# US 11,451,455 B2 · Technologies for latency based service level agreement management in remote direct memory access networks > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 11,451,455 B2 - Document ID: US11451455B2 - Status: Granted - Kind: B2 - Application number: 16/540,800 - Filed: 2019-08-14 - Published: 2019-12-05 - Granted: 2022-09-20 - Earliest claimed: 2019-08-14 - Assignees: Intel Corporation - Applicants: Intel Corporation - Inventors: Mrittika Ganguli; Arvind Srinivasan; Slawomir Putyrski; Donald E. Wood - Web page: https://patentgrep.com/patent/US11451455B2 ## CPC - H04L 12/4625: {Single bridge functionality, e.g. connection of two networks over a single bridge} - H04L 41/5009: Determining service level performance parameters or violations of service level contracts, e.g. violations of agreed response time or mean time between failures [MTBF] - G06F 13/28: using burst mode transfer, e.g. direct memory access {DMA}, cycle steal (G06F13/32 takes precedence) - G06F 15/17331: {Distributed shared memory [DSM], e.g. remote direct memory access [RDMA]} - H04L 12/06: Answer-back mechanisms or circuits - H04L 41/5022: by giving priorities, e.g. assigning classes of service - H04L 43/0852: Delays - H04L 49/9047: including multiple buffers, e.g. buffer pools - H04L 47/6215: {Individual queue per QOS, rate or priority} - H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M) - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) ## Abstract Technologies for latency based service level agreement (SLA) management in remote direct memory access (RDMA) networks include multiple compute devices in communication via a network switch. A compute device determines a service level objective (SLO) indicative of a guaranteed maximum latency for a percentage of RDMA requests of an RDMA session. The compute device receives latency data indicative of latency of an RDMA request from a host device. The compute device determines a priority associated with the RDMA request as a function of the SLO and the latency data. The compute device schedules the RDMA request based on the priority. The network switch may allocate queue resources to the RDMA request based on the priority, reclaim the queue resources after the RDMA request is scheduled, and then return the queue resources to a free pool. Other embodiments are described and claimed. ## Claims ### Claim 1 A compute device comprising: circuitry to: receive a service level objective for a remote direct memory access (RDMA) session established with a second compute device, the service level objective to indicate a latency requirement for at least a portion of RDMA requests included in the RDMA session; monitor RDMA requests in the RDMA session to determine a latency of an RDMA request included in the RDMA session; and prioritize the RDMA request for submittal to an input/output (I/O) device hosted by the compute device based on the service level objective and the determined latency of the RDMA request. ### Claim 2 (depends on 1) The compute device of claim 1, wherein to monitor RDMA requests in the RDMA session comprises the circuitry to monitor one or more counters indicative of RDMA session traffic, the one or more counters included in a network interface controller of the compute device. ### Claim 3 (depends on 1) The compute device of claim 1, wherein the I/O device hosted by the compute device comprises, a storage device, or an accelerator device. ### Claim 4 (depends on 1) The compute device of claim 1, wherein to prioritize the RDMA request for submittal to the I/O device comprises to prioritize the RDMA request compared to other requests to the I/O device based on whether a time to complete the RDMA request exceeds a time interval associated with the service level objective. ### Claim 5 (depends on 1) The compute device of claim 1, wherein to prioritize the RDMA request for submittal to the I/O device comprises to prioritize the RDMA request compared to other requests to the I/O device based on a latency of the I/O device to which the RDMA request is submitted. ### Claim 6 (depends on 5) The compute device of claim 5, wherein the I/O device hosted by the compute device comprises a Non-Volatile Memory Express (NVMe) storage device or an accelerator device. ### Claim 7 (depends on 1) The compute device of claim 1, wherein the latency requirement for at least a portion of RDMA requests included in the RDMA session comprises a guaranteed maximum latency for the at least a portion of RDMA requests included in the RDMA session. ### Claim 8 A method comprising: receiving, by circuitry of a compute device, a service level objective for a remote direct memory access (RDMA) session, the service level objective to indicate a latency requirement for at least a portion of RDMA requests included in the RDMA session; monitoring RDMA requests in the RDMA session to determine a latency of an RDMA request included in the RDMA session; and prioritizing the RDMA request for submittal to an input/output (I/O) device hosted by the compute device based on the service level objective and the determined latency of the RDMA request. ### Claim 9 (depends on 8) The method of claim 8, wherein monitoring the RDMA requests in the RDMA session comprises monitoring one or more counters indicative of RDMA session traffic, the one or more counters included in a network interface controller of the compute device. ### Claim 10 (depends on 8) The method of claim 8, wherein prioritizing the RDMA request for submittal to the I/O device comprises prioritizing the RDMA request compared to other requests to the I/O device based on whether a time to complete the RDMA request exceeds a time interval associated with the service level objective. ### Claim 11 (depends on 8) The method of claim 8, wherein prioritizing the RDMA request for submittal to the I/O device comprises prioritizing the RDMA request compared to other requests to the I/O device based on a latency of the I/O device to which the RDMA request is submitted. ### Claim 12 (depends on 11) The method of claim 11, wherein the I/O device hosted by the compute device comprises a Non-Volatile Memory Express (NVMe) storage device or an accelerator device. ### Claim 13 (depends on 12) The method of claim 12, wherein the latency requirement for at least a portion of RDMA requests included in the RDMA session comprises a guaranteed maximum latency for the at least a portion of RDMA requests included in the RDMA session. ### Claim 14 One or more non-transitory computer-readable storage media comprising a plurality of instructions stored thereon that, in response to being executed, cause a compute device to: receive a service level objective for a remote direct memory access (RDMA) session established with a second compute device, the service level objective to indicate a latency requirement for at least a portion of RDMA requests included in the RDMA session; monitor RDMA requests in the RDMA session to determine a latency of an RDMA request included in the RDMA session; and prioritize the RDMA request for submittal to an input/output (I/O) device hosted by the compute device based on the service level objective and the determined latency of the RDMA request. ### Claim 15 (depends on 14) The one or more non-transitory computer-readable storage media of claim 14, wherein to monitor RDMA requests in the RDMA session comprises to monitor one or more counters indicative of RDMA session traffic, the one or more counters included in a network interface controller of the compute device. ### Claim 16 (depends on 14) The one or more non-transitory computer-readable storage media of claim 14, wherein to prioritize the RDMA request for submittal to the I/O device comprises to prioritize the RDMA request compared to other requests to the I/O device based on whether a time to complete the RDMA request exceeds a time interval associated with the service level objective. ### Claim 17 (depends on 14) The one or more non-transitory computer-readable storage media of claim 14, wherein to prioritize the RDMA request for submittal to the I/O device comprises to prioritize the RDMA request compared to other requests to the I/O device based on a latency of the I/O device to which the RDMA request is submitted. ### Claim 18 (depends on 17) The one or more non-transitory computer-readable storage media of claim 17, wherein the I/O device hosted by the compute device comprises a storage device or an accelerator device. ### Claim 19 (depends on 18) The one or more non-transitory computer-readable storage media of claim 18, wherein the latency requirement for at least a portion of RDMA requests included in the RDMA session comprises a guaranteed maximum latency for the at least a portion of RDMA requests included in the RDMA session.