- Run date – Nov 1, 2021
- We got below configuration from AWS for this benchmarking in zone us-west-2d
- One c5.12xlarge with Ubuntu 20.04 (EBS optimized)
- One EBS volume of type IO2, 64000 IOPS configuration, no snapshot
- No optimization at OS level
- Treat this as our review of single EBS volume performance with above configuration. Note that there is a throughput limit of 1000 MB/s on single EBS volume, we could not push more than this. At least we verified if we can hit promised throughput per EBS volume.
- Used direct IO and/or sync to avoid cache.
- Picked the best possible peak numbers and corresponding lowest 99.9 percentile latencies.
- We used same workload (jobs/threads) that we published the peak numbers for 1m block size for sustained run.
- Whether an application can fully utilize the bandwidth (The advertised “up to” number) really depends on which block size, read/write ratio, sequential or random access workload, that application is using.
- We can help identify which storage system is optimal for your application in terms of performance.
- Few observations:
- Sequential write latencies seems slightly higher than random write latencies (throughput remains same though).
- Can achieve promised 1000 MB/s throughput only with large block sizes (eg. 64k, we did not try 16k and 32k).
- We wrote up to 5TB of data to benchmark sustained writes. Throughput seems to sustain at 1000 MB/s with slight fluctuation in latency which is acceptable. We thought 5TB is reasonable for sustained test. Results might change when writing much more data than this.
- Setup is pretty straight forward
- If time permits, we will run benchmarks to see how performance scales with increase in number of volumes shared with one or more EC2 instances.
- If you are interested, use with caution as ballpark performance numbers as it could vary depending on multiple factors. These are the numbers we got in our run. We are not liable for any loss or damages caused by using this data.
