SYSTEM AND METHOD FOR FLEXIBLE STORAGE AND NETWORKING PROVISIONING IN LARGE SCALABLE PROCESSOR INSTALLATIONS
A system and method for provisioning within a system design to allow the storage and IO resources to scale with compute resources are provided.
Latest Calxeda, Inc. Patents:
- PERFORMANCE AND POWER OPTIMIZED COMPUTER SYSTEM ARCHITECTURES AND METHODS LEVERAGING POWER OPTIMIZED TREE FABRIC INTERCONNECT
- SYSTEM AND METHOD FOR USING A MULTI-PROTOCOL FABRIC MODULE ACROSS A DISTRIBUTED SERVER INTERCONNECT FABRIC
- TIME SYNCHRONIZATION BETWEEN NODES OF A SWITCHED INTERCONNECT FABRIC
- System And Method For Continuous Low-Overhead Monitoring Of Distributed Applications Running On A Cluster Of Data Processing Nodes
- System, Method and Computer Readable Medium for Offloaded Computation of Distributed Application Protocols Within a Cluster of Data Processing Nodes
The disclosure relates generally to provisioning within a system design to allow the storage and networking resources to scale with compute resources.
BACKGROUNDServer systems generally provide a fixed number of options. For example, there are a fixed number of PCI Express IO slots and a fixed number of hard drive bays, which often are delivered empty as they provide future upgradability. The customer is expected to gauge future needs and select a server chassis category that will serve present and future needs. Historically, and particularly with x86-class servers, predicting the future needs has been achievable because product improvements from one generation to another have been incremental.
With the advent of scalable servers, the ability to predict future needs has become less obvious. For example, in the class of servers within a 2U chassis, it is possible to install 120 compute nodes in an incremental fashion. Using this server as a data storage device, the user may require only 4 compute nodes, but may desire 80 storage drives. Using the same server as a pure compute function focused on analytics, the user may require 120 compute nodes and no storage drives. The nature of scalable servers lends itself to much more diverse applications which require diverse system configurations. As the diversity increases over time, the ability to predict the system features that must scale becomes increasingly difficult.
An example of a typical server system is shown in
The disclosure is particular applicable to a 2U chassis which is the most widely favored form factor for PC-class servers. The concepts herein apply to any chassis form factor, such as tower and rack chassis' of varying customary sizes and any unconventional form. For example,
Computer architecture have various components and those components can be categorized in three categories: compute, storage, and IO wherein the compute category may include computing related or processor components, the storage category are storage type devices and IO are input/output components of the computer architecture. Each category can be further subdivided, and each category can be defined to contain certain element types. For example, compute can be subdivided into an ALU, cache, system memory, and local peripherals. Also for example, the storage category can contain element types of hard drives, solid state storage devices, various industry-standard form factors, or non-standard devices. For this disclosure, the component level (compute, storage, IO) are used with the understanding that each component has dimensions and attributes to which the same concepts may be applied.
The system and method of the disclosure allow the same physical space to be used by any of the computer components: compute devices, storage devices, or IO devices. This provides the greatest flexibility in configuration of systems for different applications. In addition, devices within the computer system that support all three components, such as power supplies and fans, will be assumed to be stationary for simplicity in the examples provided. It is understood that these support devices do not have to be stationary, depending on the goals in differentiation of the system design, meaning that they also can scale as needed.
In this example, a “slot” consists of physical connectors and a defined volume of space above these connectors. In one implementation, two PCI Express x16 connectors are used, along with a volume of 10″ length by 2.7″ height by 1″ width. This volume is selected based on associated component heights, the restrictions of a 2U chassis, and a length driven by the PCB space required to accommodate this implementation. It is understood that other connector types can be used, depending on the signaling frequency and quantity of pins required. It is understood that other volumes can be used, depending on the physical constraints that are acceptable for the application. The connector pin definitions are critical to accommodate the many needs of the computer components, both in power delivery and bandwidth of the electrical interfaces.
An exemplary compute module 30 is shown in
Examples of storage modules 50 that may be used in the system are shown in
In
In
The example in
An exemplary IO module 60 for the system is shown in
An exemplary hybrid module 70 is shown in
With the compute, storage, and IO module concepts described above, exemplary systems of
While the foregoing has been with reference to a particular embodiment of the invention, it will be appreciated by those skilled in the art that changes in this embodiment may be made without departing from the principles and spirit of the disclosure, the scope of which is defined by the appended claims.
Claims
1. A scalable system, comprising:
- a chassis having a predetermined physical form factor, the chassis having a plurality of slots into which modules are placed;
- one or more compute components that are capable of being housed within the chassis in the plurality of slots;
- one or more storage components that are capable of being housed within the chassis in the plurality of slots;
- one or more IO components that are capable of being housed within the chassis in the plurality of slots; and
- wherein the compute components, the storage components and the IO components housed in the chassis are determined based on a desired computing power, storage power and input/output power of the system such that the system in the chassis is scalable.
2. The system of claim 1 further comprising one or more support devices within the chassis that support the other components housed within the chassis.
