I/O request performance improvement using backend as a service
Methods and apparatuses are provided for utilizing additional and available system resources to cache one or more incoming I/O write requests. A storage processor may implement a policy including a Backend as a Service (BEaaS) to use a node pair backend, a fabric network, and other resources to absorb incoming bursts of I/O write requests from a host and direct the requests temporarily to NVMe storage devices. The policy may further use local and/or remote storage arrays connected over the fabric to alleviate strain on local processing resources. The policy may provide for using cut-through writes, including for example, dual-cast operations over PCIe connections, to send I/O requests directly to the NVMe storage devices.
1 . A computer-implemented method performed by a storage processor comprising a local cache memory, a fabric network interface, and a selection manager executing a selection policy the method comprising:
receiving, at the storage processor an input/output (I/O) request from a host device;
determining, by the selection manager a memory utilization threshold for the local cache memory;
determining by the selection manager, a local memory utilization level of the local cache memory;
in response to determining that the local memory utilization level equals or exceeds the memory utilization threshold;
invoking a backend as a service policy implement ed by the storage processor to generate a remote direct memory access (RDMA) operation encapsulating the I/O request; and
transmitting, via the fabric network interface, the RDMA operation to a second memory, thereby temporarily storing the I/O request to reduce local cache memory overflow.
2 . The method of claim 1 wherein the RDMA operation is executed by a remote processor.
3 . The method of claim 1 wherein the second memory is a RAID-1 array.
4 . The method of claim 1 wherein the I/O request is stored a controller memory buffer (CMB) of the second memory.
5 . The method of claim 1 wherein the local memory utilization level includes a write pending level.
6 . The method of claim 1 further comprising determining the local memory utilization level has decreased below the memory utilization threshold and relocating the I/O request from the second memory to the local cache memory.
7 . The method of claim 1 wherein the memory utilization threshold is determined by a number of memory devices.
8 . The method of claim 1 wherein the fabric network interface includes an InfiniBand network interface.
9 . The method of claim 1 wherein the second memory is a non-volatile memory express (NVMe).
10 . A data storage system comprising:
a memory; and
at least one processor that is operatively coupled to the memory, the at least one processor being configured to perform the operations of:
receiving an input/output (I/O) request from a host device;
determining a memory utilization threshold for a local cache memory;
determining a local memory utilization level of the local cache memory;
in response to determining that the local memory utilization level equals or exceeds the memory utilization threshold;
invoking a backend as a service policy implemented by the storage processor to generate a remote direct memory access (RDMA) operation encapsulating the I/O request; and
transmitting, via a fabric network interface, the RDMA operation to a second memory, thereby temporarily storing the I/O request to reduce local cache memory overflow.
11 . The data storage system of claim 10 wherein the RDMA operation is executed by a remote processor.
12 . The data storage system of claim 10 wherein the second memory is a RAID-1 array.
13 . The data storage system of claim 10 wherein the I/O request is stored a controller memory buffer (CMB) of the second memory.
14 . The data storage system of claim 10 wherein the local memory utilization level includes a write pending level.
15 . The data storage system of claim 10 further comprising determining the local memory utilization level has decreased below the memory utilization threshold and relocating the I/O request from the second memory to local cache memory.
16 . The data storage system of claim 10 wherein the memory utilization threshold is determined by a number of memory devices.
17 . The data storage system of claim 10 wherein the fabric network interface includes an InfiniBand network interface.
18 . A non-transitory computer-readable medium storing one or more processor-executable instructions, which when executed by at least one processor cause the at least one processor to perform the operations of:
receiving an input/output (I/O) request from a host device;
determining a memory utilization threshold for a local cache memory;
determining a local memory utilization level of the local cache memory;
in response to determining that the local memory utilization level equals or exceeds the memory utilization threshold;
invoking a backend as a service policy implemented by the storage processor to generate a remote direct memory access (RDMA) operation encapsulating the I/O request; and
transmitting, via a fabric network interface, the RDMA operation to a second memory, thereby temporarily storing the I/O request to reduce local cache memory overflow.