that may be representative of an architecture of any particular one of the racks depicted in FIGS. 1 to 4 according to some embodiments. As reflected in FIG. 6, rack architecture 600 may generally feature a plurality of sled spaces into which sleds may be inserted, each of which may be robotically-accessible via a rack access region 601. In the particular non-limiting example depicted in FIG. 6, rack architecture 600 features five sled spaces 603-1 to 603-5. Sled spaces 603-1 to 603-5 feature respective multi-purpose connector modules (MPCMs) 616-1 to 616-5.

FIG. 7 illustrates an example of a sled 704 that may be representative of a sled of such a type. As shown in FIG. 7, sled 704 may comprise a set of physical resources 705, as well as an MPCM 716 designed to couple with a counterpart MPCM when sled 704 is inserted into a sled space such as any of sled spaces 603-1 to 603-5 of FIG. 6. Sled 704 may also feature an expansion connector 717. Expansion connector 717 may generally comprise a socket, slot, or other type of connection element that is capable of accepting one or more types of expansion modules, such as an expansion sled 718. By coupling with a counterpart connector on expansion sled 718, expansion connector 717 may provide physical resources 705 with access to supplemental computing resources 705B residing on expansion sled 718. The embodiments are not limited in this context.

FIG. 8 illustrates an example of a rack architecture 800 that may be representative of a rack architecture that may be implemented in order to provide support for sleds featuring expansion capabilities, such as sled 704 of FIG. 7. In the particular non-limiting example depicted in FIG. 8, rack architecture 800 includes seven sled spaces 803-1 to 803-7, which feature respective MPCMs 816-1 to 816-7. Sled spaces 803-1 to 803-7 include respective primary regions 803-1A to 803-7A and respective expansion regions 803-1B to 803-7B. With respect to each such sled space, when the corresponding MPCM is coupled with a counterpart MPCM of an inserted sled, the primary region may generally constitute a region of the sled space that physically accommodates the inserted sled. The expansion region may generally constitute a region of the sled space that can physically accommodate an expansion module, such as expansion sled 718 of FIG. 7, in the event that the inserted sled is configured with such a module.

FIG. 9 illustrates an example of a rack 902 that may be representative of a rack implemented according to rack architecture 800 of FIG. 8 according to some embodiments. In the particular non-limiting example depicted in FIG. 9, rack 902 features seven sled spaces 903-1 to 903-7, which include respective primary regions 903-1A to 903-7A and respective expansion regions 903-1B to 903-7B. In various embodiments, temperature control in rack 902 may be implemented using an air cooling system. For example, as reflected in FIG. 9, rack 902 may feature a plurality of fans 919 that are generally arranged to provide air cooling within the various sled spaces 903-1 to 903-7. In some embodiments, the height of the sled space is greater than the conventional “1U” server height. In such embodiments, fans 919 may generally comprise relatively slow, large diameter cooling fans as compared to fans used in conventional rack configurations. Running larger diameter cooling fans at lower speeds may increase fan lifetime relative to smaller diameter cooling fans running at higher speeds while still providing the same amount of cooling. The sleds are physically shallower than conventional rack dimensions. Further, components are arranged on each sled to reduce thermal shadowing (i.e., not arranged serially in the direction of air flow). As a result, the wider, shallower sleds allow for an increase in device performance because the devices can be operated at a higher thermal envelope (e.g., 250 W) due to improved cooling (i.e., no thermal shadowing, more space between devices, more room for larger heat sinks, etc.).

MPCMs 916-1 to 916-7 may be configured to provide inserted sleds with access to power sourced by respective power modules 920-1 to 920-7, each of which may draw power from an external power source 921. In various embodiments, external power source 921 may deliver alternating current (AC) power to rack 902, and power modules 920-1 to 920-7 may be configured to convert such AC power to direct current (DC) power to be sourced to inserted sleds. In some embodiments, for example, power modules 920-1 to 920-7 may be configured to convert 277-volt AC power into 12-volt DC power for provision to inserted sleds via respective MPCMs 916-1 to 916-7. The embodiments are not limited to this example.

MPCMs 916-1 to 916-7 may also be arranged to provide inserted sleds with optical signaling connectivity to a dual-mode optical switching infrastructure 914, which may be the same as—or similar to—dual-mode optical switching infrastructure 514 of FIG. 5. In various embodiments, optical connectors contained in MPCMs 916-1 to 916-7 may be designed to couple with counterpart optical connectors contained in MPCMs of inserted sleds to provide such sleds with optical signaling connectivity to dual-mode optical switching infrastructure 914 via respective lengths of optical cabling 922-1 to 922-7. In some embodiments, each such length of optical cabling may extend from its corresponding MPCM to an optical interconnect loom 923 that is external to the sled spaces of rack 902. In various embodiments, optical interconnect loom 923 may be arranged to pass through a support post or other type of load-bearing element of rack 902. The embodiments are not limited in this context. Because inserted sleds connect to an optical switching infrastructure via MPCMs, the resources typically spent in manually configuring the rack cabling to accommodate a newly inserted sled can be saved.

FIG. 10 illustrates an example of a sled 1004 that may be representative of a sled designed for use in conjunction with rack 902 of FIG. 9 according to some embodiments. Sled 1004 may feature an MPCM 1016 that comprises an optical connector 1016A and a power connector 1016B, and that is designed to couple with a counterpart MPCM of a sled space in conjunction with insertion of MPCM 1016 into that sled space. Coupling MPCM 1016 with such a counterpart MPCM may cause power connector 1016 to couple with a power connector comprised in the counterpart MPCM. This may generally enable physical resources 1005 of sled 1004 to source power from an external source, via power connector 1016 and power transmission media 1024 that conductively couples power connector 1016 to physical resources 1005.

Sled 1004 may also include dual-mode optical network interface circuitry 1026. Dual-mode optical network interface circuitry 1026 may generally comprise circuitry that is capable of communicating over optical signaling media according to each of multiple link-layer protocols supported by dual-mode optical switching infrastructure 914 of FIG. 9. In some embodiments, dual-mode optical network interface circuitry 1026 may be capable both of Ethernet protocol communications and of communications according to a second, high-performance protocol. In various embodiments, dual-mode optical network interface circuitry 1026 may include one or more optical transceiver modules 1027, each of which may be capable of transmitting and receiving optical signals over each of one or more optical channels. The embodiments are not limited in this context.

Coupling MPCM 1016 with a counterpart MPCM of a sled space in a given rack may cause optical connector 1016A to couple with an optical connector comprised in the counterpart MPCM. This may generally establish optical connectivity between optical cabling of the sled and dual-mode optical network interface circuitry 1026, via each of a set of optical channels 1025. Dual-mode optical network interface circuitry 1026 may communicate with the physical resources 1005 of sled 1004 via electrical signaling media 1028. In addition to the dimensions of the sleds and arrangement of components on the sleds to provide improved cooling and enable operation at a relatively higher thermal envelope (e.g., 250 W), as described above with reference to FIG. 9, in some embodiments, a sled may include one or more additional features to facilitate air cooling, such as a heatpipe and/or heat sinks arranged to dissipate heat generated by physical resources 1005. It is worthy of note that although the example sled 1004 depicted in FIG. 10 does not feature an expansion connector, any given sled that features the design elements of sled 1004 may also feature an expansion connector according to some embodiments. The embodiments are not limited in this context.

FIG. 11 illustrates an example of a data center 1100 that may generally be representative of one in/for which one or more techniques described herein may be implemented according to various embodiments. As reflected in FIG. 11, a physical infrastructure management framework 1150A may be implemented to facilitate management of a physical infrastructure 1100A of data center 1100. In various embodiments, one function of physical infrastructure management framework 1150A may be to manage automated maintenance functions within data center 1100, such as the use of robotic maintenance equipment to service computing equipment within physical infrastructure 1100A. In some embodiments, physical infrastructure 1100A may feature an advanced telemetry system that performs telemetry reporting that is sufficiently robust to support remote automated management of physical infrastructure 1100A. In various embodiments, telemetry information provided by such an advanced telemetry system may support features such as failure prediction/prevention capabilities and capacity planning capabilities. In some embodiments, physical infrastructure management framework 1150A may also be configured to manage authentication of physical infrastructure components using hardware attestation techniques. For example, robots may verify the authenticity of components before installation by analyzing information collected from a radio frequency identification (RFID) tag associated with each component to be installed. The embodiments are not limited in this context.

As shown in FIG. 11, the physical infrastructure 1100A of data center 1100 may comprise an optical fabric 1112, which may include a dual-mode optical switching infrastructure 1114. Optical fabric 1112 and dual-mode optical switching infrastructure 1114 may be the same as—or similar to—optical fabric 412 of FIG. 4 and dual-mode optical switching infrastructure 514 of FIG. 5, respectively, and may provide high-bandwidth, low-latency, multi-protocol connectivity among sleds of data center 1100. As discussed above, with reference to FIG. 1, in various embodiments, the availability of such connectivity may make it feasible to disaggregate and dynamically pool resources such as accelerators, memory, and storage. In some embodiments, for example, one or more pooled accelerator sleds 1130 may be included among the physical infrastructure 1100A of data center 1100, each of which may comprise a pool of accelerator resources—such as co-processors and/or FPGAs, for example—that is globally accessible to other sleds via optical fabric 1112 and dual-mode optical switching infrastructure 1114.

In another example, in various embodiments, one or more pooled storage sleds 1132 may be included among the physical infrastructure 1100A of data center 1100, each of which may comprise a pool of storage resources that is available globally accessible to other sleds via optical fabric 1112 and dual-mode optical switching infrastructure 1114. In some embodiments, such pooled storage sleds 1132 may comprise pools of solid-state storage devices such as solid-state drives (SSDs). In various embodiments, one or more high-performance processing sleds 1134 may be included among the physical infrastructure 1100A of data center 1100. In some embodiments, high-performance processing sleds 1134 may comprise pools of high-performance processors, as well as cooling features that enhance air cooling to yield a higher thermal envelope of up to 250 W or more. In various embodiments, any given high-performance processing sled 1134 may feature an expansion connector 1117 that can accept a far memory expansion sled, such that the far memory that is locally available to that high-performance processing sled 1134 is disaggregated from the processors and near memory comprised on that sled. In some embodiments, such a high-performance processing sled 1134 may be configured with far memory using an expansion sled that comprises low-latency SSD storage. The optical infrastructure allows for compute resources on one sled to utilize remote accelerator/FPGA, memory, and/or SSD resources that are disaggregated on a sled located on the same rack or any other rack in the data center. The remote resources can be located one switch jump away or two-switch jumps away in the spine-leaf network architecture described above with reference to FIG. 5. The embodiments are not limited in this context.

