Cost Optimization - aws and cloud compute

Cost optimization AWS

1. AWS Cost Optimization Workshop Techniques : EC2 Spot Instances Speaker Name Speaker job title / team / company Date

2. Agenda Benefits of Spot Rules of Spot Handling Interruptions Workload Examples

3. On-Demand Pay for compute capacity by the second with no long-term commitments Spiky workloads, to define needs Amazon EC2 purchase options Spot Instances Spare Amazon EC2 capacity at savings of up to 90% off On-Demand prices Fault-tolerant, flexible, stateless workloads Savings Plans and Reserved Make a 1 or 3 year commitment and receive a significant discount off On-Demand prices Committed and steady-state usage

4. Why Spot Instances? Low, predictable prices Up to 90% discount over On-Demand prices Faster results Increase throughput up to 10x while staying in budget Easy to use Launch through AWS services (e.g. ECS, Batch, EMR) or integrated third-parties Spot Instances are reclaimed with a 2-minute warning and only when On-Demand needs capacity back— no bidding!

5. Spot Instances help you save money and time! Up to 90% saved monthly Improved test result response time with Spot Instances 30 minutes without Spot Instances Two days Decreased monthly computing cost while increasing power 75% compute cost reduction More compute power with Spot Instances

6. Spot placement score The Spot placement score feature can recommend an AWS Region or Availability Zone based on your Spot capacity requirements. Spot capacity fluctuates, and you can't be sure that you'll always get the capacity that you need. A Spot placement score indicates how likely it is that a Spot request will succeed in a Region or Availability Zone. Benefits To relocate and scale Spot compute capacity in a different Region, as needed, in response to increased capacity needs or decreased available capacity in the current Region. To identify the most optimal Availability Zone in which to run single-Availability Zone workloads. To simulate future Spot capacity needs so that you can pick an optimal Region for the expansion of your Spot-based workloads. To find an optimal combination of instance types to fulfill your Spot capacity needs. You can use the Spot placement score feature for the following:

7. What happens when AWS reclaims an instance? Minimal interruptions Less than 5 % of Spot instances were interrupted in the last 3 months Alerts Automation Handling Options Terminate Stop/Start Hibernate Map to Strategy EC2 Spot rebalance recommendation An EC2 Spot Instance rebalance recommendation is a signal from that notifies you when a Spot Instance is at elevated risk of interruption. The signal can arrive sooner than the two-minute Spot Instance interruption notice , giving you the opportunity to proactively manage the Spot Instance.

8. Be time flexible to account for interruptions and/or location flexible to maximize application uptime Flexibility is key to successful adoption Instance flexible More than one instance type can get the job done Instance weighting gives you more flexibility on instance types Multiple instance types are key to resilient clusters Time flexible Region flexible OR

9. Spot pricing Smooth, infrequent changes no spikes, more predictable Up to 90% off Interruptions Happen when EC2 needs capacity Spot infrastructure Is same as On-Demand and RIs Diversify Choose different instance types, size and AZ in a single fleet The simple rules of Spot

10. To optimize Amazon EC2, combine purchase options with EC2 Auto Scaling Use Savings Plan or RIs for known, steady-state workloads On-demand , for new or stateful spiky workloads Scale using Spot for fault-tolerant, flexible, stateless workloads

11. Save up to 90% using EC2 Auto Scaling and EC2 Fleet Capacity optimized Prioritize deploying Spot Instances into greater Spot pool capacity in order to lower the chance of interruptions Lowest cost Specify what percentage of your ASG capacity should be fulfilled by On-Demand and Spot Instances and the ASG will prioritize launching Spot Instances based on price Price capacity optimized – 15 nov. 2022 Makes Spot Instance allocation decisions based on both the price and the capacity availability of Spot Instances Reduce cost, optimize performance, and eliminate operational overhead Spot Instances On-Demand Instances Reserved Instances Amazon EC2 Auto Scaling

