If you would prefer to disable cookies, please click below:
Like many enterprise IT organizations, NetApp IT is looking for ways to use public cloud compute to enable cloud bursting and to reduce its data center footprint. But public cloud has been cost prohibitive. Until now.
NetApp deployed Spot by NetApp to enable cloud agility while slashing compute costs in AWS by up to 75% per instance.
I had people calling me asking if we knew anything about the acquisition. We had no idea, but the solution was so good that we were going to use it regardless.
Mekka Williams Principal Engineer, NetApp Hybrid Cloud Engineering team
Inside the NetApp Common Test Lab (CTL) cloud test ecosystem, where NetApp engineers test updates to Cloud Volumes ONTAP® for AWS, cloud compute costs were rising. EC2 usage for Cloud Volumes ONTAP for AWS testing averaged about 85 instances per day. Monthly spending was projected to top $110,000 for Cloud Volumes ONTAP for AWS alone. Eyebrows were raised.
Enter Spot (now Spot by NetApp1), a novel and ingenious way to take advantage of hyperscale spot instances at discount prices without the risk. Elastigroup by Spot uses machine learning models to monitor compute usage and predict spot interruptions in advance, migrating at-risk instances to new instances to avoid disruption.
When all is said and done, the Engineering team expects to save 60% to 65% in compute costs across the workflows it can migrate to Spot. Back-of-the-envelope calculations show that could mean $35K per month or more in savings for Cloud Volumes ONTAP for AWS testing environments. Perhaps even more significantly, Spot is driving a cultural change in the way NetApp Engineering thinks about its on-premises resources.