In various embodiments, one or more layers of abstraction may be applied to the physical resources of physical infrastructure 1100A in order to define a virtual infrastructure, such as a software-defined infrastructure 1100B. In some embodiments, virtual computing resources 1136 of software-defined infrastructure 1100B may be allocated to support the provision of cloud services 1140. In various embodiments, particular sets of virtual computing resources 1136 may be grouped for provision to cloud services 1140 in the form of SDI services 1138. Examples of cloud services 1140 may include—without limitation—software as a service (SaaS) services 1142, platform as a service (PaaS) services 1144, and infrastructure as a service (IaaS) services 1146.

In some embodiments, management of software-defined infrastructure 1100B may be conducted using a virtual infrastructure management framework 1150B. In various embodiments, virtual infrastructure management framework 1150B may be designed to implement workload fingerprinting techniques and/or machine-learning techniques in conjunction with managing allocation of virtual computing resources 1136 and/or SDI services 1138 to cloud services 1140. In some embodiments, virtual infrastructure management framework 1150B may use/consult telemetry data in conjunction with performing such resource allocation. In various embodiments, an application/service management framework 1150C may be implemented in order to provide QoS management capabilities for cloud services 1140. The embodiments are not limited in this context.

As shown in FIG. 12, an illustrative system 1210 for assigning workloads among a set of managed nodes 1260 based on a hierarchical model of the managed nodes includes an orchestrator server 1240 in communication with the set of managed nodes 1260. Each managed node 1260 may be embodied as an assembly of resources (e.g., physical resources 206), such as compute resources (e.g., physical compute resources 205-4), storage resources (e.g., physical storage resources 205-1), accelerator resources (e.g., physical accelerator resources 205-2), or other resources (e.g., physical memory resources 205-3) from the same or different sleds (e.g., the sleds 204-1, 204-2, 204-3, 204-4, etc.) or racks (e.g., one or more of racks 302-1 through 302-32). Each managed node 1260 may be established, defined, or “spun up” by the orchestrator server 1240 at the time a workload is to be assigned to the managed node 1260 or at any other time, and may exist regardless of whether any workloads are presently assigned to the managed node 1260. The system 1210 may be implemented in accordance with the data centers 100, 300, 400, 1100 described above with reference to FIGS. 1, 3, 4, and 11. In the illustrative embodiment, the set of managed nodes 1260 includes managed nodes 1250, 1252, and 1254. While three managed nodes 1260 are shown in the set, it should be understood that in other embodiments, the set may include a different number of managed nodes 1260 (e.g., tens of thousands). The system 1210 may be located in a data center and provide storage and compute services (e.g., cloud services) to a client device 1220 that is in communication with the system 1210 through a network 1230. The orchestrator server 1240 may support a cloud operating environment, such as OpenStack, and the managed nodes 1250 may execute one or more applications or processes (i.e., workloads), such as in virtual machines or containers, on behalf of a user of the client device 1220. As discussed in more detail herein, the orchestrator server 1240, in operation, is configured to receive resource allocation objective data indicative of thresholds or goals (“objectives”) to be satisfied during the execution of the workloads (e.g., a target power usage, a target speed at which to execute the workloads, a target temperature of the managed nodes 1260, etc.). Additionally, the orchestrator server 1240 is configured to assign workloads to the managed nodes 1260 and receive telemetry data, which may be embodied as data indicative of the performance and conditions of each managed node 1260 as the managed nodes 1260 execute the workloads assigned to them.

Additionally, in the illustrative embodiment, the orchestrator server 1240 is configured to organize the telemetry data into a hierarchical model that is indicative of a relationship between the managed nodes (e.g., a spatial relationship such as the physical locations of the managed nodes within the data center 1100 and/or a functional relationship, such as groupings of the managed nodes by the customers the nodes provide services for, the types of functions typically performed by the managed nodes, managed nodes that typically share or exchange workloads among each other, etc.). Based on differences in the physical locations and hardware in the managed nodes, a given workload may exhibit different resource utilizations (e.g., cause a different internal temperature, use a different percentage of processor or memory capacity) across different managed nodes 1260. The orchestrator server 1240, in the illustrative embodiment, is configured to determine the differences based on the telemetry data stored in the hierarchical model and factor the differences into a prediction of future resource utilization of a workload if the workload is reassigned from one managed node to another managed node. By taking into account these differences, the orchestrator server 1240 may more accurately balance resource utilization among the workloads and increase the achievement of one or more of the resource allocation objectives without decreasing the achievement of any of the other resource allocation objectives. In the illustrative embodiment, the achievement of a resource allocation objective may be measured, equal to, or otherwise defined as the degree to which a measured value from one or more managed nodes 1260 satisfies a target value associated with the resource allocation objective. For example, in the illustrative embodiment, increasing the achievement may be performed by decreasing the error (e.g., difference) between the measured value (e.g., an operating temperature of a managed node 1260) and the target value (e.g., a target operating temperature). Conversely, decreasing the achievement may be performed by increasing the error (e.g., difference) between the measured value and the target value.

Referring now to FIG. 13, the orchestrator server 1240 may be embodied as any type of compute device capable of performing the functions described herein, including issuing a request to have cloud services performed, receiving results of the cloud services, assigning workloads to compute devices, analyzing telemetry data indicative of performance and conditions (e.g., resource utilization, one or more temperatures, fan speeds, etc.) as the workloads are executed, generating a hierarchical model indicative of a relationship between the managed nodes and differences in workload resource utilizations among the managed nodes, and adjusting the assignments of the workloads to balance resource utilization and manage the achievement of multiple resource allocation objectives as the workloads are performed, using the hierarchical model. For example, the orchestrator server 1240 may be embodied as a computer, a distributed computing system, one or more sleds (e.g., the sleds 204-1, 204-2, 204-3, 204-4, etc.), a server (e.g., stand-alone, rack-mounted, blade, etc.), a multiprocessor system, a network appliance (e.g., physical or virtual), a desktop computer, a workstation, a laptop computer, a notebook computer, a processor-based system, or a network appliance. As shown in FIG. 13, the illustrative orchestrator server 1240 includes a central processing unit (CPU) 1302, a main memory 1304, an input/output (I/O) subsystem 1306, communication circuitry 1308, and one or more data storage devices 1312. Of course, in other embodiments, the orchestrator server 1240 may include other or additional components, such as those commonly found in a computer (e.g., display, peripheral devices, etc.). Additionally, in some embodiments, one or more of the illustrative components may be incorporated in, or otherwise form a portion of, another component. For example, in some embodiments, the main memory 1304, or portions thereof, may be incorporated in the CPU 1302.

The CPU 1302 may be embodied as any type of processor capable of performing the functions described herein. The CPU 1302 may be embodied as a single or multi-core processor(s), a microcontroller, or other processor or processing/controlling circuit. In some embodiments, the CPU 1302 may be embodied as, include, or be coupled to a field programmable gate array (FPGA), an application specific integrated circuit (ASIC), reconfigurable hardware or hardware circuitry, or other specialized hardware to facilitate performance of the functions described herein. As discussed above, the managed node 1260 may include resources distributed across multiple sleds and in such embodiments, the CPU 1302 may include portions thereof located on the same sled or different sled. Similarly, the main memory 1304 may be embodied as any type of volatile (e.g., dynamic random access memory (DRAM), etc.) or non-volatile memory or data storage capable of performing the functions described herein. In some embodiments, all or a portion of the main memory 1304 may be integrated into the CPU 1302. In operation, the main memory 1304 may store various software and data used during operation such as telemetry data, resource allocation objective data, workload labels, workload classifications, a hierarchical model, workload adjustment data, operating systems, applications, programs, libraries, and drivers. As discussed above, the managed node 1260 may include resources distributed across multiple sleds and in such embodiments, the main memory 1304 may include portions thereof located on the same sled or different sled.

The I/O subsystem 1306 may be embodied as circuitry and/or components to facilitate input/output operations with the CPU 1302, the main memory 1304, and other components of the orchestrator server 1240. For example, the I/O subsystem 1306 may be embodied as, or otherwise include, memory controller hubs, input/output control hubs, integrated sensor hubs, firmware devices, communication links (e.g., point-to-point links, bus links, wires, cables, light guides, printed circuit board traces, etc.), and/or other components and subsystems to facilitate the input/output operations. In some embodiments, the I/O subsystem 1306 may form a portion of a system-on-a-chip (SoC) and be incorporated, along with one or more of the CPU 1302, the main memory 1304, and other components of the orchestrator server 1240, on a single integrated circuit chip.

The communication circuitry 1308 may be embodied as any communication circuit, device, or collection thereof, capable of enabling communications over the network 1230 between the orchestrator server 1240 and another compute device (e.g., the client device 1220, and/or the managed nodes 1260). The communication circuitry 1308 may be configured to use any one or more communication technology (e.g., wired or wireless communications) and associated protocols (e.g., Ethernet, Bluetooth®, Wi-Fi®, WiMAX, etc.) to effect such communication.

The illustrative communication circuitry 1308 includes a network interface controller (NIC) 1310, which may also be referred to as a host fabric interface (HFI). The NIC 1310 may be embodied as one or more add-in-boards, daughtercards, network interface cards, controller chips, chipsets, or other devices that may be used by the orchestrator server 1240 to connect with another compute device (e.g., the client device 1220 and/or the managed nodes 1260). In some embodiments, the NIC 1310 may be embodied as part of a system-on-a-chip (SoC) that includes one or more processors, or included on a multichip package that also contains one or more processors. In some embodiments, the NIC 1310 may include a local processor (not shown) and/or a local memory (not shown) that are both local to the NIC 1310. In such embodiments, the local processor of the NIC 1310 may be capable of performing one or more of the functions of the CPU 1302 described herein. Additionally or alternatively, in such embodiments, the local memory of the NIC 1310 may be integrated into one or more components of the orchestrator server 1240 at the board level, socket level, chip level, and/or other levels. As discussed above, the managed node 1260 may include resources distributed across multiple sleds and in such embodiments, the communication circuitry 1308 may include portions thereof located on the same sled or different sled.

