Rhizomenet scales massively with PowerDown Spot management

zome

Rhizomenet provides an Augmented Reality Message application called Zome-AR which offers an immersive real-time space and time based message app.

Rhizomenet is a startup and needs to be able to massively scale their application to support a large and rapidly growing base of users. While AWS EC2 and ECS provides the ability to scale, doing so while containing costs is a challenge.

Zome Scaling

The Zome application backend service is hosted via an AWS ECS containerized application running on an EC2 cluster behind an Auto Scale group. The cluster instances are the newer C5.*Large with Auto Scale rules providing automatic scaling based on Cloud Watch CPU metrics. The service is highly available and scalable.

Spot Instances

Rhizomenet looked at using Spot instances in their application cluster as a way to boost capacity and contain costs. Amazon EC2 Spot Instances offer spare compute capacity available in the cloud at steep discounts compared to On-Demand instances, often up to 80%. This allows AWS to sell their excess compute capacity, but with a major restriction — Spot Instances can be terminated when AWS needs the capacity for On-Demand servers with only 2 minutes of warning.

To overcome the transient nature of Spot instances, it would be ideal to mix Spot and On-Demand instances in one Auto Scale group. However, the native AWS Auto Scale groups do not support this ability. Rhizomenet investigated using AWS Spot Fleets which can support On-Demand instances with the Spot fleet, but in the event that AWS reclaims the Spot instances, capacity can be greatly reduced, impacting performance and availability. So Spot fleets were not the ideal solution either.

Transparent Spot Management

PowerDown solved Rhizomenet's scaling problem by augmenting the native AWS Auto Scale groups with PowerDown's ability to mix and match Spot and On-Demand instances in one Auto Scaling group.

The PowerDown Spot manager extends the AWS Spot facility by directly managing the lifecycle of Spot instances. Rhizomenet defines a desired percentage of Spot and On-Demand instances for their Zome Auto Scale groups and PowerDown monitors the group and ensures the group composition matches this desired percentage. Once the Spot market recovers, PowerDown resumes launching Spot instances to maintain the correct balance.

PowerDown creates Spot instances and migrates workloads from the previous On-Demand instances without downtime or impacting availability. If AWS needs to reclaim Spot instances, AWS will issue a warning 2 minutes prior to terminating the Spot instance. PowerDown listens for these warnings and proactively launches replacement On-Demand servers before the Spot servers are terminated. In this manner, PowerDown preserves both availability and capacity with no intervention required by Rhizomenet.

Configuration

Rhizomenet configured PowerDown to maintain a 75% Spot and 25% On-Demand instance balance. This reduced their EC2 compute costs by 59% with no impact to availability or performance.

zome-scale

Results

Theodore Yue Wang of Rhizomenet said: "With PowerDown, we are saving about 60% of our cloud costs each month." "It took only a couple of minutes to configure and we did not have to change any of our AWS configuration or processes. PowerDown is super simple and works without interaction in the background."

Learn More

To learn more about PowerDown, please visit https://www.sensedeep.com/powerdown. PowerDown offers a free 30 day trial.

References

© SenseDeep® LLC. All rights reserved. Privacy Policy and Terms of Use.

Consent

This web site uses cookies to provide you with a better viewing experience. Without cookies, you will not be able to view videos, contact chat or use other site features. By continuing, you are giving your consent to cookies being used.

OK