Formation of compute units from converged and disaggregated component pools
Deployment of arrangements of computing components coupled over communication fabrics are presented herein. In one example, a method includes forming a compute unit to include at least a selected disaggregated computing component and at least a selected converged computing element. The selected disaggregated computing component is added to the compute unit from a pool of disaggregated computing components having a granularity comprising the disaggregated computing components arbitrarily and separately coupled to a communication fabric. The selected converged computing element is added to the compute unit from a pool of converged computing elements having a granularity comprising sets of computing components with fixed relationships to associated processors.
1. A method comprising:
in a user interface, receiving user instructions to form a compute unit to include at least a selected disaggregated computing component and at least a selected converged computing element;
wherein the selected disaggregated computing component is added to the compute unit with a granularity of an individual computing component from a pool of disaggregated computing components arbitrarily and individually segregable into logical groupings within a communication fabric; and
wherein the selected converged computing element is added to the compute unit with a granularity having a fixed set of computing components from a pool of converged computing elements comprising sets of computing components having fixed relationships to associated processors.
2. The method of claim 1 , wherein the selected converged computing element is communicatively coupled to an additional communication fabric different from the communication fabric.
3. The method of claim 2 , wherein the communication fabric comprises a first communication fabric type corresponding to communication protocols from at least one among Peripheral Component Interconnect Express (PCIe), Gen-Z, and Compute Express Link (CXL), and wherein the additional communication fabric comprises a second communication fabric type corresponding to at least one among Ethernet, and InfiniBand, and FibreChannel.
4. The method of claim 2 , wherein forming the compute unit comprises establishing one or more logical domains within at least one among the communication fabric and the additional communication fabric that make selected computing components for the compute unit accessible to each other.
5. The method of claim 1 , wherein adding the selected disaggregated computing component into the compute unit comprises individually associating the selected disaggregated computing component into a logical arrangement comprising the compute unit; and
wherein adding the selected converged computing element into the compute unit comprises associating into the logical arrangement a set of computing components having a fixed relationship to an associated processor.
6. The method of claim 1 , wherein the converged computing elements correspond to computing assemblies that each include at least a processing component, a data storage component, and a network interface component that communicatively couples each computing assembly using a network communication protocol; and
wherein the disaggregated computing components correspond to at least individual ones among processing components, data storage components, and network interface components arbitrarily and individually coupled to the communication fabric using a point-to-point communication protocol.
7. The method of claim 1 , comprising:
in the user interface, receiving additional user instructions to deploy software for execution by at least a corresponding processor component of the compute unit.
8. The method of claim 1 , comprising:
in the user interface, receiving additional user instructions to assign network addressing for a network interface component of the selected converged computing element, wherein traffic for the selected converged computing element is directed through the network interface component.
9. The method of claim 1 , further comprising:
performing a discovery process over the communication fabric associated with the selected disaggregated computing component and an additional communication fabric associated with the converged disaggregated computing element to identify first computing components as disaggregated computing components and identify second computing components as converged computing components; and
for the converged computing components, determining ones among the converged computing components that share a chassis as the sets of computing components;
maintaining one or more data structures which indicates a communication fabric type for each of the first computing components and each of the second computing components and indicates the sets of computing components that share corresponding chassis.
10. The method of claim 1 , wherein the pool of disaggregated computing components and the pool of converged computing elements each include one or more among graphics processing units (GPUs), central processing units (CPUs), network interface cards (NICs), field-programmable gate arrays (FPGAs), data storage devices, and memory devices.
11. An apparatus, comprising:
one or more computer readable storage media; and
program instructions stored on the one or more computer readable storage media, the program instructions executable by a processing system to direct the processing system to at least:
a user interface configured to receive user instructions to form a compute unit to include at least a selected disaggregated computing component and at least a selected converged computing element;
wherein the selected disaggregated computing component is added to the compute unit with a granularity of an individual computing component from a pool of disaggregated computing components arbitrarily and individually segregable into logical groupings within a communication fabric; and
wherein the selected converged computing element is added to the compute unit with a granularity having a fixed set of computing components from a pool of converged computing elements comprising sets of computing components having fixed relationships to associated processors.
12. The apparatus of claim 11 , wherein the selected converged computing element is communicatively coupled to an additional communication fabric different from the communication fabric.
13. The apparatus of claim 12 , wherein the communication fabric comprises a first communication fabric type corresponding to communication protocols from at least one among Peripheral Component Interconnect Express (PCIe), Gen-Z, and Compute Express Link (CXL), and wherein the additional communication fabric comprises a second communication fabric type corresponding to at least one among Ethernet, and InfiniBand, and FibreChannel.
14. The apparatus of claim 12 , wherein forming the compute unit comprises establishing one or more logical domains within at least one among the communication fabric and the additional communication fabric that make selected computing components for the compute unit accessible to each other.
15. The apparatus of claim 11 , comprising program instructions stored on the one or more computer readable storage media, the program instructions executable by the processing system to direct the processing system to at least:
the user interface configured to receive the user instructions to add the selected disaggregated computing component into the compute unit by at least individually associating the selected disaggregated computing component into a logical arrangement comprising the compute unit; and
the user interface configured to receive the user instructions to add the selected converged computing element into the compute unit by at least associating into the logical arrangement a set of computing components having a fixed relationship to an associated processor.
16. The apparatus of claim 11 , wherein the converged computing elements correspond to computing assemblies that each include at least a processing component, a data storage component, and a network interface component that communicatively couples each computing assembly using a network communication protocol; and
wherein the disaggregated computing components correspond to at least individual ones among processing components, data storage components, and network interface components arbitrarily and individually coupled to the communication fabric using a point-to-point communication protocol.
17. The apparatus of claim 11 , comprising program instructions stored on the one or more computer readable storage media, the program instructions executable by the processing system to direct the processing system to at least:
the user interface configured to receive additional user instructions to deploy software for execution by at least a corresponding processor component of the compute unit.
18. The apparatus of claim 11 , comprising program instructions stored on the one or more computer readable storage media, the program instructions executable by the processing system to direct the processing system to at least:
the user interface configured to receive additional user instructions to assign network addressing for a network interface component of the selected converged computing element, wherein traffic for the selected converged computing element is directed through the network interface component.
19. The apparatus of claim 11 , comprising program instructions stored on the one or more computer readable storage media, the program instructions executable by the processing system to direct the processing system to at least:
perform a discovery process over the communication fabric associated with the selected disaggregated computing component and an additional communication fabric associated with the converged disaggregated computing element to identify first computing components as disaggregated computing components and identify second computing components as converged computing components; and
for the converged computing components, determine ones among the converged computing components that share a chassis as the sets of computing components;
maintain one or more data structures which indicates a communication fabric type for each of the first computing components and each of the second computing components and indicates the sets of computing components that share corresponding chassis.
20. The apparatus of claim 11 , wherein the pool of disaggregated computing components and the pool of converged computing elements each include one or more among graphics processing units (GPUs), central processing units (CPUs), network interface cards (NICs), field-programmable gate arrays (FPGAs), data storage devices, and memory devices.