The one or more illustrative data storage devices 1312, may be embodied as any type of devices configured for short-term or long-term storage of data such as, for example, memory devices and circuits, memory cards, hard disk drives, solid-state drives, or other data storage devices. Each data storage device 1312 may include a system partition that stores data and firmware code for the data storage device 1312. Each data storage device 1312 may also include an operating system partition that stores data files and executables for an operating system.

Additionally, the orchestrator server 1240 may include a display 1314. The display 1314 may be embodied as, or otherwise use, any suitable display technology including, for example, a liquid crystal display (LCD), a light emitting diode (LED) display, a cathode ray tube (CRT) display, a plasma display, and/or other display usable in a compute device. The display 1314 may include a touchscreen sensor that uses any suitable touchscreen input technology to detect the user's tactile selection of information displayed on the display including, but not limited to, resistive touchscreen sensors, capacitive touchscreen sensors, surface acoustic wave (SAW) touchscreen sensors, infrared touchscreen sensors, optical imaging touchscreen sensors, acoustic touchscreen sensors, and/or other type of touchscreen sensors.

Additionally or alternatively, the orchestrator server 1240 may include one or more peripheral devices 1316. Such peripheral devices 1316 may include any type of peripheral device commonly found in a compute device such as speakers, a mouse, a keyboard, and/or other input/output devices, interface devices, and/or other peripheral devices.

The client device 1220 and the managed nodes 1260 may have components similar to those described in FIG. 13. The description of those components of the orchestrator server 1240 is equally applicable to the description of components of the client device 1220 and the managed nodes 1260 and is not repeated herein for clarity of the description. Further, it should be appreciated that any of the client device 1220 and the managed nodes 1260 may include other components, sub-components, and devices commonly found in a computing device, which are not discussed above in reference to the orchestrator server 1240 and not discussed herein for clarity of the description.

As described above, the client device 1220, the orchestrator server 1240 and the managed nodes 1260 are illustratively in communication via the network 1230, which may be embodied as any type of wired or wireless communication network, including global networks (e.g., the Internet), local area networks (LANs) or wide area networks (WANs), cellular networks (e.g., Global System for Mobile Communications (GSM), 3G, Long Term Evolution (LTE), Worldwide Interoperability for Microwave Access (WiMAX), etc.), digital subscriber line (DSL) networks, cable networks (e.g., coaxial networks, fiber networks, etc.), or any combination thereof.

Referring now to FIG. 14, in the illustrative embodiment, the orchestrator server 1240 may establish an environment 1400 during operation. The illustrative environment 1400 includes a network communicator 1420, a telemetry monitor 1430, and a resource manager 1440. Each of the components of the environment 1400 may be embodied as hardware, firmware, software, or a combination thereof. As such, in some embodiments, one or more of the components of the environment 1400 may be embodied as circuitry or a collection of electrical devices (e.g., network communicator circuitry 1420, telemetry monitor circuitry 1430, resource manager circuitry 1440, etc.). It should be appreciated that, in such embodiments, one or more of the network communicator circuitry 1420, telemetry monitor circuitry 1430, or resource manager circuitry 1440 may form a portion of one or more of the CPU 1302, the main memory 1304, the I/O subsystem 1306, and/or other components of the orchestrator server 1240. In the illustrative embodiment, the environment 1400 includes telemetry data 1402 which may be embodied as data indicative of the performance and conditions (e.g., resource utilization, operating frequencies, power usage, one or more temperatures, fan speeds, etc.) of each managed node 1260 as the managed nodes 1260 execute the workloads assigned to them. Additionally, the illustrative environment 1400 includes resource allocation objective data 1404 indicative of user-defined thresholds or goals (“objectives”) to be satisfied during the execution of the workloads. In the illustrative embodiment, the objectives pertain to power consumption, life expectancy, heat production, and performance of the components of the managed nodes 1260. Further, the illustrative environment 1400 includes workload labels 1406 which may be embodied as any identifiers (e.g., process numbers, executable file names, alphanumeric tags, etc.) that uniquely identify each workload executed by the managed nodes 1260. In addition, the illustrative environment 1400 includes workload classifications 1408 which may be embodied as any data indicative of the general resource utilization tendencies of each workload (e.g., processor intensive, memory intensive, network bandwidth intensive, etc.).

In the illustrative embodiment, the environment 1400 also includes a hierarchical model 1410, which may be embodied as any data indicative of the telemetry data 1402 associated with each managed node 1260, a relationship between the nodes, such as a spatial relationship or a functional relationship, and differences between the resource utilization of a workload when executed by one managed node as compared to a reference (e.g., an average) resource utilization of the workload. For example, the spatial relationship may be indicative of the physical location of each managed node 1260 (e.g., a location in the data center 1100), such as an access pathway number, rack number, and sled space index, or a set of spatial coordinates (e.g., latitude, longitude, and altitude, or coordinates relative to a known location in the data center). A functional relationship may be embodied as an indication that a particular managed node 1260 is a member of a group of managed nodes 1260 that perform a related set of functions (e.g., functions for a particular customer or class of customers, functions that have similar resource utilization profiles, etc.). The differences between the resource utilization of a workload executed by one managed node 1260 and a reference resource utilization indicative of resource utilization of the workload on a reference managed node 1260 may be embodied as weights (e.g., coefficients) and may be a result of the idiosyncrasies of the hardware components of the managed node 1260 and/or its relationship to the other managed nodes 1260 (e.g., the managed node 1260 is located in a portion of the data center 1100 that typically has an above-average ambient temperature, etc.).

Further, the illustrative environment 1400 includes workload adjustment data 1412 which may be embodied as any data indicative of reassignments (e.g., live migrations) of one or more workloads from one managed node 1260 to another managed node 1260 and/or adjustments to settings for components within each managed node 1260, such as target power usage of the components, processor capacity (e.g., a number of cores to be used, a clock speed, a percentage of available processor cycles, etc.) available to one or more workloads, memory resource capacity (e.g., amount of memory to be used and/or frequency of memory accesses to volatile memory and/or non-volatile memory) available to one or more workloads, communication circuitry capacity (e.g., network bandwidth) available to one or more workloads, and/or target operating temperatures and fan speeds.

In the illustrative environment 1400, the network communicator 1420, which may be embodied as hardware, firmware, software, virtualized hardware, emulated architecture, and/or a combination thereof as discussed above, is configured to facilitate inbound and outbound network communications (e.g., network traffic, network packets, network flows, etc.) to and from the orchestrator server 1240, respectively. To do so, the network communicator 1420 is configured to receive and process data packets from one system or computing device (e.g., the client device 1220) and to prepare and send data packets to another computing device or system (e.g., the managed nodes 1260). Accordingly, in some embodiments, at least a portion of the functionality of the network communicator 1420 may be performed by the communication circuitry 1308, and, in the illustrative embodiment, by the NIC 1310.

The telemetry monitor 1430, which may be embodied as hardware, firmware, software, virtualized hardware, emulated architecture, and/or a combination thereof as discussed above, is configured to collect the telemetry data 1402 from the managed nodes 1260 as the managed nodes 1260 execute the workloads assigned to them. The telemetry monitor 1430 may actively poll each of the managed nodes 1260 for updated telemetry data 1402 on an ongoing basis or may passively receive telemetry data 1402 from the managed nodes 1260, such as by listening on a particular network port for updated telemetry data 1402. The telemetry monitor 1430 may further parse and categorize the telemetry data 1402, such as by separating the telemetry data 1402 into an individual file or data set for each managed node 1260.

The resource manager 1440, which may be embodied as hardware, firmware, software, virtualized hardware, emulated architecture, and/or a combination thereof, is configured to generate data analytics from the telemetry data 1402, identify the workloads, classify the workloads, generate the hierarchical model 1410, determine whether the assignment of the workloads can be adjusted in view of the differences in resource utilizations of workload across different managed nodes 1260 to improve the achievement of at least one resource allocation objective (e.g., increase resource utilization) without adversely affecting the achievement of any other resource allocation objectives, and apply the adjustments. To do so, the resource manager 1440 includes a hierarchical modeler 1442 that includes a workload labeler 1444, a workload classifier 1446, a workload behavior predictor 1448, and a multi-objective analyzer 1450. The hierarchical modeler 1442, in the illustrative embodiment, is configured to generate the hierarchical model 1410 and coordinate the functions of the workload labeler 1444, the workload classifier 1446, the workload behavior predictor 1448, and the multi-objective analyzer 1450 to more accurately determine and apply adjustments to the assignments to the workloads using information indicative of the differences in how a workload affects the resource utilization on one managed node 1260 versus another managed node 1260 that may be located in a different place (e.g., a cooler section of the data center 1100) or include different hardware components (e.g., more memory, a more efficient processor, etc). The workload labeler 1444, in the illustrative embodiment, is configured to assign a workload label 1406 to each workload presently performed or scheduled to be performed by one or more of the managed nodes 1260. The workload labeler 1444 may generate the workload label 1406 as a function of an executable name of the workload, a hash of all or a portion of the code of the workload, or based on any other method to uniquely identify each workload. The workload classifier 1446, in the illustrative embodiment, is configured to categorize each labeled workload based on the average resource utilization of each workload (e.g., generally utilizes 65% of processor capacity, generally utilizes 40% of memory capacity, etc.).

