SNIA SSS Performance Test Specification (PTS) Testing Service
The SNIA Solid State Storage Initiative (SNIA SSSI) is pleased to recommend a testing service where interested parties may submit their SSD products for testing to the SSS Performance Test Specification.
PTS test services support most form factor, interface and protocol SSDs including mSATA, SATA, SAS, M.2, U.2, PCIe x8x16 SSDs, NVMe, NVDIMM-N and others. The tested device must be recognized as a logical Block IO device by CentOS 6.x and must support PURGE (via SECURITY ERASE, FORMAT UNIT, or equivalent proprietary method of returning the Flash LBAs to a state as if no writes had occured).
Testing is based on the SSS PTS-E version 1.1 and PTS-C v1.2. The following tests are available. Visit the SSS Performance Test Specification page for more information on the tests.
1. SNIA BASIC Benchmark
- IOPS: RND IOs Per Second (56 R/W x Block Size matrix) - Pre conditioned (PC) with Workload Dependent Block Size (BS)/ RW mix Loop to Steady State (SS)
- Throughput (TP): Large Block SEQ R/W in MB/sec - Pre conditioned (PC) with Workload Dependent BS/RW loop to Steady State (SS)
- Latency (LAT): Ave & Max Response Times at 3 BS & 3 R/W Mixes at Thread Count 1 / Queue Depth 1 (TC1/QD1) - Pre conditioned (PC) with Workload Dependent BS/RW loop to Steady State (SS)
2. SNIA Advanced: Saturation, Optimized and Garbage Collection
- Write Saturation (WSAT): Continuous application of workload to see performance behavior evolution over Time, TGBW or Drive Fills - PURGE, FOB, Workload Dependent Pre conditioning (PC), to Steady State (SS)
- Demand Intensity Response Time Histograms (DIRTH): Profile IOPS and Response Time Quality of Service (QoS) at varying Outstanding IOs (OIO) - PURGE, PC with WSAT Workload Dependent Access Pattern to SS
- Cross Stimulus Recovery (XSR): Assess impact of changes of sustained workloads on Bandwidth and Response Time performance recovery (Garbage Collection (GC)) - PURGE, PC with super saturated Workload Dependent Access Pattern to SS
- Host Idle Recovery (HIR): Assess impact of increasing host idle times on performance recovery (GC) - PURGE, PC with super saturated Workload Dependent Access Pattern to SS, varying periods of host idle periods
- Enterprise Composite Workload (ECW): Apply JEDEC composite workload - PURGE, PC with WSAT Mixed BS/RW mix probability streams to SS in order to profile IOPS and Response Time QoS at varying OIO thresholds.
Testing will be conducted by Calypso Systems, a certified SSS PTS testing facility. Contact firstname.lastname@example.org for pricing.
Participants must submit two (2) samples of each SSD model to be tested and provide prepaid return express shipment bills (FedEx, DHL or UPS).
Testing will take approximately 3-4 weeks to complete.
Any failed test, or test that will not complete, will be tested twice and error logs will be provided.
All product test result data will be kept confidential.
Test results are provided in standard SNIA Report Format as specified in the SSS Performance Test Specification.
For more details, contact email@example.com.