NetApp Cloud Volumes Service
for AWS: Benchmarks

Putting Cloud Volumes Service for AWS to the test. These benchmarks
show the performance Cloud Volumes Service for AWS brings to the cloud

img-workloads
logo-aws

What does Cloud Volumes Service for AWS offer your workloads?

  • Managed cloud service providing NFS or SMB protocols, or both at the same time. No need to manage infrastructure of Windows or Linux servers just to provide file services
  • On-demand scaling, without having to create new data stores
  • Consistent high performance, over 200k IOPS
  • Designed for five nines of availability and eight nines of durability
  • Save money by adjusting performance tier without moving data
  • Protect your database with zero-impact incremental snapshots, only keep and pay for new writes
  • Sync or migrate your data from on-prem to the cloud without the need to reconfigure your applications
  • Faster time to market by accelerating development and test using quick copies from snapshots
  • File Services
  • Oracle
  • MySQL
  • AWS EFS

File Services

File Services Workloads - IOPS

Our test indicates performance a customer could achieve with single or multiple clients against a single instance of Cloud Volumes Service for AWS. The test reveals that the maximum I/O that can drive from a single Cloud Volumes Service instance to a single client is ~60,000, whether 1K, 4K or 8K random read, and 50,000 down to ~40,000 for the writes. However, with multiple clients, a single instance of Cloud Volumes Service can provide more than 200k IOPS.

image-graph-1-outlines

File Services Workloads - Throughput

Our sequential test was done using the exact same tests and procedures. In this case the maximum amount of I/O that a single client is able to drive is right around 4-4.5Gbps reads or writes. Like in the previous sample, a single client is not able to drive the maximum capability of a Cloud Volumes Service volume, which can achieve around 2,250MBps for 16K block sizes, and over 3,000MBps for 32K and 64K block sizes.

For a deeper understanding of how to best run Cloud Volumes Service on AWS check out our 3 part blog series on the subject.

image-graph-2 (1)

Oracle Slob2 workloads

Oracle IOPS with EBS and Cloud Volumes Service

The graphs compare the performance when using EBS and when using NetApp Cloud Volumes Service for AWS. The graph on the right shows that Oracle is able to drive 250,000 file system IOPS at 2ms when using the c5.18xlarge instance and a single volume provisioned from the Cloud Volumes Service, or 144,000 file system operations at below 2ms using the c5.9xlarge.

guide-graph-3

image-graph-3-1

Oracle IOPS at various read/write ratios

The graph to the left provides more performance examples of how Oracle workloads behave on Cloud Volumes Service for AWS when run on the same c5.18xlarge EC2 instance shown above. Note that Cloud Volumes Service provides a high level of IOPS at 2 ms or below across a range of read/write ratios. Also note that in all cases adding a second volume increases the IOPS provided.

For more information on Oracle Performance and Storage Comparison in AWS: Cloud Volumes Service, EBS, EFS.

guide-graph-4

image-graph-4 (2)

MySQL Workloads

MySQL gets 25,000 IOPS

The graph to the right shows the performance of a MySQL workload running on c5.18xlarge. Run against the single Cloud Volumes Service instance, we were able to generate close to 25,000 IOPS at 4 ms latency, and 22,000 IOPS at 3 ms latency.

guide-graph-3-outlines

image-graph-5-3

AWS Elastic File Storage vs. Cloud Volumes Service for AWS

I/O Comparison of Cloud Volumes Service and AWS EFS

The graphs demonstrate how Cloud Volumes Service for AWS and EFS compare when running random and sequential workloads.

The following are the instance and limits in the test:

Instance limits:

Elastic File System: Maximum 250/MB/s per instance throughput
Cloud Volumes: 1GB/s maximum per instance throughput (512MB/s read + 512MB/s write)

Limits:

Elastic File System: Maximum 7,000 IOPS per volume (as documented by AWS)
Cloud Volumes: ~200,000 maximum IOPS per volume as tested

image-graph-6 (2)