The workload behavior predictor 1448, in the illustrative embodiment, is configured to analyze the telemetry data 1402 to identify different phases of resource utilization within the telemetry data 1402 for each workload. Each resource utilization phase may be embodied as a period of time in which the resource utilization of a one or more resources of a managed node satisfies a predefined threshold. For example, a utilization of at least 85% of the available processor capacity may be indicative of a high processor utilization phase, and a utilization of at least 85% of the memory capacity may be indicative of a high memory utilization phase. In the illustrative embodiment, the workload behavior predictor 1448 is further to identify patterns in the resource utilization phases of the workloads (e.g., a high processor utilization phase, followed by a high memory utilization phase, followed by a phase of low resource utilization, which is then followed by the high processor utilization phase again). The workload behavior predictor 1448 may be configured to utilize the identifications of the resource utilization phase patterns, determine a present resource utilization phase of a given workload, predict the next resource utilization phase based on the patterns, and determine an amount of remaining time until the workload transitions to the next resource utilization phase. The multi-objective analyzer 1450, in the illustrative embodiment, is configured to balance the resource allocation objectives defined in the resource allocation objective data 1404, determine, based on the telemetry data 1402, whether the present allocation of the resources in the managed nodes 1260 is Pareto-efficient (e.g., that no adjustment can be made without decreasing the achievement of one or more other resource allocation objectives), and if not, determine an adjustment that provides a Pareto improvement (e.g., an increase in the achievement of at least one of the objectives without decreasing the achievement of any of the other objectives). To do so, in the illustrative embodiment, the multi-objective analyzer 1450 includes a workload placer 1452 and a node settings adjuster 1454.

In the illustrative embodiment, the multi-objective analyzer 1450 is configured to determine, as a function of the telemetry data 1402, including the present resource utilizations of the workloads and the predicted behavior of the workloads, whether an adjustment can be made to any of the assignments of the workloads and/or the settings of the components of the managed nodes 1260 to increase the achievement of one or more of the resource allocation objectives without decreasing the achievement of the other resource allocation objectives. In the illustrative embodiment, the multi-objective analyzer 1450 may do so by modeling or simulating the set of managed nodes 1260, including simulating the differences (e.g., weights), represented in the hierarchical model 1410, indicative of how a workload affects the resource utilization on one managed node 1260 versus another managed node 1260, to determine the power consumption, heat generation, compute capacity, and other factors in response to various adjustments to the assignments of workloads and/or the settings of components within the managed nodes 1260, define a Pareto frontier indicative of a set of resource allocations that are all Pareto-efficient, determine whether the present resource allocation is already on the Pareto frontier, and if not, what adjustment to the allocations would reach the Pareto frontier. The multi-objective analyzer 1450 may determine the Pareto frontier, P(Y), as follows:

ƒ: Rⁿ→R^m  (Equation 1)

In the above equation, ƒ is a function of the set of managed nodes 1260, modeled by the multi-objective analyzer 1450, that is indicative of the response of the managed nodes 1260 to adjustments to the assignments of workloads. Rⁿ is a metric space of possible allocations (i.e., assignments of workloads) and R^m represents a set of criterion vectors. In the following equation, X is a compact set of feasible decisions in the metric space, Rⁿ, and Y is the feasible set of criterion vectors in R^m, as follows:

Y={yεR^m: y=ƒ(x),xεX}  (Equation 2)

Furthermore, a point y″ defined in Equation 3 below dominates another point y′ defined in Equation 4.

y″εR^m  (Equation 3)

y′εR^m  (Equation 4)

As such, the Pareto frontier may be represented as follows:

P(Y)={y′εY:{y″εY: y″>y′,y″≠y′}=Ø}  (Equation 5)

The workload placer 1452, in the illustrative embodiment, is configured to initially assign workloads to the various managed nodes 1260 and reassign the workloads among the managed nodes 1260 to provide a Pareto improvement (e.g., an adjustment that improves the achievement of at least one resource allocation objective without decreasing the achievement of any other resource allocation objectives). In doing so, the workload placer 1452 may determine reassignments of workloads among the managed nodes 1260 and/or time offsets to apply to workloads to align the timing of the resource utilization phases identified by the workload behavior predictor 1448. Similarly, the node settings adjuster 1454, in the illustrative embodiment, is configured to apply one or more adjustments to the settings within the managed nodes 1260 to provide or restrict the resources available to the workloads in accordance with the determined Pareto improvement. The settings may be associated with the operating system and/or the firmware or drivers of the components of the managed nodes 1260.

It should be appreciated that each of the hierarchical modeler 1442, workload labeler 1444, the workload classifier 1446, the workload behavior predictor 1448, the multi-objective analyzer 1450, the workload placer 1452, and the node settings adjuster 1454 may be separately embodied as hardware, firmware, software, virtualized hardware, emulated architecture, and/or a combination thereof. For example, the hierarchical modeler 1442 may be embodied as a hardware component, while the workload labeler 1444, the workload classifier 1446, the workload behavior predictor 1448, the multi-objective analyzer 1450, the workload placer 1452, and the node settings adjuster 1454 are embodied as virtualized hardware components or as some other combination of hardware, firmware, software, virtualized hardware, emulated architecture, and/or a combination thereof.

Referring now to FIG. 15, in use, the orchestrator server 1240 may execute a method 1500 for managing the assignment of workloads among the managed nodes 1260 based on the hierarchical model 1410. The method 1500 begins with block 1502, in which the orchestrator server 1240 determines whether to manage workloads. In the illustrative embodiment, the orchestrator server 1240 determines to manage workloads if the orchestrator server 1240 is powered on, in communication with the managed nodes 1260, and has received at least one request from the client device 1220 to provide cloud services (i.e., to perform one or more workloads). In other embodiments, the orchestrator server 1240 may determine whether to manage workloads based on other factors. Regardless, in response to a determination to manage workloads, in the illustrative embodiment, the method 1500 advances to block 1504 in which the orchestrator server 1240 receives resource allocation objective data (e.g., the resource allocation objective data 1404). In doing so, the orchestrator server 1240 may receive the resource allocation objective data 1404 from a user (e.g., an administrator) through a graphical user interface (not shown), from a configuration file, or from another source. In receiving the resource allocation objective data 1404, the orchestrator server 1240 may receive power consumption objective data indicative of a target power usage or threshold amount of power usage of the managed nodes 1260 as they execute the workloads, as indicated in block 1506. The orchestrator server 1240, in the illustrative embodiment, may also receive performance objective data, indicative of a target speed at which workloads are to be executed (e.g., a processor clock speed, a memory clock speed, I/O operations per second, a target time period in which to complete execution of a workload, etc.), as indicated in block 1508. Additionally or alternatively, the orchestrator server 1240 may receive reliability objective data indicative of a target life cycle of one or more of the managed nodes 1260 or components therein (e.g., a target life cycle of a data storage device, a target life cycle of a cooling fan, etc.), as indicated in block 1510. As indicated in block 1512, the orchestrator server 1240 may also receive thermal objective data indicative of one or more target temperatures in the managed nodes 1260.

After receiving the resource allocation objective data 1404, in the illustrative embodiment, the method 1500 advances to block 1514 in which the orchestrator server 1240 assigns initial workloads to the managed nodes 1260. In the illustrative embodiment, the orchestrator server 1240 has not received telemetry data 1402 that would inform a decision as to where the workloads are to be assigned among the managed nodes 1260. As such, the orchestrator server 1240 may assign the workloads to the managed nodes 1260 based on any suitable method, such as by assigning each workload to the first available managed node that is idle (i.e., is not presently executing a workload), randomly assigning the workloads, or by any other method.

Having assigned the workloads, the method 1500 advances to block 1516 in which the orchestrator server 1240 receives telemetry data 1402 from the managed nodes 1260 as the workloads are performed (i.e., executed). As described herein, depending on the idiosyncrasies of the components and location of each managed node 1260, the same workload may affect one managed node 1260 differently than it would affect another managed node 1260. By collecting the telemetry data 1402, the orchestrator server 1240 may later determine the differences in how the same workload affects managed nodes 1260 differently, as described herein. In receiving the telemetry data 1402, the orchestrator server 1240 may receive power consumption data indicative of an amount of power (e.g., Watts) consumed by each managed node 1260, as indicated in block 1518. The orchestrator server 1240 may also receive performance data from each managed node 1260 indicative of a speed at which the workloads are performed, as indicated in block 1520. As such, the performance data may be embodied as an amount of time consumed to complete a function or task associated with a workload. Further, as indicated in block 1522, the orchestrator server 1240 may receive processor utilization data indicative of an amount of processor usage consumed by each workload performed by each managed node 1260. Moreover, as indicated in block 1524, the orchestrator server 1240 may receive memory utilization data for each managed node 1260. The memory utilization data may be embodied as Intel Cache Allocation Technology (CAT) data, Intel Cache Monitoring Technology (CMT) data, Intel Memory Bandwidth Monitoring (MBM) data, and/or other data indicative of an amount or frequency of memory use by each workload performed by each managed node 1260.

In receiving the memory utilization data, the orchestrator server 1240 may receive cache utilization data indicative of a frequency of cache accesses associated with a workload and/or cache miss rate information, as indicated in block 1526. Additionally or alternatively, as indicated in block 1528, in receiving the memory utilization data, the orchestrator server 1240 may receive volatile memory utilization data indicative of an amount volatile memory (e.g., the main memory 1304) used, a frequency of accesses to the volatile memory, page fault data, and/or other information indicative of the utilization of the volatile memory within each managed node 1260. The orchestrator server 1240 may additionally or alternatively receive non-volatile memory utilization data indicative of the amount of data stored and/or retrieved from the data storage devices 1312 and/or a frequency at which each workload issues write requests and/or read requests to the data storage devices 1312 in each managed node 1260, as indicated in block 1530.

In receiving the telemetry data 1402, the orchestrator server 1240 may additionally or alternatively receive network utilization data indicative of an amount of network bandwidth (e.g., capacity of the communication circuitry) used by each workload performed by each managed node 1260, as indicated in block 1532. The orchestrator server 1240 may additionally receive temperature data from each managed node 1260 indicative of one or more temperatures within the managed nodes 1260, as indicated in block 1534. After receiving the telemetry data 1402, the method 1500 advances to block 1536 of FIG. 16, in which the orchestrator server 1240 generates the hierarchical model 1410 from the telemetry data 1402 as the workloads are being performed.

