Sign in to my dashboard Create an account
Menu

Saving 73% on compute costs in the cloud with Spot Ocean

sea turtle
Contents

Share this page

Karthikeyan Nagalingam
Karthikeyan Nagalingam
128 views

To reduce capital expenditures (capex) and to better align costs to revenue generation, customers like you are moving workloads to cloud. And to make better use of cloud, customers are shifting to microservice-style deployments, which are typically container based and which replace some capex with operational expenditures (opex).

The performance of these workloads is based on compute power, and more compute usage costs you more. So, the question is “How can I reduce my opex in the cloud for batch and stream processing?”

And the answer is the Spot Ocean by NetApp® application-scaling service.

Proof of concept using Spot Ocean

NetApp recently worked with a customer who was moving their workloads to the cloud for their production and their R&D applications. The customer’s main challenges were the cost of compute, the cost of data storage, and how to reduce the company’s reliance on a single cloud provider.

This customer ingests data from Internet of Things (IoT) devices by using inline streaming applications for batch and stream processing. Batch processing is performed on a data lake that’s running in Kubernetes, and stream processing is carried out in a multicloud environment by using Amazon FSx for NetApp ONTAP. The archived data is then moved to NetApp StorageGRID® object storage. The AI and analytics team processes the data from the NetApp object storage and the multicloud attached storage.

To optimize cloud usage and to reduce costs, we conducted a proof of concept with the customer on their platform-as-a-service (PaaS) and infrastructure-as-a-service (IaaS) environments.

For the PaaS environment, the customer was using AWS big data analytics products. To optimize the compute for those products, we used Spot Ocean by NetApp.

For the IaaS environment, the customer was using Amazon Elastic Compute Cloud (Amazon EC2) instances with Amazon Elastic Block Store (Amazon EBS) volumes, Amazon FSx, and Amazon Simple Storage Service (Amazon S3). We optimized all those elements with a combination of Ocean and Amazon FSx for NetApp ONTAP.

Real-life proof that Ocean reduces costs

In our proof of concept, we found that:

  • Ocean optimized compute with app-driven provisioning of Amazon EC2 instances. By optimizing containers, Ocean reduced costs and optimized the entire PaaS environment. The customer reduced reserved instances by 70% . The result was a significant cost savings in Amazon EC2 instances for the customer.
Data graph
  • Amazon FSx for NetApp ONTAP optimized data storage through several techniques (thin provisioning, deduplication, space-free snapshots, and cloning), and by automatically tiering unused or infrequently accessed data to Amazon S3. At the same time, it provided access to the same volumes through both Windows (SMB protocol) and Linux (NFS protocol) environments. FSx for ONTAP supported integration into Amazon Elastic Kubernetes Service (Amazon EKS) and Amazon Elastic Container Service (Amazon ECS) environments, all while delivering high performance with minimal latency.

Overall, this proof of concept showed that Ocean cut costs by 73% for this customer’s container environment in AWS. With this Spot by NetApp technology, the customer significantly reduced their opex for stream and batch processing in the cloud.

See for yourself how Ocean can optimize your cloud compute and storage and lower your opex. Request a free trial today.

Karthikeyan Nagalingam

Karthikeyan Nagalingam is a Principal Technical Marketing Engineer at NetApp for NetApp XCP, Fpolicy, Filesystem Analytics and Antivirus. His previous roles in Emerging Technology Solutions involved in Pre-Sales and Post-Sales technical activities with fields, partners and customers. He holds an Master of Science in Software Systems from Birla Institute of Technology and Science.

View all Posts by Karthikeyan Nagalingam

Next Steps

Drift chat loading