Highlights from AWS FEST: Compute Edition

With nearly 650 participants, 9 expert speakers, and 6 topics under the "Compute Optimization" theme, this AWS FEST recap is worth the read!
Share This Post

Q2 2023 AWS FEST is behind us but that doesn’t mean we’ve moved on! With nearly 650 participants, 9 expert speakers, 6 topics all focused on Compute Optimization, over 11,000 emojis sent (a record, by the way), we lost track of the math needed to add up all of the amazing knowledge put out there and are still digesting.

We focused a lot on compute FinOps/Cost Optimization, particularly due to the fact that IT budgets are continuing to shrink and cloud waste has been on the rise… but not only. We got some great tools on how to make your Lambda runtimes more efficient, must-have AWS tools and third party services to help with usage transparency and simplifying management of your provisioned resources, how to maximize your sustainability, essential KPIs, and how to manage multi-cloud operations effectively.

If you couldn’t make it to this session or are looking to review all of the valuable tips and tools discussed, we’ve compiled some highlights below.

Highlights from AWS FEST: Compute Edition – May 2023

Modern Compute and Optimization in AWS with Jon Hammant

We were so happy to have Jon join us for the event. He leads the Compute business for Amazon Web Services in UK & Ireland and as part of this role, he heads a number of businesses, including EC2, Modern Applications, Graviton, Containers/Servers and networking. He has a wealth of knowledge on how AWS is really stepping on the gas in terms of helping customers focus on cost reduction.

What are some practical approaches to optimize AWS spend, according to Jon?

First, Jon ran through the 5 pillars of Cost Optimization for those who have completed their ‘lift and shift’ which included:

  • Rightsizing and modernizing your instances – Take advantage of the 475 available instances on AWS. Make sure you have that correct combination of CPU, memory, storage and networking. Don’t fire up a large instance by default. Jon gave some great tips on how to stop paying for idle instances, including creating schedules within AWS Instance Scheduler and tagging non production instances.
  • Leveraging the right pricing model – Mix in upfront payment commitments. Jon recommends logging into AWS Cost Explorer and reviewing your recommendations, perhaps utilizing flexible purchase options in order to pay lower rates.

Jon also recommends moving to Graviton if you are running Java, Python or Node and if you are using x86 based instances. He provides steps to getting started which include test running your application to reach a 40% savings potential.

Here’s Jon’s full session where he gives many other great tips and tools on modernizing your compute.

AWS Application Composer: the app building future we need with Danielle Heberling

As a Senior Software Engineer and AWS Community Builder in the Serverless space, we loved having Danielle on, hearing that she not only spends her time helping AWS customers optimize their serverless workloads, she mentors folks from underrepresented groups that would like to make a career switch into tech.

Danielle was on the original builders team at Stackery for what is now AWS Application Composer, and she brought her wealth of expertise on the app that just launched at AWS re:Invent 2022, which lets you build serverless applications through a visual designer.

What are the challenges of AWS Lambda and how does AWS Application Composer help, according to Danielle?

Danielle recommends building your serverless apps using Infrastructure As Code (IaC) in order to have a deployable written representation of your resources. Also it is helpful to have it in version control to make it easy to revert to a previous version of your infrastructure when mistakes happen. IaC also beats clicking through the console for multiple environments, making sure you have the right DevOps order/settings.

Danielle ran through the challenges of getting started with IaC and why this made her start searching for a solution which was more visual design oriented (drag & drop) rather than typescript oriented. Specifically, getting IaC right for many individual resources and connecting those resources together (i.e. IAM resources, event triggers) proved to be difficult.

Enter AWS Application Composer, which makes it incredibly easy to drag and drop resources to generate IaC (i.e. drag and connect API Gateway –> AWS Lambda function –> DynamoDB table)

Here’s Danielle’s full session where she demonstrates AWS Application Composer and talks about additional features she would love see going forward.

Your Lambdas, in Rust! with Luciano Mammino

Luciano is currently a Senior Architect at fourTheorem and he loves talking full-stack web, Node.js and Serverless. He was the perfect person to run through how you can get the maximum out of AWS Lambda.

What should we know about AWS Lambda and how can Rust help us optimize it and efficiently build apps according to Luciano?

  • AWS Lambda is a great tool for event-driven computation, allowing us to perform actions such as HTTP requests using API Gateway, create workflows for files in S3, orchestrate Step Functions or even run things on a schedule like website crawling
  • The cost of running AWS Lambda can seem inexpensive at first looking at stand-alone executions, but if you are running thousands of Lambdas per second, costs can escalate
  • In AWS Lambda, you don’t get to choose virtual CPU. It depends on the amount of memory. Therefore, if you need to allocate more CPU, it will cost you.

Luciano shows us how to run AWS Lambda in Node.js and notes that Rust isn’t an official runtime of Lambda. However we can write our own custom Lambda easily using a custom runtime without starting from scratch. Luciano runs through how to use an open source runtime built to run Rust.

And finally, why even bother with Rust? Luciano explains in detail what Rust focuses on and what it brings to the table.

Be sure to watch Luciano’s full session where he deep dives into the many advantages of Rust including its focus on memory safety and performance, its utility as a general-purpose language and its provision of high-level constructs. He runs through examples using Cargo (a built-in package manager), and SAM (Serverless Application Model) and finally shows us an app he built using all of these tools to monitor earthquakes in his hometown in Italy.

