Flux7 Blog

Andy Jassy, AWS CEO, mentioned in his keynote talk that AWS released about 2,500 new features and services last year. Not to be outdone, anchoring the vast number of new solutions AWS announced this year is a long list from re:Invent — and that’s only after two days.  While we’ve rounded up all the announcements in one place for you here, two big takeaways emerged from the first two keynotes as well.

Specifically, Andy Jassy’s keynote along with the Monday Night Live talk hosted by Peter Desantis, AWS VP Global Infrastructure and Customer Support, illustrates the big bet AWS is placing on the combination of High Performance Computing (HPC) and Artificial Intelligence/Machine Learning (AI/ML). As shown through new services unveiled during both talks, AWS continues to make it easier and easier for people to adopt ML without a bench of data scientists sitting at the ready. The spate of new announcements at re:Invent thus far strengthens an already strong portfolio here.

The second particularly interesting development this week was the announcement of Local Zones. Putting resources closer to “home” for companies doing hybrid is going to be very helpful in terms of having latency-sensitive workloads be located closer to a home base. While we think that the initial location in Los Angeles is a good start, AWS will definitely need to add more Local Zones as time goes on. 

Easily build and train cloud machine learning models

Andy Jassy rightfully points out that there are not a plethora of ML experts simply sitting around; therefore, to make ML more expansive, you have to make it more accessible. AWS began this journey with SageMaker (which it announced a couple of years ago) and continued on that path at re:Invent with several SageMaker announcements, the most notable of which was Amazon SageMaker Studio, what AWS touts as the first end-to-end Integrated development environment for ML. It allows developers to build, train, tune, and deploy models from a single interface; organize, search, and access notebooks, datasets, code and settings; create project folders to organize all of their ML projects; share projects, folders and content; discuss notebooks and their results collaboratively and more. Amazon SageMaker Studio is available today in US East.

In addition, AWS announced:

• SageMaker Notebooks
  This new notebook experience allows developers to quickly create ML notebooks which can be shared with a single click. Available via SageMaker Studio, SageMaker Notebooks are in preview in US East.

• • SageMaker Experiments
    This new SageMaker tool lets developers easily organize, track, compare and evaluate ML experiments and model versions. It captures input parameters, configurations and results automatically helping developers actively keep track of the many ML models saved by SageMaker studio as an experiment. Amazon SageMaker Experiments is available today in all commercial regions where Amazon SageMaker is available.

• • SageMaker Debugger
    On by default, SageMaker Debugger automatically collects and monitors key training metrics; provides real-time metrics of model training performance and; gives you feature prioritization for better interpretation of ML. Automatically identifying complex issues developing in your ML training jobs, its debugging results can be viewed through Sagemaker Notebooks. Amazon SageMaker Debugger is available today in all commercial regions where Amazon SageMaker is available.

• Amazon SageMaker Model Monitor  Detect concept drift in deployed models automatically with SageMaker Model Monitor. Further, it allows developers to create baseline statistics about data during training, allowing them to actively monitor models with a single click. Amazon SageMaker Model Monitor is available today in all commercial regions where Amazon SageMaker is available.

• Amazon SageMaker Autopilot
  Autopilot provides automatic training, automatically creating the best possible classification and regression machine learning models while allowing full control and visibility. Amazon SageMaker Autopilot is available today across a wide variety of commercial regions.

• Amazon SageMaker Operators for Kubernetes
  This new capability announced by DeSantis makes it easier for developers and data scientists using Kubernetes to train, tune, and deploy ML models in SageMaker. Install Operators on a Kubernetes cluster to create Amazon SageMaker jobs natively using the Kubernetes API and command-line Kubernetes tools such as ‘kubectl’. Amazon SageMaker Operators for Kubernetes are GA in the US East (Ohio), US East (N. Virginia), US West (Oregon), and EU (Ireland) AWS Regions.

AWS Machine Learning Use Cases

In addition to a host of SageMaker services, AWS announced several services that directly address specific ML use cases. Namely:

• Amazon Fraud Detector, a machine learning-driven fraud management service.
• Amazon Code Guru, a ML service that assesses code and gives developers human-readable comments, down to line-of-code issues.
• Contact Lens for Amazon Connect uses ML for contact center analytics, automatically transcribing and analyzing customer calls, sentiment, and more.
• Amazon Kendra lets you reimagine search with ML.
• And, AWS DeepComposer which allows you to compose Music with Generative Machine Learning Models.

Local Zones

As Amazon helps enterprises address hurdles to cloud migration, one issue they encounter is around workloads that can not — for one reason or another — be migrated to the cloud. For these workloads, AWS announced AWS Outposts last year. This week Andy Jassy announced that the service has reached GA, allowing on-premises workloads to use the same tools and services.

Additionally, to address highly-demanding applications that are particularly sensitive to latency, AWS announced a new type of infrastructure deployment — Local Zones. The goal is to bring select AWS services very close to a particular geographic area. The first area announces is Los Angeles, with others to be added shortly. Local Zones provide compute, storage and database services close to large cities. 

At Flux7, we too have a placed a wager on HPC, ML and the cloud. YOu can find more on our new HPC practice on our resource page. We’ll continue to track re:Invent news for you as well. Don’t miss it; sign up for the Flux7 blog today.

Written by Flux7 Labs

Flux7 is the only Sherpa on the DevOps journey that assesses, designs, and teaches while implementing a holistic solution for its enterprise customers, thus giving its clients the skills needed to manage and expand on the technology moving forward. Not a reseller or an MSP, Flux7 recommendations are 100% focused on customer requirements and creating the most efficient infrastructure possible that automates operations, streamlines and enhances development, and supports specific business goals.