Referring now to FIG. 16, in generating the hierarchical model 1410, the orchestrator server 1240 identifies each managed node 1260 associated with the telemetry data 1402, as indicated in block 1538. In the illustrative embodiment, the orchestrator server 1240 may associate an identifier (e.g., an Internet Protocol (IP) address, a media access control (MAC) address, or other unique identifier) included in the telemetry data 1402 with each corresponding managed node 1260. Additionally, in the illustrative embodiment, the orchestrator server 1240 stores the telemetry data 1402 in the hierarchical model 1410 in a format indicative of a relationship between the managed nodes 1260. In doing so, as indicated in block 1542, the orchestrator server 1240 may store the telemetry data 1402 in the hierarchical model 1410 in a format indicative of spatial locations of the managed nodes 1260. For example, and with reference to the diagram 1800 shown in FIG. 18, the hierarchical model 1410 may be organized with the orchestrator server 1240 as a root 1810, with access pathways 1820, 1822 associated with the root, racks 1830, 1832, associated with the access pathways 1820, 1822, and sled space indexes 1840, 1842 associated with the racks 1830, 1832. Alternatively, the spatial relationship may include the data center 1100 as a root and iteratively smaller subsections of the data center 1100 connected in a hierarchy. In the illustrative embodiment, the orchestrator server 1240 stores the telemetry data 1402 for each managed node 1260 in association with the location of the managed node 1260 in the hierarchy (e.g., stored in association with the entry indicative of the sled space index of the corresponding managed node 1260, an entry indicative of spatial coordinates where the corresponding managed node 1260 is physically located, etc.). The orchestrator server 1240 may alternatively store the telemetry data 1402 in the hierarchical model 1410 in a format indicative of a membership of each managed node 1260 in a corresponding functional group, as indicated in block 1544. Each functional group may be embodied as a set of managed nodes 1260 associated with a unique identifier (e.g., an alphanumeric code), that perform similar functions (e.g., encryption, compression, etc.), perform different functions for the same customer or class of customer (e.g., a customer who paid for a particular quality of service (QoS)), or based on another functional relationship. The storage of the telemetry data 1402 in the format indicative of the relationship between the managed nodes 1260 is the beginning of the formation of the hierarchical model 1410. In the illustrative embodiment, the orchestrator server 1240 subsequently generates data analytics which enables the orchestrator server 1240 to add additional data (e.g., weights 1850 shown in FIG. 18) to the hierarchical model 1410 indicative of the differences in how each managed node 1260 is affected by workloads, as described herein.

As indicated in block 1546, the orchestrator server 1240, in the illustrative embodiment, generates data analytics as the workloads are performed. In generating the data analytics, the orchestrator server 1240, in the illustrative embodiment, generates profiles of the workloads, as indicated in block 1548. In doing so, in the illustrative embodiment, the orchestrator server 1240 generates the labels 1406 for the workloads to uniquely identify each workload, as indicated in block 1550. Additionally, in the illustrative embodiment, the orchestrator server 1240 generates the classifications 1408 of the workloads, as indicated in block 1552. Further, in the illustrative embodiment, the orchestrator server 1240 identifies the patterns in the resource utilization phases for each of the workloads, as indicated in block 1554. In doing so, the orchestrator server 1240 may determine that a particular workload experiences a phase of high processor utilization and low memory utilization that is typically followed by a phase of low processor utilization and high memory utilization, and that another workload experiences similar phases, but at a different frequency or at a time offset from the other workload.

Additionally, as indicated in block 1556, the orchestrator server 1240 may predict future resource utilization phases of the workloads, such as by comparing a present resource utilization of each workload to the identified patterns in the resource utilization phases to determine the present phase of each workload, and then identifying the upcoming phases of the workloads from the patterns. In doing so, as indicated in block 1558, the orchestrator server 1240 may predict future contention for resources. For example, the orchestrator server 1240 may identify two workloads executed concurrently by the same managed node 1260 that are predicted to both enter high processor utilization phases, causing contention for that resource of the managed node 1260. Similarly, the orchestrator server 1240 may identify workloads executed by a particular managed node 1260 that are predicted to concurrently enter phases of high memory utilization or high network bandwidth utilization, causing contention for those resources. In doing so, the orchestrator server 1240 may determine whether, at the heightened resource utilization, the available capacity of the resource subject to contention will be less than the amount requested by the workloads executed by that managed node 1260, and if so, store an indicator to potentially adjust the assignment of those workloads among the managed nodes 1260. In block 1560, the orchestrator server 1240 stores the data analytics in association with each of the managed nodes 1260 represented in the hierarchical model 1410.

In block 1562, the orchestrator server 1240, in the illustrative embodiment, determines differences in the data analytics for each workload as a function of the managed node 1260 that performed the workload. For example, while a particular labeled workload, or workloads that have the same classification 1408, should exhibit the same resource utilization on any given managed node 1260 if the managed nodes 1260 were identical and located in exactly the same place, the data analytics may indicate that the resource utilizations actually differed (e.g., due to differences in the components of the managed nodes 1260 and/or their relationship to other managed nodes 1260, such as being near other managed nodes that generate an above-average amount of heat). The orchestrator server 1240, in the illustrative embodiment, may determine a reference amount of resource utilization (e.g., an average across the managed nodes 1260 that performed the workload) for both the classification (e.g., the general resource utilization) of the workload and each resource utilization phase of the workload, and then determine differences from those reference amounts based on the telemetry data reported by each managed node 1260 that executed the workload. As indicated in block 1564, the orchestrator server 1240 may then store the determined differences as weights 1850 (e.g., coefficients) in the hierarchical model 1410 in association with each managed node 1260 represented therein. As an example, one managed node 1260 located in a particularly hot area of the data center 1100 (e.g., near managed nodes that perform high processor utilization workloads) may report a higher fan utilization than another managed node 1260 located in a relatively cooler section of the data center 1100 for the same workload (e.g., having the same workload label 1406) or same workload classification (e.g., having the same workload classification 1408). In such an example, the orchestrator server 1240 may store a fan utilization weight of 1.2 for the managed node 1260 in the hot area of the data center and a fan utilization weight of 0.8 for the other managed node 1260 in the cooler area.

After generating the data analytics, the method 1500 advances to block 1566, in which the orchestrator server 1240 determines whether to adjust the workload assignments. In doing so, the orchestrator server 1240, in the illustrative embodiment, determines whether the workload assignments are Pareto-efficient, as indicated in block 1568. In the illustrative embodiment, in determining whether the workload assignments are Pareto-efficient the orchestrator server 1240 determines whether an adjustment can be made to the workload assignments to increase the achievement of a resource allocation objective (e.g., a target performance), without decreasing the achievement of any other resource allocation objective (e.g., a target temperature, a target power consumption, etc.). As described above, the orchestrator server 1240 may determine the Pareto frontier based on a model of the reactions of the managed nodes 1260 to adjustments to the assignment of workloads, and determine whether the present state of the allocation of resources is already on the Pareto frontier. Afterwards, the method 1500 advances to block 1570 of FIG. 17, in which the orchestrator server 1240 determines the subsequent steps based on whether the orchestrator server 1240 determined to adjust the workload assignments.

Referring now to FIG. 17, if the orchestrator server 1240 determines not to adjust the workload assignments, the method 1500 loops back to block 1516 of FIG. 15, in which the orchestrator server 1240 again receives telemetry data from the managed nodes 1260 as the workloads continue to be performed. Otherwise, the method 1500 advances to block 1572, in which the orchestrator server 1240 determines adjustments to the workload assignments.

Referring now to FIG. 17, the illustrative orchestrator server 1240, in block 1572, determines, as a function of the data analytics and the resource allocation objective data 1404, adjustments to the workload assignments as the workloads are performed. In doing so, the orchestrator server 1240 determines one or more adjustments to improve the achievement of at least one of the resource allocation objectives without decreasing the achievement of any of the other resource allocation objectives, as indicated in block 1574. As indicated in block 1576, the orchestrator server 1240 determines workload reassignments, to move at least one workload from one managed node 1260 to another managed node 1260. In doing so, the orchestrator server 1240 predicts, as function of the node-specific weights determined in block 1564 of FIG. 16, a change in the resource utilization of a workload to be reassigned from one managed node 1260 (e.g., a source managed node 1260) to another managed node 1260 (e.g., a destination managed node 1260). For example, if the hierarchical model 1410 indicates that the source managed node 1260 has weight of 1.6 associated with processor utilization (e.g., because the processor in the source managed node 1260 is less efficient than an average processor in the managed nodes 1260), and the hierarchical model 1410 indicates that the destination managed node 1260 has a weight of 1 associated with processor utilization (e.g., because the processor in the destination managed node 1260 has an average efficiency compared to the other managed nodes 1260), then the orchestrator server 1240 may predict that the processor utilization for the workload will decrease by approximately 38% when the workload is reassigned to the destination managed node 1260.

The orchestrator server 1240 may additionally determine node-specific adjustments, as indicated in block 1580. The node-specific adjustments may be embodied as changes to settings within one or more of the managed nodes 1260, such as in the operating system, the drivers, and/or the firmware of components (e.g., the CPU 1302, the memory 1304, the communication circuitry 1308, the one or more data storage devices 1312, etc.) to improve resource utilization. As such, in the illustrative embodiment, in determining the node-specific adjustments, the orchestrator server 1240 may determine processor throttle adjustments, such as clock speed and/or processor affinity for one or more workloads, as indicated in block 1582. Additionally or alternatively, the orchestrator server 1240 may determine memory usage adjustments, such as allocations of volatile memory (e.g., the memory 1304) and/or data storage capacity (e.g., capacity of the one or more data storage devices 1312), memory bus speeds, and/or other memory-related settings, as indicated in block 1584. Additionally or alternatively, the orchestrator server 1240 may determine one or more fan speed adjustments to increase or decrease the cooling within the managed node 1260, as indicated in block 1586. Additionally or alternatively, the orchestrator server 1240 may determine network bandwidth adjustments, such as an available bandwidth of the communication circuitry 1308 to be allocated to each workload in the managed node 1260, as indicated in block 1588.

