Abstract: Technologies for managing cache quality of service (QoS) include a compute node that includes a network interface controller (NIC) configured to identify a total amount of available shared cache ways of a last level cache (LLC) of the compute node and identify a destination address for each of a plurality of virtual machines (VMs) managed by the compute node. The NIC is further configured to calculate a recommended amount of cache ways for each workload type associated with VMs based on network traffic to be received by the NIC and processed by each of the VMs, wherein the recommended amount of cache ways includes a recommended amount of hardware I/O LLC cache ways and a recommended amount of isolated LLC cache ways usable to update a cache QoS register that includes the recommended amount of cache ways for each workload type. Other embodiments are described herein.

Abstract: The disclosure provides an approach for dynamic centralized model compilation. Embodiments include receiving, from a client, a request for a machine learning model, wherein the request indicates either one or more attributes comprising one or more of a hardware characteristic, a target precision, or a compiler characteristic, or that one or more default behaviors should be used to compile the machine learning model. Embodiments include determining a compiler for the machine learning model based on the one or more attributes or the one or more default behaviors, wherein the compiler is stored in a registry. Embodiments include compiling the machine learning model using the compiler. Embodiments include providing the compiled machine learning model to the client in response to the request.

Abstract: Various systems and methods are described for implementing cloud-to-edge (C2E) security are disclosed, including systems and methods for the execution of various workloads that are distributed among multiple edge computing nodes. An example technique for managing distributed workloads includes: identifying characteristics of a distributed workload from an execution of the distributed workload, for a distributed workload that is partitioned among multiple computing nodes; evaluating a trust status of the distributed workload in response to a change in the execution of the distributed workload, including verifying resources to execute the distributed workload and verifying security policies associated with the resources; and controlling the execution of the distributed workload among the multiple computing nodes, based on the characteristics and the evaluated trust status.

Abstract: Systems, apparatus, articles of manufacture, and methods are disclosed to improve container security. An example apparatus includes interface circuitry to access network resources, instructions, and programmable circuitry to at least one of instantiate or execute the instructions to parse an inventory data structure to identify (a) first data structures corresponding to a container and (b) second data structures corresponding to an attestation circuit interface, and verify a combination of validation elements corresponding to the first and second data structures.

Abstract: Operations are described for executing application images in a trusted execution environment (TEE). These operations can include retrieving a user application image and generating a bundle for the application image by mounting an overlay onto the application image. The overlay can include library functionality for operating in the TEE. The operations can further include providing the bundle for execution in the TEE.

Abstract: A system and computer-implemented method for container provenance tracking uses a build instruction file of a container image to output a new provenance document associated with the container image for distribution. For each file system layer of the container image specified in the build instruction file, an existing provenance document for the file system layer is inserted into the new provenance document. If there is no existing provenance document, information about each software component included in the file system layer is retrieved and inserted into the new provenance document.

Abstract: A system and computer-implemented method for container provenance tracking uses a build instruction file of a container image to output a new provenance document associated with the container image for distribution. For each file system layer of the container image specified in the build instruction file, an existing provenance document for the file system layer is inserted into the new provenance document. If there is no existing provenance document, information about each software component included in the file system layer is retrieved and inserted into the new provenance document.

Abstract: Technologies for managing cache quality of service (QoS) include a compute node that includes a network interface controller (NIC) configured to identify a total amount of available shared cache ways of a last level cache (LLC) of the compute node and identify a destination address for each of a plurality of virtual machines (VMs) managed by the compute node. The NIC is further configured to calculate a recommended amount of cache ways for each workload type associated with VMs based on network traffic to be received by the NIC and processed by each of the VMs, wherein the recommended amount of cache ways includes a recommended amount of hardware I/O LLC cache ways and a recommended amount of isolated LLC cache ways usable to update a cache QoS register that includes the recommended amount of cache ways for each workload type. Other embodiments are described herein.

Abstract: Disclosed is a file system that may support data management for a distributed data storage and computing system, such as Apache™ Hadoop®. The file system may include an expandable tree-based indexing framework that enables convenient expansion of the file system. As a non-limiting example, the file system disclosed herein may enable indexing, storage, and management of a billion or more files, which is 1,000 times the capacity of currently available file systems. The file system includes a root index system and a number of leaf index systems that are organized in a tree data structure. The leaf index systems provide heartbeat information to the root index system to enable the root index system to maintain a lightweight and searchable index of file references and leaf index references. Each of the leaf indexes maintains an index or mapping of file references to file block addresses within data storage devices that store files.

Abstract: Two techniques are commonly used when developing database applications. First, script files containing batched database commands are frequently used to establish the schema of database tables. Second, the software design pattern “Data Access Objects” are sometimes used to contain programmatic database requests while providing an application programmer a more abstract, easier to use interface to the database. Both of these techniques require essentially the same information: an understanding of the organization of particular database tables. The present invention reduces the labor associated with maintaining synchronicity between these two components by a method that allows both the schema and the format of the Data Access Object to be determined by evaluating the properties of an object to be stored in the database.

Abstract: Creating directed circuits including encoding of policy and state information associated with directed circuits that may be established between two network entities on separate protected networks. The policies are maintained and implemented according to security methods that initially belong to network apparatus on either of the two protected networks; however the security methods may be relinquished to trusted third parties.