What Happens in Vegas, not always stays in Vegas

This year the Noovolari team went to the AWS re:Invent in Las Vegas, and we want to share our experiences and thoughts on this re:Invent on the perspective of the Noovolari suite.

This year the Noovolari team went to the AWS re:Invent in Las Vegas, and we want to share our experiences and thoughts on this re:Invent on the perspective of the Noovolari suite.

It was an amazing journey! 

Announcements

We had 77 new products launch to play with, but we tried to make a list of what excited us the most. Ready? Go!

Machine learning, machine learning, machine learning…

With 20 announcements, Machine Learning wins the podium!

What we saw is that this year, AWS is focused on the Machine Learning topic, putting a lot of effort into extending the capabilities of it’s Amazon SageMaker: SageMaker Processing, SageMaker Experiments, SageMaker AutoPilot, SageMaker Debugger, SageMaker Model Monitor, SageMaker Notebooks, SageMaker Studio, and SageMaker Operators for Kubernetes.

Autopilot

There is nearly not enough time to go deep inside all the fantastic tools that AWS launched to provide us with better machine learning workflows. The only one I want to quote explicitly is the Autopilot service. This excellent tool lets you have machine learning models automatically generated from data! For people that are not data scientists, it’s a massive hit and ease of use as the first step in machine learning.

Aurora integration

Amazon Aurora can now be directly integrated inside AWS Sagemaker. It enables you to write SQL functions in queries to apply a machine learning model. Moreover, you can store the output of large queries, including the additional information from machine learning services, in a new table. Amazing!

And compute…

Sightly behind and earning the second rank with 16 announcements is the Compute topic. But with more significant improvements.

AWS Outpost

The Outpost announcement was significant in many different aspects. For whoever missed this, it’s a fully-managed compute, networking, and storage rack built with AWS-designed hardware that allows customers to run AWS services on-premises and is connected to the AWS public cloud. Great, isn’t it?

It’s astonishing as the price of the entry-level outpost (at a whopping 250.000$) we are sure that it will enable more and more legacy companies to migrate to a hybrid cloud solution. It may not be for everyone, but from a technical perspective, it opens up a lot of exciting scenarios!

ECS

And finishing on the containers topic, the ECS Auto Scaling, together with Capacity Providers and Fargate Spot, make using containers a first-class citizen on AWS and improves the reliability, scalability, and cost of running containerized workloads on ECS. Not bad at all for our container guys!

EBS direct APIs

AWS finally opened the APIs that leverage EBS snapshots! It enables a lot of more new integration with Noovolari Smart Backup. Stay tuned because we will work on this soon! ​

Serverless!!!

But the thing we are most excited about is the serverless part. As using a serverless paradigm consistently for the vast majority of workloads, we are happy to see some big news on this part!

Provisioned Concurrency

First of all, the Provisioned Concurrency clears the most annoying problem of using for AWS Lambda: cold starts!

​To understand this, we need to deep dive into how Lambda manages invocation environments: basically, the invocation is run in an execution environment that processes the request. When certain conditions are met (function not used for some time, a lot of concurrent invocations, or when you update the lambda code), the service automatically creates new execution environments. In the execution environment are installed the function code, and the runtime starts. Depending on the size of the deployment package and the initialization time of the runtime, new execution environments get some latency.

This feature enables us to have tighter control over our application performance, always leaving a certain number of execution environments and preventing them from hitting the cold start. Something we were looking forward to for specific workloads! But beware because it comes with fixed costs whether your functions will run or not, so be sure to weight down the benefits before applying it everywhere!

HTTP Rest API

Only in preview, but this one kept us very excited! API Gateway was notoriously verbose and hard to deal with for the number of resources that needed correct configuration, even a simple HTTP API. Now it’s possible to create a simple, single-resource API endpoint with minimal configuration.

Of course, it only routes requests in and passes them along to a backend to handle itself, but this was a widespread scenario, and we are glad that AWS simplified and reduced the cost of the most straightforward approach by 70%! Besides, this change extends the use cases with frameworks and tools that may prefer to set up routes themselves at the application level, without using the API Gateway configuration. Way to go!

Security

Amazon Detective

Amazon Detective is a tool to help security teams conduct faster and more effective investigations. The service automatically uses data from AWS CloudTrail and Amazon VPC Flow Logs to create a model that summarizes resource behaviors and interactions across the AWS environment. While collecting the logs, it uses machine learning and statistical analysis to deliver tailored visualizations to assist customers in detecting unusual behaviors. It’s a bit pricey but worth it!

IAM Analyzer

This tool is a little gem. It enables you to examine all the organization’s resource policies around AWS for various AWS resources and it determines all the different access methods allowed by the various policies. So security admins can review policies and ensure only intended public or cross-account access.

Moreover, it can create a report of all your publicly accessible AWS resources and also includes when a service was last accessed, by what users and when. With that information, it’s possible to reduce permissions to only what each user needs. And not only is it runs on demand, but it also provides continuous monitoring.

And the best part? It doesn’t use machine learning at all! IAM Access Analyzer team found a way to solve this problem only with math! What if you will find it integrated with Noovolari LookAuth soon? Let us know below and stay tuned!

Conclusion

That’s all for our trip to re:Invent! It’s not comprehensive, but we just wanted to share with you what are the topics on which we are most excited about and what will you eventually find in new Noovolari releases!

So stay tuned, and happy coding!


Notice: Undefined variable: post in /var/www/noovolari.com/wp/wp-content/themes/noovolari-eddie-website/functions.php on line 24

Notice: Trying to get property of non-object in /var/www/noovolari.com/wp/wp-content/themes/noovolari-eddie-website/functions.php on line 24