After determining the adjustments, the method advances to block 1590 in which the orchestrator server 1240 applies the determined adjustments. In doing so, the orchestrator server 1240 may issue one or more requests to perform a live migration of a workload between two managed nodes 1260 (i.e., a workload reassignment), as indicated in block 1592. In the illustrative embodiment, the migration is live because, rather than waiting until the workloads have been completed to analyze the telemetry data 1402, the orchestrator server 1240 collects and analyzes the telemetry data 1402, and makes adjustments online (i.e., as the workloads are being performed), as described above. The orchestrator server 1240 may also issue requests to suspend or resume execution of workloads, as indicated in block 1594. In doing so, the orchestrator server 1240 may suspend the execution of a workload until a determined time offset has elapsed, then resume execution, thereby shifting a resource utilization phase in time, to align with a complementary resource utilization phase of another workload executed on the same managed node 1260. In the illustrative embodiment, complementary resource utilization phases may be embodied as two or more resource utilization phases of different workloads that primarily use different resources of a managed node 1260, such that concurrent execution of the phases does not cause resource contention. For example, a high processor and low memory utilization phase may be complementary with a low processor and high memory utilization phase. In applying the determined adjustments, as indicated in block 1596, the orchestrator server 1240 may also issue one or more requests to one or more of the managed nodes 1260 to apply the node-specific adjustments described above with reference to block 1580. After applying the adjustments, the method 1500 loops back to block 1516 of FIG. 15, in which the orchestrator server 1240 receives additional telemetry data 1402 from the managed nodes 1260.

EXAMPLES

Illustrative examples of the technologies disclosed herein are provided below. An embodiment of the technologies may include any one or more, and any combination of, the examples described below.

Example 1 includes an orchestrator server to allocate resources of a set of managed nodes to workloads with a hierarchical model, the orchestrator server comprising one or more processors; one or more memory devices having stored therein a plurality of instructions that, when executed by the one or more processors, cause the orchestrator server to receive resource allocation objective data indicative of multiple resource allocation objectives to be satisfied; determine an initial assignment of a set of workloads among the managed nodes; receive telemetry data from the managed nodes, wherein the telemetry data is indicative of resource utilization by the managed nodes as the workloads are performed; generate, as a function of the telemetry data, a hierarchical model indicative of the resource utilization of each managed node and a relationship between the managed nodes; determine, with the hierarchical model, differences in resource utilization for each workload as a function of the managed node that performed the workload; determine, as a function of the telemetry data and the determined differences, an adjustment to the assignment of the workloads to increase an achievement of at least one of the resource allocation objectives without decreasing the achievement of any of the other resource allocation objectives; and apply the adjustment to the assignments of the workloads among the managed nodes as the workloads are performed.

Example 2 includes the subject matter of Example 1, and wherein to generate a hierarchical model comprises to generate a hierarchical model indicative of a spatial relationship between the managed nodes.

Example 3 includes the subject matter of any of Examples 1 and 2, and wherein to generate a hierarchical model comprises to generate a hierarchical model indicative of a membership of each managed node in a functional group.

Example 4 includes the subject matter of any of Examples 1-3, and wherein the plurality of instructions, when executed, further cause the orchestrator server to determine node-specific weights indicative of the determined differences in resource utilization; and wherein to determine the adjustment to the workload assignments comprises to predict a change in resource utilization of a workload as a function of the node-specific weights associated with two different managed nodes.

Example 5 includes the subject matter of any of Examples 1-4, and wherein the plurality of instructions, when executed, further cause the orchestrator server to generate data analytics as a function of the telemetry data; and store the data analytics in association with each managed node in the hierarchical model.

Example 6 includes the subject matter of any of Examples 1-5, and wherein to generate the data analytics comprises to generate profiles of the workloads, wherein the profiles are indicative of an identity of each workload and a resource usage classification of each workload.

Example 7 includes the subject matter of any of Examples 1-6, and wherein to generate the data analytics comprises to predict future resource utilization of the workloads.

Example 8 includes the subject matter of any of Examples 1-7, and wherein to predict future resource utilization of the workloads comprises to identify potential contention for resources.

Example 9 includes the subject matter of any of Examples 1-8, and wherein to receive resource allocation objective data comprises to receive two or more of power consumption objective data indicative of a target power usage of one or more of the managed nodes, performance objective data indicative of a target speed at which to perform the workloads, reliability objective data indicative of a target life cycle of one or more of the managed nodes, or thermal objective data indicative of a target temperature of one or more of the managed nodes.

Example 10 includes the subject matter of any of Examples 1-9, and wherein to receive telemetry data from the managed nodes comprises to receive at least one of power consumption data indicative of an amount of power consumed by each managed node, performance data indicative of a speed at which the workloads are executed by each managed node, temperature data indicative of a temperature within each managed node, processor utilization data indicative of an amount of processor usage consumed by each workload performed by each managed node, memory utilization data indicative of an amount or frequency of memory use by each workload performed by each managed node, or network utilization data indicative of an amount of network bandwidth used by each workload performed by each managed node.

Example 11 includes the subject matter of any of Examples 1-10, and wherein the plurality of instructions, when executed, further cause the orchestrator server to determine whether the assignment of the workloads is Pareto-efficient; and wherein to determine an adjustment to the assignment of the workloads comprises to determine, in response to a determination that the assignment of the workloads is not Pareto-efficient, an adjustment to the assignment of the workloads.

Example 12 includes the subject matter of any of Examples 1-11, and wherein to determine the adjustments comprises to determine one or more node-specific adjustments indicative of changes to an availability of one or more resources of at least one of the managed nodes to one or more of the workloads performed by the managed node.

Example 13 includes the subject matter of any of Examples 1-12, and wherein to determine the node-specific adjustments comprises to determine at least one of a processor throttle adjustment, a memory usage adjustment, a network bandwidth adjustment, or a fan speed adjustment.

Example 14 includes the subject matter of any of Examples 1-13, and wherein to apply the determined adjustments comprises to issue a request to perform a live migration of a workload between the managed nodes.

Example 15 includes the subject matter of any of Examples 1-14, and wherein to apply the determined adjustments comprises to issue a request to one of the managed nodes to apply one or more node-specific adjustments indicative of changes to an availability of one or more resources of the managed node to one or more of the workloads performed by the managed node.

Example 16 includes a method for allocating resources of a set of managed nodes to workloads with a hierarchical model, the method comprising receiving, by an orchestrator server, resource allocation objective data indicative of multiple resource allocation objectives to be satisfied; determining, by the orchestrator server, an initial assignment of a set of workloads among the managed nodes; receiving, by the orchestrator server, telemetry data from the managed nodes, wherein the telemetry data is indicative of resource utilization by the managed nodes as the workloads are performed; generating, by the orchestrator server and as a function of the telemetry data, a hierarchical model indicative of the resource utilization of each managed node and a relationship between the managed nodes; determining, by the orchestrator server and with the hierarchical model, differences in resource utilization for each workload as a function of the managed node that performed the workload; determining, by the orchestrator server and as a function of the telemetry data and the determined differences, an adjustment to the assignment of the workloads to increase an achievement of at least one of the resource allocation objectives without decreasing the achievement of any of the other resource allocation objectives; and applying, by the orchestrator server, the adjustment to the assignments of the workloads among the managed nodes as the workloads are performed.

Example 17 includes the subject matter of Example 16, and wherein generating a hierarchical model comprises generating a hierarchical model indicative of a spatial relationship between the managed nodes.

Example 18 includes the subject matter of any of Examples 16 and 17, and wherein generating a hierarchical model comprises generating a hierarchical model indicative of a membership of each managed node in a functional group.

Example 19 includes the subject matter of any of Examples 16-18, and further including determining, by the orchestrator server, node-specific weights indicative of the determined differences in resource utilization, wherein determining the adjustment to the workload assignments comprises predicting a change in resource utilization of a workload as a function of the node-specific weights associated with two different managed nodes.

Example 20 includes the subject matter of any of Examples 16-19, and further including generating, by the orchestrator server, data analytics as a function of the telemetry data; and storing, by the orchestrator server, the data analytics in association with each managed node in the hierarchical model.

Example 21 includes the subject matter of any of Examples 16-20, and wherein generating the data analytics comprises generating profiles of the workloads, wherein the profiles are indicative of an identity of each workload and a resource usage classification of each workload.

Example 22 includes the subject matter of any of Examples 16-21, and wherein generating the data analytics comprises predicting future resource utilization of the workloads.

Example 23 includes the subject matter of any of Examples 16-22, and wherein predicting future resource utilization of the workloads comprises identifying potential contention for resources.

Example 24 includes the subject matter of any of Examples 16-23, and wherein receiving resource allocation objective data comprises receiving two or more of power consumption objective data indicative of a target power usage of one or more of the managed nodes, performance objective data indicative of a target speed at which to perform the workloads, reliability objective data indicative of a target life cycle of one or more of the managed nodes, or thermal objective data indicative of a target temperature of one or more of the managed nodes.

Example 25 includes the subject matter of any of Examples 16-24, and wherein receiving telemetry data from the managed nodes comprises receiving at least one of power consumption data indicative of an amount of power consumed by each managed node, performance data indicative of a speed at which the workloads are executed by each managed node, temperature data indicative of a temperature within each managed node, processor utilization data indicative of an amount of processor usage consumed by each workload performed by each managed node, memory utilization data indicative of an amount or frequency of memory use by each workload performed by each managed node, or network utilization data indicative of an amount of network bandwidth used by each workload performed by each managed node.

Example 26 includes the subject matter of any of Examples 16-25, and further including determining, by the orchestrator server, whether the assignment of the workloads is Pareto-efficient, wherein determining an adjustment to the assignment of the workloads comprises determining, in response to a determination that the assignment of the workloads is not Pareto-efficient, an adjustment to the assignment of the workloads.

Example 27 includes the subject matter of any of Examples 16-26, and wherein determining the adjustments comprises determining one or more node-specific adjustments indicative of changes to an availability of one or more resources of at least one of the managed nodes to one or more of the workloads performed by the managed node.

