Part 2: Identification of candidates for Spot

Written by Krishnan S G


In the previous part, we touched upon what spot is and why migrating to spot provides great savings. This part will try to realize the important factors we need to understand to identify instances that would give the highest ROI on migration to Spot.

Identifying the instances which would be suitable for Spot is a hectic and mammoth task to master. Taking help from Cloudwiry’s Automated Spot module will help you save time and money. The tool performs statistical analysis of usage patterns, identifying ideal candidates for Spot with the highest ROI and least efforts from the user’s side.

(Here is a screenshot from our app where clients can see an overview of the candidates for spot migration identified for respective services.)

 

 

A deep dive into the identification

This section discusses a few workload types that we believe are easiest to migrate and have a high ROI.

 

Auto Scaling Groups

EC2 Auto Scaling groups are best suitable for Spot since these workloads are Stateless and fault-tolerant.

(Here is a screenshot from our app where clients can see an overview of the candidates for spot migration identified for ASG)

 

Cloudwiry’s Spot Module identifies the best instance types with the best configuration-price ratio while providing you with these instance type recommendations, maximizing savings and minimizing interruption risk.

 

 

To minimize risk further, Cloudwiry’s Spot module ideally recommends using a mixed instance policy, which allows having multiple instance types and a combination of OnDemand and Spot.

 

Elastic Map Reduce

Due to the fault-tolerant nature of big data workloads on EMR, they can continue processing, even when interrupted. Running EMR on Spot Instances drastically reduces the cost of big data, allows for significantly higher compute capacity, and reduces the time to process big data sets.

 

Cloudwiry’s Spot module recommends the ideal number of instances for Master, Core, and Task nodes, balancing the number of instances on OnDemand and Spot, maximizing savings, and minimizing interruption risk.

 

CICD

CI/CD workloads are fault-tolerant and are built to take advantage of Spot Instances’ cost savings. You can launch your CI/CD pipelines in the AWS cloud quickly and cost-effectively with the EC2 Spot Fleet plug-in.

 

 

Cloudwiry’s Spot module ideally recommends you the best spot instance fleet with the best configuration-price ratio while providing you with these instance type recommendations, maximizing savings and minimizing interruption risk.

Container Service and Batch

Containers and Batch are often stateless and fault-tolerant and are a great fit for Spot Instances. You can create Amazon Batch, Amazon ECS, Amazon EKS, or self-managed Kubernetes clusters on Spot to run containerized workloads of any scale. AWS integration makes the migration easy.

Cloudwiry helps identify and migrate the ECS containers and Fargate tasks to Spot instances and Fargate Spot tasks, respectively.

 

SageMaker

Amazon SageMaker is a fully managed service that provides every developer and data scientist with the ability to build, train, and deploy machine learning (ML) models quickly. Amazon SageMaker makes it easy to train machine learning models using managed Amazon EC2 Spot instances. Managed spot training can optimize the cost of training models up to 90% over on-demand instances. Amazon SageMaker manages the Spot interruptions on your behalf.

Managed Spot Training uses Amazon EC2 Spot instance to run training jobs instead of on-demand instances. You can specify which training jobs use spot instances and a stopping condition that specifies how long Amazon SageMaker waits for a job to run using Amazon EC2 Spot instances.

Jobs that are interrupted, automatically resume the training from the last checkpoint instead of restarting.

Cloudwiry identifies & analyses your SageMaker training workloads and provides spot recommendations which would provide savings up to 90% on your model training cost.

 

 

This brings us to the end of the second part of our series on Spot Instances and Migration. Identification is just the stepping stone to complete migration to Spot. The migration is equally hectic and a mammoth task. In the next part, we’ll demonstrate how Cloudwiry’s Automated Spot module can be used to easily migrate and monitor the migration.

For more strategies on optimizing your cloud, schedule a call with our cloud experts or drop us an email at hello@cloudwiry.com.

GET STARTED
cloudwiryqa
Share
This