AWS This Week

AWS This Week: AWS DeepRacer quota management, AWS Mainframe Modernization in GA

Episode description

Scott is back with your AWS news! This week, AWS Mainframe Modernization lets you migrate all those tens of thousands of lines of PL/1 or COBOL code in your mainframe, and Amazon Lookout for Metrics enables historical anomaly detection. Also, AWS DeepRacer gets an update to help curb large ML bills.

Introduction to AWS latest updates (0:00)
AWS Mainframe Modernization now GA (0:25)
Amazon Lookout for Metrics to Work Faster (1:53)
DeepRacer gets guardrails (3:04)
Miso Robotics to use AWS RoboMaker (4:02)

Sign up for a free A Cloud Guru plan to get access to free courses, quizzes, learning paths, and web series
https://bit.ly/AWSThisWeek

Subscribe to A Cloud Guru for AWS latest updates and service announcements, every week
https://www.youtube.com/channel/UCp8lLM2JP_1pv6E0NQ38pqw/?sub_confirmation=1

Like us on Facebook!
https://www.facebook.com/acloudguru

Follow us on Twitter!
https://twitter.com/acloudguru

Join the conversation on Discord!
http://discord.gg/acloudguru

Series description

Join our ACG hosts as they recap the most important developments in the AWS world from the past week. Keeping up with ever-changing world of cloud can be difficult, so let us do the hard work sifting through announcements to bring you the best of what's new with AWS This Week.

Okay. Hello Cloud Gurus, how y'all doing? Scott Pletcher here with another carefully curated set of AWS news suitable for framing. This week, AWS continues to build bridges to mainframe shops, AWS Lookout for metrics gets a rear view mirror, and our little buddy DeepRacer gets some guardrails. This is AWS this week. How often have you been sitting around the conference room and thought, 'Hey, let's migrate our mainframe to AWS. Just fire up our 3270 terminals, bring everybody outta retirement and start copy pasting'? If only it were that easy, right? Announced back at re:Invent 2021, AWS Mainframe Modernization service is now generally available.

But what in the world does AWS know about mainframes? Well through acquisitions and partnerships with companies who specialize in mainframe migration, AWS's assembled a pretty decent crew. Blue Age, which AWS acquired last year, are the people who brought us the ability to run COBOL code on Lambda. AWS has also partnered with Microfocus who offer a suite of mainframe migration tools, but tools alone won't get the job done. Most mainframe shops have decades of business logic embedded in tens of thousands of lines of COBOL or PL/1 code. We're not exactly talking a simple lift and shift here, but a major replatforming effort. To this end,

AWS has also established a Migration Acceleration Program for Mainframe and created a Mainframe Modernization Partner Competency. Come on our mainframes really still a thing? Absolutely. Especially in the banking, insurance, and healthcare industries. According to Reuters, about 80% of the in-person commerce transactions and 95% of ATM transactions still rely on COBOL code running on mainframes. The Amazon Lookout family of services are a perfect example of taking some complicated machine learning processes and abstracting them into easily accessible real world value add turnkey services. At the core,

Amazon Lookout services use anomaly detection algorithms to identify things that are out of the ordinary and bring those things to your attention. Amazon lookout for equipment can pick up on subtle changes in equipment performance, and allow for more proactive maintenance. Lookout for Vision can perform visual inspections at scale and help with quality assurance. Amazon Lookout for Metrics can provide that same anomaly detection across a variety of AWS and third party services. Recently, AWS announced that Amazon Lookout for Metrics can now backtest on CloudWatch.

Previously, we could only check the anomalies in CloudWatch from the point at which we initiated coverage forward. This backtesting mode allows Amazon Lookout for Metrics to look back in time, to create a more accurate baseline. And that's real useful for organizations that have seasonal fluctuations or other ebbs and flows that aren't necessarily anomalies. Unintended consequences - we've probably all had our share of those things. For example, let's say you launch a DeepRacer league within your organization as a fun way to increase employee engagement and learn some new skills.

Let's say your league takes off and gets super competitive between the folks over on the application development team and the network engineering team. Let's say at the end of the month, you get a panicked call from your accounts payable people asking why your AWS bill is three times higher than it's ever been. I get it. It happens fortunately, you now have some guardrails to apply to your DeepRacer dabbling, if you'd like. AWS DeepRacer multi-mode now allows us to apply quotas to our participants, reigning in the number of training hours, model counts, and even turn on and turn off the ability to train. No doubt this feature originated from organizations who likely found themselves in the same place as our completely hypothetical example.

And before we go, one last little tidbit, you may soon be thanking AWS for improving the quality of your fast food french fries. Miso Robotics, the company behind Flippy, a robotic fry station, has announced a partnership with AWS to use AWS RoboMaker to vastly accelerate their simulation and development processes. The Flippy 2 fry robots have already been ordered by several quick-service restaurant chains as they prefer to be called, including my own guilty indulgence of choice, White Castle. I, for one, welcome our new fry robot overlords and thank them in advance for their tireless service. Thank you as well for joining me this week, as that's just about all the AWS news it's fit to print, stay safe, take care of one another and keep being awesome Cloud Gurus.

More videos in this series

Master the Cloud with ACG

Sign In
Welcome Back!

Psst…this one if you’ve been moved to ACG!

Get Started
Who’s going to be learning?