Example 28 includes the subject matter of any of Examples 16-27, and wherein determining the node-specific adjustments comprises determining at least one of a processor throttle adjustment, a memory usage adjustment, a network bandwidth adjustment, or a fan speed adjustment.

Example 29 includes the subject matter of any of Examples 16-28, and wherein applying the determined adjustments comprises issuing a request to perform a live migration of a workload between the managed nodes.

Example 30 includes the subject matter of any of Examples 16-29, and wherein applying the determined adjustments comprises issuing a request to one of the managed nodes to apply one or more node-specific adjustments indicative of changes to an availability of one or more resources of the managed node to one or more of the workloads performed by the managed node.

Example 31 includes one or more machine-readable storage media comprising a plurality of instructions stored thereon that in response to being executed, cause an orchestrator server to perform the method of any of Examples 16-30.

Example 32 includes an orchestrator server to allocate resources of a set of managed nodes to workloads with a hierarchical model, the orchestrator server comprising one or more processors; one or more memory devices having stored therein a plurality of instructions that, when executed by the one or more processors, cause the orchestrator server to perform the method of any of Examples 16-30.

Example 33 includes an orchestrator server to allocate resources of a set of managed nodes to workloads with a hierarchical model, the orchestrator server comprising means for performing the method of any of Examples 16-30.

Example 34 includes an orchestrator server to allocate resources of a set of managed nodes to workloads with a hierarchical model, the orchestrator server comprising resource manager circuitry to receive resource allocation objective data indicative of multiple resource allocation objectives to be satisfied and determine an initial assignment of a set of workloads among the managed nodes; and telemetry monitor circuitry to receive telemetry data from the managed nodes, wherein the telemetry data is indicative of resource utilization by the managed nodes as the workloads are performed; wherein the resource manager circuitry is further to generate, as a function of the telemetry data, a hierarchical model indicative of the resource utilization of each managed node and a relationship between the managed nodes, determine, with the hierarchical model, differences in resource utilization for each workload as a function of the managed node that performed the workload, determine, as a function of the telemetry data and the determined differences, an adjustment to the assignment of the workloads to increase an achievement of at least one of the resource allocation objectives without decreasing the achievement of any of the other resource allocation objectives, and apply the adjustment to the assignments of the workloads among the managed nodes as the workloads are performed.

Example 35 includes the subject matter of Example 34, and wherein to generate a hierarchical model comprises to generate a hierarchical model indicative of a spatial relationship between the managed nodes.

Example 36 includes the subject matter of any of Examples 34 and 35, and wherein to generate a hierarchical model comprises to generate a hierarchical model indicative of a membership of each managed node in a functional group.

Example 37 includes the subject matter of any of Examples 34-36, and wherein the resource manager circuitry is further to determine node-specific weights indicative of the determined differences in resource utilization; and wherein to determine the adjustment to the workload assignments comprises to predict a change in resource utilization of a workload as a function of the node-specific weights associated with two different managed nodes.

Example 38 includes the subject matter of any of Examples 34-37, and wherein the resource manager circuitry is further to generate data analytics as a function of the telemetry data; and store the data analytics in association with each managed node in the hierarchical model.

Example 39 includes the subject matter of any of Examples 34-38, and wherein to generate the data analytics comprises to generate profiles of the workloads, wherein the profiles are indicative of an identity of each workload and a resource usage classification of each workload.

Example 40 includes the subject matter of any of Examples 34-39, and wherein to generate the data analytics comprises to predict future resource utilization of the workloads.

Example 41 includes the subject matter of any of Examples 34-40, and wherein to predict future resource utilization of the workloads comprises to identify potential contention for resources.

Example 42 includes the subject matter of any of Examples 34-41, and wherein to receive resource allocation objective data comprises to receive two or more of power consumption objective data indicative of a target power usage of one or more of the managed nodes, performance objective data indicative of a target speed at which to perform the workloads, reliability objective data indicative of a target life cycle of one or more of the managed nodes, or thermal objective data indicative of a target temperature of one or more of the managed nodes.

Example 43 includes the subject matter of any of Examples 34-42, and wherein to receive telemetry data from the managed nodes comprises to receive at least one of power consumption data indicative of an amount of power consumed by each managed node, performance data indicative of a speed at which the workloads are executed by each managed node, temperature data indicative of a temperature within each managed node, processor utilization data indicative of an amount of processor usage consumed by each workload performed by each managed node, memory utilization data indicative of an amount or frequency of memory use by each workload performed by each managed node, or network utilization data indicative of an amount of network bandwidth used by each workload performed by each managed node.

Example 44 includes the subject matter of any of Examples 34-43, and wherein the resource manager circuitry is further to determine whether the assignment of the workloads is Pareto-efficient; and wherein to determine an adjustment to the assignment of the workloads comprises to determine, in response to a determination that the assignment of the workloads is not Pareto-efficient, an adjustment to the assignment of the workloads.

Example 45 includes the subject matter of any of Examples 34-44, and wherein to determine the adjustments comprises to determine one or more node-specific adjustments indicative of changes to an availability of one or more resources of at least one of the managed nodes to one or more of the workloads performed by the managed node.

Example 46 includes the subject matter of any of Examples 34-45, and wherein to determine the node-specific adjustments comprises to determine at least one of a processor throttle adjustment, a memory usage adjustment, a network bandwidth adjustment, or a fan speed adjustment.

Example 47 includes the subject matter of any of Examples 34-46, and wherein to apply the determined adjustments comprises to issue a request to perform a live migration of a workload between the managed nodes.

Example 48 includes the subject matter of any of Examples 34-47, and wherein to apply the determined adjustments comprises to issue a request to one of the managed nodes to apply one or more node-specific adjustments indicative of changes to an availability of one or more resources of the managed node to one or more of the workloads performed by the managed node.

Example 49 includes an orchestrator server to allocate resources of a set of managed nodes to workloads with a hierarchical model, the orchestrator server comprising circuitry for receiving resource allocation objective data indicative of multiple resource allocation objectives to be satisfied; circuitry for determining an initial assignment of a set of workloads among the managed nodes; circuitry for receiving telemetry data from the managed nodes, wherein the telemetry data is indicative of resource utilization by the managed nodes as the workloads are performed; means for generating, as a function of the telemetry data, a hierarchical model indicative of the resource utilization of each managed node and a relationship between the managed nodes; means for determining, with the hierarchical model, differences in resource utilization for each workload as a function of the managed node that performed the workload; means for determining, as a function of the telemetry data and the determined differences, an adjustment to the assignment of the workloads to increase an achievement of at least one of the resource allocation objectives without decreasing the achievement of any of the other resource allocation objectives; and means for applying the adjustment to the assignments of the workloads among the managed nodes as the workloads are performed.

Example 50 includes the subject matter of Example 49, and wherein the means for generating a hierarchical model comprises means for generating a hierarchical model indicative of a spatial relationship between the managed nodes.

Example 51 includes the subject matter of any of Examples 49 and 50, and wherein the means for generating a hierarchical model comprises means for generating a hierarchical model indicative of a membership of each managed node in a functional group.

Example 52 includes the subject matter of any of Examples 49-51, and further including means for determining node-specific weights indicative of the determined differences in resource utilization, wherein the means for determining the adjustment to the workload assignments comprises means for predicting a change in resource utilization of a workload as a function of the node-specific weights associated with two different managed nodes.

Example 53 includes the subject matter of any of Examples 49-52, and further including means for generating data analytics as a function of the telemetry data; and means for storing the data analytics in association with each managed node in the hierarchical model.

Example 54 includes the subject matter of any of Examples 49-53, and wherein the means for generating the data analytics comprises means for generating profiles of the workloads, wherein the profiles are indicative of an identity of each workload and a resource usage classification of each workload.

Example 55 includes the subject matter of any of Examples 49-54, and wherein the means for generating the data analytics comprises means for predicting future resource utilization of the workloads.

Example 56 includes the subject matter of any of Examples 49-55, and wherein the means for predicting future resource utilization of the workloads comprises means for identifying potential contention for resources.

Example 57 includes the subject matter of any of Examples 49-56, and wherein the circuitry for receiving resource allocation objective data comprises circuitry for receiving two or more of power consumption objective data indicative of a target power usage of one or more of the managed nodes, performance objective data indicative of a target speed at which to perform the workloads, reliability objective data indicative of a target life cycle of one or more of the managed nodes, or thermal objective data indicative of a target temperature of one or more of the managed nodes.

Example 58 includes the subject matter of any of Examples 49-57, and wherein the circuitry for receiving telemetry data from the managed nodes comprises circuitry for receiving at least one of power consumption data indicative of an amount of power consumed by each managed node, performance data indicative of a speed at which the workloads are executed by each managed node, temperature data indicative of a temperature within each managed node, processor utilization data indicative of an amount of processor usage consumed by each workload performed by each managed node, memory utilization data indicative of an amount or frequency of memory use by each workload performed by each managed node, or network utilization data indicative of an amount of network bandwidth used by each workload performed by each managed node.

Example 59 includes the subject matter of any of Examples 49-58, and further including means for determining whether the assignment of the workloads is Pareto-efficient, wherein the means for determining an adjustment to the assignment of the workloads comprises means for determining, in response to a determination that the assignment of the workloads is not Pareto-efficient, an adjustment to the assignment of the workloads.

Example 60 includes the subject matter of any of Examples 49-59, and wherein the means for determining the adjustments comprises means for determining one or more node-specific adjustments indicative of changes to an availability of one or more resources of at least one of the managed nodes to one or more of the workloads performed by the managed node.

Example 61 includes the subject matter of any of Examples 49-60, and wherein the means for determining the node-specific adjustments comprises means for determining at least one of a processor throttle adjustment, a memory usage adjustment, a network bandwidth adjustment, or a fan speed adjustment.

Example 62 includes the subject matter of any of Examples 49-61, and wherein the means for applying the determined adjustments comprises means for issuing a request to perform a live migration of a workload between the managed nodes.

Example 63 includes the subject matter of any of Examples 49-62, and wherein the means for applying the determined adjustments comprises means for issuing a request to one of the managed nodes to apply one or more node-specific adjustments indicative of changes to an availability of one or more resources of the managed node to one or more of the workloads performed by the managed node.