AWS Cost Optimization – Maximizing your Effective Savings Rate with a Dynamic Discount Portfolio with Erik Carlin

Erik is the Co-Founder and Chief Product Officer of ProsperOps, a software service that uses AI, optimization algorithms, and a real-time execution engine to deliver maximum, hands-free savings. He brings his extensive experience to help hundreds of complex, at-scale customers optimize their costs on cloud.

What are some of the best ways to get rid of cloud waste according to Erik?

  • Erik runs through the main two avenues to cut down on cloud waste:
    • Usage optimization (using less ‘stuff’, rightsizing, transitioning to Fargate/Lambda).
    • Rate optimization (for the resources you are running, how to pay less for them).
  • Erik defines Effective Savings Rate (ESR) which is possibly the most important AWS KPI you aren’t tracking. ESR represents what you have actually saved using rate optimization.
  • ESR represents actual discounts you have achieved and takes into account coverage utilization and discount rate (prepaying, discounting Lambda/EC2, which region).

Erik explains how to increase and drive ESR including

  • Not compromising on performance and maintaining flexibility by avoiding overcommitment (such as only ‘investing’ in Convertible and Standard Reserve Instances (RIs) ).
  • How to achieve a blended commitment portfolio with a mix of instruments to generate compute discounts and manage risk.

Watch Erik’s full session where he runs through a ProsperOps real-life example to explain how one particularly company increased their ESR from 24% to >40% by reducing commitment, maintaining both high utilization rates and blended discount rate.

What’s new and upcoming in Cloud Computing Optimization, and how do you measure its Environmental Impact with Jean Latiere

We then brought on Jean, who opened up about life after a career at AWS (it exists!) Jean, with his many years of experience managing teams specialized in Fortune 500 consulting, software development, and cloud services -IaaS, PaaS, SaaS- has launched OptimNow because he believes that Cloud Optimization and CFM are most impactful and profitable when simple and fast.

Jean explains why it’s important to rely on innovation and third party tools when it comes to cloud optimization and sustainability. What does he suggest?

  • Jean recommends MemVerge for High Performance Computing (HPC) / High-throughput Computing (HTP) optimization. It uses unique patented technology called AppCapsule which captures all the data at once. You can successfully migrate from one VM to another VM during the runtime, monitor in real time and scale up or down with no interruption in workflow.
  • MemVerge will soon run on both AWS and GCP, so it’s perfect for multi cloud environments.
  • Another third party recommendation from Jean is Cloudscanner by Boavizta.org, an open-source solution that scans your cloud and lets you change settings and see in real-time how your workflows are impacting your cloud footprint (for free!).

See Jean’s full session where he also delves into what everyone seems to be talking about – Generative AI – and how it will impact not only FinOps, but GreenOps and how we manage our cloud resources in the future.

Top 3 (Simple) Best Practices to Minimize Your Cloud Waste according to Saif Kabishi and Jessica Eisenberg

We rounded out the event with our very own Saif and Jessica, two veterans of cloud backup and disaster recovery technology. Being on the N2WS team, which for over a decade has been the leader in optimizing recovery and DR in AWS and providing peace of mind to thousands of AWS customers, they were able to provide simple tricks and best practices to keep your cloud waste in check.

Jessica and Saif give these great tips to easily minimize cloud waste, particularly during this recent period of shrinking IT budgets.

Even if you are aware of how your cloud waste is spiking and what to do, Jessica points out that a bigger challenge is how to get every team on board to prevent future costs. If there’s a lack of awareness within an organization and lack of incentive, how do you make sure everyone is on the same page.?

  • Come with data and bring current spend reports.
  • Constantly develop KPIs and call out those who have succeeded in saving costs.
  • Implement gaming strategies: offer rewards, run a hackathons, etc.

Saif recommends requiring tagging and stopping any instance that is not tagged. This way anyone held accountable for creating a resource is also accountable for rightsizing it. Strictly follow global tagging policies so it’s easy to find resources.

Saif also recommends controlling who can provision resources using IAM roles in AWS. If you feel someone is over-permissioned, use Access Advisor.

And finally, identify and stop idle instances effectively. N2WS Resource Control gives you a leg up by providing an accessible ‘lightswitch’ (no permission needed in the AWS console to control resources). N2WS also lets your development team have access to power on/off your environment even outside a predetermined schedule.

See the full session where Saif runs through more advantages of Resource Control like having a fully visible and real-time cost savings calculator. He then demonstrates how to use N2WS Resource Control and delves into a recent case study of Gett Taxi where they were able to save over $100,000 simply by implementing the tool.

Excited as we are about the next event? We are already planning it… (but spoiler alert: we are currently undergoing a name change to focus on multi-cloud!)

Be sure to get all updates. Stay tuned by following us on LinkedIn.

New to N2WS? You can start backing up your AWS and Azure environment within minutes for absolutely free with N2WS Backup & Recovery 30-day Free Trial. Get access to enterprise features for 30 days including support for Amazon EC2, EBS, RDS, Redshift, Aurora, EFS, FSx, SAP Hana and DynamoDB. Trial automatically converts to our Free Edition + takes <14 minutes to launch, configure and backup. No credit card needed.

Next step

Cut compute costs with the flip of a switch

Allowed us to save over $1 million in the management of AWS EBS snapshots...

Try N2WS for Free