Includes AWS Glue Data Catalog Integration and Graceful Scaledown
Starburst Enterprise(SEP) Cloud 203e is now available on AWS Marketplace. With Starburst Enterprise Cloud 203e, we’ve made it even easier and more economical to set up and scale SEP to query data on AWS.
Starburst Enterprise is now based on Starburst 203e, which we released for download in July. As a result, SEP Cloud is now enhanced with:
- AWS Glue Data Catalog integration
- Auto scaling with “Graceful Scaledown”
As with the previous release, SEP Cloud also includes the state of the art Cost Based Optimizer (CBO) for superior performance which ultimately reduces your costs.
AWS Glue Data Catalog Integration for Faster Setup
We’ve added AWS Glue Data Catalog support into 203e so you can seamlessly query your S3-backed tables defined in the catalog. The advantage of using AWS Glue Data Catalog with SEP is simple — it provides an alternative to managing your own Hive Metastore to work with your S3 data. Using SEP Cloud 203e, it’s a single click in the Cloud Formation template, and SEP will be deployed and configured to immediately starting querying S3 data defined by tables in Glue.
You can read more about our AWS Glue Data Catalog support in the SEP AWS and Hive Connector documentation.
Shrink SEP Clusters to Save Costs–Without Interrupting Users
In SEP Cloud 203e, we have added what we refer to as “Graceful Scaledown.” Using the auto scaling feature in SEP Cloud, you can elastically grow and shrink your cluster based on business needs. When there is a desire to shrink the cluster (e.g., a weekend drop in usage), EC2 instances running SEP will be terminated.
However, it’s often the case there will still be users running queries on the SEP cluster, and terminating instances will kill those queries, which is undesirable; As always, we’ve listened to the many SEP users who requested an improvement, and now gracefully shutdown the nodes as the cluster is scaled down.
When it comes time to shrink the cluster, the SEP worker nodes selected to be terminated are notified. This could be a small or large subset of the workers or even the entirety of the workers. Once the selected workers are notified, they will finish the current work scheduled before shutting down. This also means the workers will stop servicing new requests and any new requests will be processed by the workers that will remain up.
This feature is very important as it allows you to manage your cluster costs while also not disrupting user’s work in progress.
Getting Started with SEP Cloud
Our mission is to make SEP easy to use at a low cost. SEP Cloud is simple to deploy and available with an hourly, low cost consumption-based pricing. Because of SEP’s separation of storage and compute, you can scale the cluster up and down and pay for only what you use. And with Cost-based query optimization enabled, SEP Cloud delivers up to 18x faster SEP performance than AWS EMR on SEP, and at 1/7th the cost!
You can run Starburst Cloud from the AWS Marketplace. Try it today!
Need help with getting SEP set up or want to learn more?
Contact us any time!