Claims

1. An orchestrator server to allocate resources of a set of managed nodes to workloads with a hierarchical model, the orchestrator server comprising:

one or more processors;

one or more memory devices having stored therein a plurality of instructions that, when executed by the one or more processors, cause the orchestrator server to: receive resource allocation objective data indicative of multiple resource allocation objectives to be satisfied; determine an initial assignment of a set of workloads among the managed nodes; receive telemetry data from the managed nodes, wherein the telemetry data is indicative of resource utilization by the managed nodes as the workloads are performed; generate, as a function of the telemetry data, a hierarchical model indicative of the resource utilization of each managed node and a relationship between the managed nodes; determine, with the hierarchical model, differences in resource utilization for each workload as a function of the managed node that performed the workload; determine, as a function of the telemetry data and the determined differences, an adjustment to the assignment of the workloads to increase an achievement of at least one of the resource allocation objectives without decreasing the achievement of any of the other resource allocation objectives; and apply the adjustment to the assignments of the workloads among the managed nodes as the workloads are performed.

2. The orchestrator server of claim 1, wherein to generate a hierarchical model comprises to generate a hierarchical model indicative of a spatial relationship between the managed nodes.

3. The orchestrator server of claim 1, wherein to generate a hierarchical model comprises to generate a hierarchical model indicative of a membership of each managed node in a functional group.

4. The orchestrator server of claim 1, wherein the plurality of instructions, when executed, further cause the orchestrator server to:

determine node-specific weights indicative of the determined differences in resource utilization; and

wherein to determine the adjustment to the workload assignments comprises to predict a change in resource utilization of a workload as a function of the node-specific weights associated with two different managed nodes.

5. The orchestrator server of claim 1, wherein the plurality of instructions, when executed, further cause the orchestrator server to:

generate data analytics as a function of the telemetry data; and

store the data analytics in association with each managed node in the hierarchical model.

6. The orchestrator server of claim 5, wherein to generate the data analytics comprises to generate profiles of the workloads, wherein the profiles are indicative of an identity of each workload and a resource usage classification of each workload.

7. The orchestrator server of claim 5, wherein to generate the data analytics comprises to predict future resource utilization of the workloads.

8. The orchestrator server of claim 7, wherein to predict future resource utilization of the workloads comprises to identify potential contention for resources.

9. The orchestrator server of claim 1, wherein to receive resource allocation objective data comprises to receive two or more of power consumption objective data indicative of a target power usage of one or more of the managed nodes, performance objective data indicative of a target speed at which to perform the workloads, reliability objective data indicative of a target life cycle of one or more of the managed nodes, or thermal objective data indicative of a target temperature of one or more of the managed nodes.

10. The orchestrator server of claim 1, wherein to receive telemetry data from the managed nodes comprises to receive at least one of power consumption data indicative of an amount of power consumed by each managed node, performance data indicative of a speed at which the workloads are executed by each managed node, temperature data indicative of a temperature within each managed node, processor utilization data indicative of an amount of processor usage consumed by each workload performed by each managed node, memory utilization data indicative of an amount or frequency of memory use by each workload performed by each managed node, or network utilization data indicative of an amount of network bandwidth used by each workload performed by each managed node.

11. The orchestrator server of claim 1, wherein the plurality of instructions, when executed, further cause the orchestrator server to determine whether the assignment of the workloads is Pareto-efficient; and

wherein to determine an adjustment to the assignment of the workloads comprises to determine, in response to a determination that the assignment of the workloads is not Pareto-efficient, an adjustment to the assignment of the workloads.

12. The orchestrator server of claim 1, wherein to determine the adjustments comprises to determine one or more node-specific adjustments indicative of changes to an availability of one or more resources of at least one of the managed nodes to one or more of the workloads performed by the managed node.

13. One or more machine-readable storage media comprising a plurality of instructions stored thereon that, in response to being executed, cause an orchestrator server to:

receive resource allocation objective data indicative of multiple resource allocation objectives to be satisfied;

determine an initial assignment of a set of workloads among the managed nodes;

receive telemetry data from the managed nodes, wherein the telemetry data is indicative of resource utilization by the managed nodes as the workloads are performed;

generate, as a function of the telemetry data, a hierarchical model indicative of the resource utilization of each managed node and a relationship between the managed nodes;

determine, with the hierarchical model, differences in resource utilization for each workload as a function of the managed node that performed the workload;

determine, as a function of the telemetry data and the determined differences, an adjustment to the assignment of the workloads to increase an achievement of at least one of the resource allocation objectives without decreasing the achievement of any of the other resource allocation objectives; and

apply the adjustment to the assignments of the workloads among the managed nodes as the workloads are performed.

14. The one or more machine-readable storage media of claim 13, wherein to generate a hierarchical model comprises to generate a hierarchical model indicative of a spatial relationship between the managed nodes.

15. The one or more machine-readable storage media of claim 13, wherein to generate a hierarchical model comprises to generate a hierarchical model indicative of a membership of each managed node in a functional group.

16. The one or more machine-readable storage media of claim 13, wherein the plurality of instructions, when executed, further cause the orchestrator server to:

determine node-specific weights indicative of the determined differences in resource utilization; and

wherein to determine the adjustment to the workload assignments comprises to predict a change in resource utilization of a workload as a function of the node-specific weights associated with two different managed nodes.

17. The one or more machine-readable storage media of claim 13, wherein the plurality of instructions, when executed, further cause the orchestrator server to:

generate data analytics as a function of the telemetry data; and

store the data analytics in association with each managed node in the hierarchical model.

18. The one or more machine-readable storage media of claim 17, wherein to generate the data analytics comprises to generate profiles of the workloads, wherein the profiles are indicative of an identity of each workload and a resource usage classification of each workload.

19. The one or more machine-readable storage media of claim 17, wherein to generate the data analytics comprises to predict future resource utilization of the workloads.

20. The one or more machine-readable storage media of claim 19, wherein to predict future resource utilization of the workloads comprises to identify potential contention for resources.

21. The one or more machine-readable storage media of claim 13, wherein to receive resource allocation objective data comprises to receive two or more of power consumption objective data indicative of a target power usage of one or more of the managed nodes, performance objective data indicative of a target speed at which to perform the workloads, reliability objective data indicative of a target life cycle of one or more of the managed nodes, or thermal objective data indicative of a target temperature of one or more of the managed nodes.

22. The one or more machine-readable storage media of claim 13, wherein to receive telemetry data from the managed nodes comprises to receive at least one of power consumption data indicative of an amount of power consumed by each managed node, performance data indicative of a speed at which the workloads are executed by each managed node, temperature data indicative of a temperature within each managed node, processor utilization data indicative of an amount of processor usage consumed by each workload performed by each managed node, memory utilization data indicative of an amount or frequency of memory use by each workload performed by each managed node, or network utilization data indicative of an amount of network bandwidth used by each workload performed by each managed node.

23. The one or more machine-readable storage media of claim 13, wherein the plurality of instructions, when executed, further cause the orchestrator server to determine whether the assignment of the workloads is Pareto-efficient; and

wherein to determine an adjustment to the assignment of the workloads comprises to determine, in response to a determination that the assignment of the workloads is not Pareto-efficient, an adjustment to the assignment of the workloads.

24. The one or more machine-readable storage media of claim 13, wherein to determine the adjustments comprises to determine one or more node-specific adjustments indicative of changes to an availability of one or more resources of at least one of the managed nodes to one or more of the workloads performed by the managed node.

25. An orchestrator server to allocate resources of a set of managed nodes to workloads based on resource utilization phase residencies, the orchestrator server comprising:

circuitry for receiving resource allocation objective data indicative of multiple resource allocation objectives to be satisfied;

circuitry for determining an initial assignment of a set of workloads among the managed nodes;

circuitry for receiving telemetry data from the managed nodes, wherein the telemetry data is indicative of resource utilization by each of the managed nodes as the workloads are performed;

means for determining, as a function of the telemetry data, phase residency data indicative of temporal lengths of resource utilization phases of the workloads, wherein each resource utilization phase is indicative of a utilization of a managed node component that satisfies a threshold amount;

means for determining, as a function of at least the phase residency data and the resource allocation objective data, an adjustment to the assignment of the workloads to increase an achievement of at least one of the resource allocation objectives without decreasing the achievement of any of the other resource allocation objectives; and

means for applying the adjustment to the assignments of the workloads among the managed nodes as the workloads are performed.

26. A method for allocating resources of a set of managed nodes to workloads based on resource utilization phase residencies, the method comprising:

receiving, by an orchestrator server, resource allocation objective data indicative of multiple resource allocation objectives to be satisfied;

determining, by the orchestrator server, an initial assignment of a set of workloads among the managed nodes;

receiving, by the orchestrator server, telemetry data from the managed nodes, wherein the telemetry data is indicative of resource utilization by each of the managed nodes as the workloads are performed;

determining, by the orchestrator server and as a function of the telemetry data, phase residency data indicative of temporal lengths of resource utilization phases of the workloads, wherein each resource utilization phase is indicative of a utilization of a managed node component that satisfies a threshold amount;

determining, by the orchestrator server and as a function of at least the phase residency data and the resource allocation objective data, an adjustment to the assignment of the workloads to increase an achievement of at least one of the resource allocation objectives without decreasing the achievement of any of the other resource allocation objectives; and

applying, by the orchestrator server, the adjustment to the assignments of the workloads among the managed nodes as the workloads are performed.

27. The method of claim 26, wherein determining the adjustment to the assignment of the workloads comprises:

identifying complementary workload utilization phases indicative of resource utilization phases of different managed node components by two or more workloads; and

determining an alignment of the complementary workload utilization phases to cause the complementary workload utilization phases to be performed concurrently by the same managed node.

28. The method of claim 27, wherein:

determining the alignment of complementary workload phases comprises determining a time offset for execution of one or more of the workloads; and

applying the adjustments to the assignments comprises:

temporarily suspending execution of the one or more of the workloads; and

resuming execution of the one or more of the workloads after the time offset has elapsed.