Disaggregated Fabric-Switched Computing User Interfaces
Disaggregated computing architectures, platforms, and systems are provided herein. In one example, a method of operating at graphical user interface system to a disaggregated computing platform is provided. The method includes receiving user input for a cluster interface portion to establish compute clusters comprising compute units for executing user software thereon, and receiving further user input to add physical resources into one or more compute units that are coupled over a Peripheral Component Interconnect Express (PCIe) fabric and associate the one or more compute units to associated ones of the compute clusters, the physical resources comprising processing resources, storage resources, and network interface resources. The method also includes presenting indications of the formation of the physical resources into the compute units and the compute units into the compute clusters, and presenting telemetry data related to operation of at least the compute units.
1 . A graphical user interface system to a disaggregated computing platform, comprising:
a cluster interface portion configured to accept user input to create compute units for executing user software thereon;
a computing block interface portion configured to receive further user input to add physical resources into one or more compute units that are coupled over a Peripheral Component Interconnect Express (PCIe) fabric and associate the one or more compute units to selected computing clusters, the physical resources comprising processing resources, storage resources, and network interface resources; and
a monitoring interface portion configured to present telemetry data related to operation of at least the compute units.
2 . The graphical user interface system of claim 1 , comprising:
the computing block interface portion configured to present graphical representations of at least the processing elements, the storage elements, and the network interface elements for incorporation by a user into the compute units responsive to associated user input.
3 . The graphical user interface system of claim 2 , comprising:
the computing block interface portion configured to present further graphical representations of at least graphics processing units (GPU) resources for incorporation into the compute units responsive to the user input.
4 . The graphical user interface system of claim 2 , comprising:
the cluster interface portion configured to receive, via the user input, user instructions to incorporate one or more compute units into a cluster for executing the user software thereon; and
responsive to receiving the user instructions, a management interface portion configured to indicate the user instructions to a management processor for implementation of the user instructions to alter logical partitioning among the PCIe fabric.
5 . The graphical user interface system of claim 1 , comprising:
the computing block interface portion configured to present options to the user to alter the compute units after formation of the compute units to selectively add additional physical resources into the compute units or remove existing physical resources from the compute units; and
responsive to receiving user instructions to alter the compute units, a management interface portion configured to indicate the user instructions to a management processor for implementation of the user instructions to alter logical partitioning among the PCIe fabric.
6 . The graphical user interface system of claim 1 , comprising:
the computing block interface portion configured to receive user defined targets describing usage requirements for one or more compute units;
the computing block interface portion configured to responsively process the user defined targets to determine ones of the physical resources to include in at least one associated compute unit to meet the usage requirements; and
a management interface portion configured to indicate the ones of the physical resources to a management processor for implementation of the one or more compute units using logical partitioning among the PCIe fabric.
7 . The graphical user interface system of claim 1 , comprising:
the computing block interface portion configured to receive user thresholds related to operation of the compute units, wherein the user thresholds indicate activity levels of the compute units for which to add one or more physical resources into ones of the compute units that exceed the user thresholds.
8 . The graphical user interface system of claim 1 , comprising:
the cluster interface portion configured to deploy the user software to at least one of the compute units responsive to creation of the at least one of the compute units.
9 . The graphical user interface system of claim 1 , wherein the telemetry data indicates at least usage levels among the physical resources assigned to the compute units, and indicates a quantity of free physical resources ready for assignment to one or more compute units.
10 . A method of operating at graphical user interface system to a disaggregated computing platform, the method comprising:
receiving user input for a cluster interface portion to establish compute clusters comprising compute units for executing user software thereon;
receiving further user input to add physical resources into one or more compute units that are coupled over a Peripheral Component Interconnect Express (PCIe) fabric and associate the one or more compute units to associated ones of the compute clusters, the physical resources comprising processing resources, storage resources, and network interface resources;
responsive to formation of the compute units and the compute clusters among the PCIe fabric, presenting indications of the formation of the physical resources into the compute units and the compute units into the compute clusters; and
presenting telemetry data related to operation of at least the compute units.
11 . The method of claim 10 , further comprising:
presenting graphical representations of at least the processing elements, the storage elements, and the network interface elements for incorporation by the user into the compute units responsive to associated user input.
12 . The method of claim 11 , further comprising:
presenting further graphical representations of at least graphics processing units (GPU) resources for incorporation into the compute units.
13 . The method of claim 11 , further comprising:
receiving user instructions indicating at least one compute unit for incorporation a target compute cluster for executing the user software thereon; and
responsive to receiving the user instructions, indicating the user instructions to a management processor for implementation of the user instructions to alter logical partitioning among the PCIe fabric in accordance with the user instructions.
14 . The method of claim 10 , further comprising:
presenting options to the user to alter the compute units after formation of the compute units to selectively add additional physical resources into the compute units or remove existing physical resources from the compute units; and
responsive to receiving user instructions to alter the compute units, indicating the user instructions to a management processor for implementation of the user instructions to alter logical partitioning among the PCIe fabric in accordance with the user instructions.
15 . The method of claim 10 , further comprising:
receiving user defined targets describing usage requirements for one or more compute units;
processing the user defined targets to determine ones of the physical resources to include in at least one associated compute unit to meet the usage requirements; and
indicating the ones of the physical resources to a management processor for implementation of the one or more compute units using logical partitioning among the PCIe fabric.
16 . The method of claim 10 , further comprising:
receiving user thresholds related to operation of the compute units, wherein the user thresholds indicate activity levels of the compute units for which to add one or more physical resources into ones of the compute units that exceed the user thresholds.
17 . The method of claim 10 , further comprising:
deploying the user software to at least one of the compute units responsive to creation of the at least one of the compute units.
18 . The method of claim 10 , wherein the telemetry data indicates at least usage levels among the physical resources assigned to the compute units, and indicates a quantity of free physical resources ready for assignment to one or more compute units.
19 . A computing apparatus comprising:
one or more computer readable storage media;
a processing system operatively coupled with the one or more computer readable storage media; and
program instructions stored on the one or more computer readable storage media, that when executed by the processing system, direct the processing system to at least:
receive user input for a cluster interface portion to establish compute clusters comprising compute units for executing user software thereon;
receive further user input to add physical resources into one or more compute units that are coupled over a Peripheral Component Interconnect Express (PCIe) fabric and associate the one or more compute units to associated ones of the compute clusters, the physical resources comprising processing resources, storage resources, and network interface resources;
responsive to formation of the compute units and the compute clusters among the PCIe fabric, present indications of the formation of the physical resources into the compute units and the compute units into the compute clusters; and
present telemetry data related to operation of at least the compute units.
20 . The computing apparatus of claim 19 , comprising further program instructions, when executed by the processing system, direct the processing system to at least:
present options to the user to alter the compute units after formation of the compute units to selectively add additional physical resources into the compute units or remove existing physical resources from the compute units; and
responsive to receiving user instructions to alter the compute units, indicate the user instructions to a management processor for implementation of the user instructions to alter logical partitioning among the PCIe fabric in accordance with the user instructions.