12. Amazon EC2 Spot integrations AWS Batch Amazon EMR Amazon Elastic Container Service AWS CloudFormation Amazon SageMaker AWS Elastic Beanstalk EC2 Auto Scaling AWS Fargate AWS Gamelift Amazon Elastic Kubernetes Service

13. Spot Instances are perfect for fault-tolerant Lean on Spot for these workloads! Big data HPC CI/CD Web services C ontainers Machine Learning Batch

14. Containers + Spot = match made in heaven Containers are often stateless, fault-tolerant, and a great fit for Spot Instances Deploy containerized workloads and easily manage clusters at any scale at a fraction of the cost with Spot Instances Spot instances can be used with ECS or Kubernetes to run any containerized workload Skyscanner is a travel fare aggregator website and travel metasearch engine based in Edinburgh, Scotland “We are currently tracking 74% saving over all regions.” —Paul Gillespie, Principal Architect/Tribe Lead

15. Workload example: Big data Spot Instances provide acceleration, scale, and cost savings to run hyper-scale workloads for data analysis Scale to large numbers of parallel nodes via Fleet Use Spot Instances with Amazon EMR, Hadoop, or Spark to process massive amounts of data Amazon EMR “ A job that took weeks in our data center, due to limited resources, took hours on Spot thanks to the great parallelism, in a very cost-efficient price.” Shay Asoolin , Sr. Director Development infrastructure, Mobileye

16. Workload example: CI/CD Configure Jenkins with the EC2 spot plug-in to automatically scale a fleet of spot instances based on the number CI/CD jobs Increase cost savings by leveraging older generation instances for CI, as these processes do not require a lot of power for testing “By using AWS Spot instances, we've been able to save up to 75 percent a month simply by changing four lines of code. It makes perfect sense for saving money when you're running continuous integration workloads or pipeline processing.” —Matthew Leventi , Lead Engineer, Lyft

17. Workload example: Web services Scale, throughput and deep cost savings for large-scale web operations Launch and manage a collection of diversified Spot Instances across pools via EC2 Fleet and ASG NEW! Include Spot with RIs and On-Demand in a single ASG Quantcast Scales Ad Services Saves 60% Using Amazon EC2 Spot Instances “As we roll out more infrastructure to AWS, Amazon EC2 Spot Instances are helping us control costs and scale our systems to meet demand.“ —Leah Blank, Senior Systems Engineer, Quantcast Amazon EC2 Auto Scaling Amazon EC2 Fleet

18. Workload example: HPC Accelerate HPC workloads such as genomic sequencing, CFD and algorithmic trading by running massively parallel jobs Run multiple projects simultaneously; launch & de-commission 1000’s of nodes Spot Auto Scaling groups; F1(FPGA), eg1 (Elastic GPUs), Cluster GPU instances to accelerate processing Illumina saves nearly $400,000 monthly, Speeds Genomics Analysis using Spot Instances “We are able to offer our customers a lower cost, high-performance genomic-analysis platform, which can help them speed their time to answers.“ —Andy Nelson, Associate Director, Informatics & Cloud Operations, Illumina Amazon EC2 Fleet AWS CloudFormation AWS Batch

19. Consumer apps B2B enterprise tech Research Sports, media, & entertainment Financial services AdTech & MarTech Customers across different industries and verticals use Spot

20. resources https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/spot-best-practices.html https://aws.amazon.com/blogs/compute/diving-deep-into-ec2-spot-instance-cost-and-operational-practices/ https://aws.amazon.com/blogs/compute/efficiently-scaling-kops-clusters-with-amazon-ec2-spot-instances/ https://aws.amazon.com/ec2/spot/pricing/ https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/ec2-fleet-allocation-strategy.html https://aws.amazon.com/blogs/compute/introducing-price-capacity-optimized-allocation-strategy-for-ec2-spot-instances/

21. Q&A