3. The system of claim 2, wherein the one or more support devices are one of a fan and a power supply.
4. The system of claim 1, wherein the chassis has a set of physical connectors and a volume of space.
5. The system of claim 4, wherein the set of physical connectors is one or more PCIe connectors.
6. The system of claim 5, wherein the set of physical connectors are two PCI Express x16 connectors and the volume of space is 10″ length by 2.7″ height by 1″ width.
7. The system of claim 1, wherein the chassis is one of a 2U chassis and a vertical chassis.
8. The system of claim 1, wherein each compute component further comprises one or more compute nodes,
9. The system of claim 8, wherein each node has a system on chip, a set of system memory that is accessible by the system on chip, a local storage space for the system on chip and connectivity.
10. The system of claim 1, wherein each storage component is one of 2.5″ cased SATA drive, caseless SATA SSD (solid state device) and mSATA (Modular SATA) SSD.
11. The system of claim 1, wherein each IO component has set of connectors and a translation circuit that translates between IO protocols.
12. The system of claim 1 further comprising one or more hybrid components that are capable of being housed within the chassis in the plurality of slots, wherein each hybrid component has one or more of the compute component, the storage component and the IO component.
13. The system of claim 1 further comprising a straddle slot that spans one or more boards and accepts different components.
14. A method for building a scalable system in a fixed area, the method comprising:
- providing a chassis having a predetermined physical form factor, the chassis having a plurality of slots into which modules are placed;
- providing one or more compute components that are capable of being housed within the chassis in the plurality of slots, one or more storage components that are capable of being housed within the chassis in the plurality of slots and one or more IO components that are capable of being housed within the chassis in the plurality of slots; and
- determining, for a system with a desired computing power, storage power and input/output power, one of the one or more compute components, the one or more storage components and the one or more IO components that are housed within the chassis such that the system in the chassis is scalable.
15. The method of claim 14 further comprising providing one or more support devices within the chassis that support the other components housed within the chassis.
16. The method of claim 15, wherein the one or more support devices are one of a fan and a power supply.
17. The method of claim 14, wherein the chassis has a set of physical connectors and a volume of space.
18. The method of claim 17, wherein the set of physical connectors is one or more PCIe connectors.
19. The method of claim 18, wherein the set of physical connectors are two PCI Express x16 connectors and the volume of space is 10″ length by 2.7″ height by 1″ width.
20. The method of claim 14, wherein the chassis is one of a 2U chassis and a vertical chassis.
21. The method of claim 14, wherein each compute component further comprises one or more compute nodes,
22. The method of claim 21, wherein each node has a method on chip, a set of method memory that is accessible by the method on chip, a local storage space for the method on chip and connectivity.
23. The method of claim 14, wherein each storage component is one of 2.5″ cased SATA drive, caseless SATA SSD (solid state device) and mSATA (Modular SATA) SSD.
24. The method of claim 14, wherein each IO component has set of connectors and a translation circuit that translates between IO protocols.
25. The method of claim 14 further comprising providing one or more hybrid components that are capable of being housed within the chassis in the plurality of slots, wherein each hybrid component has one or more of the compute component, the storage component and the IO component.
26. A printed circuit board, comprising:
- one or more PCIe connectors through which power is routed;
- one or more regulators connected to the printed circuit board that are powered by the one or more PCIe connectors and generate a regulated voltage;
- one of a SATA, mSATA and miniSATA connector connected to the printed circuit board that are powered by the regulated voltage; and
- wherein a storage component can be connected to the connector to power the storage component.
27. The printed circuit board of claim 26, wherein the storage component is one of a 2.5″ cased SATA drive, caseless SATA solid state device and an mSATA solid state device.
28. The printed circuit board of claim 26 further comprising one of a SATA connector, mSATA connector and a miniSATA connector connected to the storage component through which a set of SATA signals from the storage component are communicated.
29. The printed circuit board of claim 26 further comprising one of a SATA connector, mSATA connector and a miniSATA connector connected to the storage component through which a set of SATA signals from the storage component are communicated and the set of SATA signals are routed on the printed circuit board to the PCIe connectors.
30. The printed circuit board of claim 28 further comprising a compute component connected to the printed circuit board using a SATA connector and the set of SATA signals are communicated to the compute component.
31. The printed circuit board of claim 26 further comprising one or more digital enables that are routable through the PCIe connectors to allow external control of the one or more regulators.
32. The printed circuit board of claim 26 further comprising one or more of a power good signal and an acknowledge signal are routable through the PCIe connectors from the one or more regulators.
33. The printed circuit board of claim 31 further comprising a compute component connected to the printed circuit board and the compute component controls the digital enables.
34. The printed circuit board of claim 26 further comprising a temperature sensor attached to the printed circuit board and a temperature sensor interface is routed through the PCIe connector.
Type: Application
Filed: Oct 28, 2011
Publication Date: May 2, 2013
Applicant: Calxeda, Inc. (Austin, TX)
Inventors: Arnold Thomas Schnell (Pflugerville, TX), Richard Owen Waldorf (Austin, TX), David Borland (Austin, TX)
Application Number: 13/284,855
International Classification: G06F 1/18 (20060101); H05K 13/00 (20060101); G06F 1/20 (20060101);