Unlocking computing resources for decomposable data centers
Techniques for sending Compute Express Link (CXL) packets over Ethernet (CXL-E) in a composable data center that may include disaggregated, composable servers. The techniques may include receiving, from a first server device, a request to bind the first server device with a multiple logical device (MLD) appliance. Based at least in part on the request, a first CXL-E connection may be established for the first server device to export a computing resource to the MLD appliance. The techniques may also include receiving, from the MLD appliance, an indication that the computing resource is available, and receiving, from a second server device, a second request for the computing resource. Based at least in part on the second request, a second CXL-E connection may be established for the second server device to consume or otherwise utilize the computing resource of the first server device via the MLD appliance.
1. A method comprising:
receiving, at a controller associated with a composable data center, a request associated with an application to be hosted on resources of the composable data center;
determining, by the controller, a first group of servers of the composable data center with available compute resources to support a number of cores requested for the application;
determining, by the controller, a second group of servers from the first group of servers that satisfy a memory-latency constraint to support the application; and
selecting, by the controller, a first server from the second group of servers to host the application based at least in part on a first resource score associated with the first server being greater than a second resource score associated with a second server from the second group of servers.
2. The method of claim 1 , wherein the request includes a domain label and each server of the first group of servers is tagged with the domain label.
3. The method of claim 1 , wherein the request includes a domain label and each one of the available compute resources is tagged with the domain label.
4. The method of claim 1 , wherein determining the second group of servers that satisfy the memory-latency constraint comprises determining that each server of the second group of servers has an available amount of a memory resource that meets or exceeds a threshold amount, the memory resource including at least one of double data rate (DDR) dynamic random-access memory (DRAM), storage class memory, or remote persistent memory.
5. The method of claim 1 , further comprising determining a memory-latency constraint score associated with each server of the second group of servers, a value of the memory-latency constraint score being based at least in part on an available amount of a memory resource for each server, wherein the first resource score and the second resource score are based at least in part on the memory-latency constraint score associated with the first server and the second server.
6. The method of claim 1 , further comprising determining an accelerator constraint score associated with each server of the second group of servers, a value of the accelerator constraint score being based at least in part on graphics processing unit (GPU) offloads for each server, wherein the first resource score and the second resource score are based at least in part on the accelerator constraint score associated with the first server and the second server.
7. The method of claim 1 , further comprising determining a storage constraint score associated with each server of the second group of servers, a value of the storage constraint score being based at least in part on storage requirements associated with each server, wherein the first resource score and the second resource score are based at least in part on the storage constraint score associated with the first server and the second server.
8. The method of claim 1 , further comprising establishing a connection with the first server for the application based at least in part on selecting the first server.
9. The method of claim 8 , wherein the connection is at least one of a compute express link (CXL) connection or a CXL over ethernet (CXL-E) connection.
10. A system comprising:
one or more processors; and
one or more non-transitory computer-readable media storing instructions that, when executed, cause the one or more processors to perform operations comprising:
receiving, at a controller associated with a composable data center, a request associated with an application to be hosted on resources of the composable data center;
determining, by the controller, a first group of servers of the composable data center with available compute resources to support a number of cores requested for the application;
determining, by the controller, a second group of servers from the first group of servers that satisfy a memory-latency constraint to support the application; and
selecting, by the controller, a first server from the second group of servers to host the application based at least in part on a first resource score associated with the first server being greater than a second resource score associated with a second server from the second group of servers.
11. The system of claim 10 , wherein the request includes a domain label and each server of the first group of servers is tagged with the domain label.
12. The system of claim 10 , wherein the request includes a domain label and each one of the available compute resources is tagged with the domain label.
13. The system of claim 10 , wherein determining the second group of servers that satisfy the memory-latency constraint comprises determining that each server of the second group of servers has an available amount of a memory resource that meets or exceeds a threshold amount, the memory resource including at least one of double data rate (DDR) dynamic random-access memory (DRAM), storage class memory, or remote persistent memory.
14. The system of claim 10 , the operations further comprising determining a memory-latency constraint score associated with each server of the second group of servers, a value of the memory-latency constraint score being based at least in part on an available amount of a memory resource for each server, wherein the first resource score and the second resource score are based at least in part on the memory-latency constraint score associated with the first server and the second server.
15. The system of claim 10 , the operations further comprising determining an accelerator constraint score associated with each server of the second group of servers, a value of the accelerator constraint score being based at least in part on graphics processing unit (GPU) offloads for each server, wherein the first resource score and the second resource score are based at least in part on the accelerator constraint score associated with the first server and the second server.
16. The system of claim 10 , the operations further comprising determining a storage constraint score associated with each server of the second group of servers, a value of the storage constraint score being based at least in part on storage requirements associated with each server, wherein the first resource score and the second resource score are based at least in part on the storage constraint score associated with the first server and the second server.
17. The system of claim 10 , the operations further comprising establishing a connection with the first server for the application based at least in part on selecting the first server.
18. One or more non-transitory computer-readable media storing instructions that, when executed, cause one or more processors to perform operations comprising:
receiving, at a controller associated with a composable data center, a request associated with an application to be hosted on resources of the composable data center;
determining, by the controller, a first group of servers of the composable data center with available compute resources to support a number of cores requested for the application;
determining, by the controller, a second group of servers from the first group of servers that satisfy a memory-latency constraint to support the application; and
selecting, by the controller, a first server from the second group of servers to host the application based at least in part on a first resource score associated with the first server being greater than a second resource score associated with a second server from the second group of servers.
19. The one or more non-transitory computer-readable media of claim 18 , wherein the request includes a domain label and each server of the first group of servers is tagged with the domain label.
20. The one or more non-transitory computer-readable media of claim 18 , wherein determining the second group of servers that satisfy the memory-latency constraint comprises determining that each server of the second group of servers has an available amount of a memory resource that meets or exceeds a threshold amount, the memory resource including at least one of double data rate (DDR) dynamic random-access memory (DRAM), storage class memory, or remote persistent memory.