AWS re:Invent is the biggest cloud computing community event in the world. Last week, Las Vegas was a place where AWS brought together over 50k professionals eager to learn, network and share their experiences. 4 of us from Semantive had an immense pleasure to be a part of the event and want to share some of our views on this year’s AWS announcements.
With the conference getting bigger and bigger each year and new services and enhancements being introduced regularly, one may think that there is nothing innovative left that can impress us. However, as usual, AWS didn’t disappoint us and made a lot of significant announcements. We are going to highlight some of the biggest and most interesting ones that are the closest to our professional interests and problems we are solving in our day-to-day work.
Amazon Forecast is a fully managed service based on the same technology that is used at amazon.com. It combines both deep learning and traditional statistical methods and can provide a high accuracy by inferring the best algorithm for the data it’s dealing with. We are looking forward to comparing Amazon’s solution with the ones that we have built for our clients over the years!
Recommendations and predictions are one of most common problems that are being solved by machine learning. We can find them almost everywhere, from online shops to streaming services. Similarly, to Amazon Forecast, Amazon Personalize is based on the same technology that is used at amazon.com. That being said, we can expect a potent tool.
Amazon SageMaker Ground Truth
High-quality training data is crucial for any machine learning problem. We all know how hard it is to have perfectly labeled datasets. Amazon SageMaker Ground Truth is designed to reduce the amount of effort needed to create training data as it can use machine learning to automatically label data.
Amazon SageMaker RL
Reinforcement learning is an extremely hot topic in the ML world. No wonder that AWS has decided to expand SageMaker capabilities with RL algorithms support.
Amazon Marketplace for Machine Learning
Starting now, AWS offers a marketplace for machine learning algorithms. Algorithms and ML packages can be deployed directly on SageMaker and are provided in free, trial and paid tiers. We firmly believe that it is a great opportunity for many companies to share and sell their models and ML solutions.
DynamoDB on demand
This feature has been highly anticipated by anyone who is dealing with unpredictable workloads (we know that problem too). Now, with DynamoDB being able to automatically adapt to a workload level, there is no need for capacity planning and managing. You pay for what you use.
Timestream is a scalable (it’s is said to handle trillions of events daily) and fully managed time series database with built-in analytics capabilities such as interpolation and smoothing. This can be a perfect fit for fast growing IoT businesses, DevOps use cases and industrial telemetry.
AWS Control Tower
Control Tower is an AWS Landing Zone solution available as a service where the Landing Zone is a framework to set up a multi-account baseline for your organization. It is done according to AWS best practices which lay the foundations for further development. For example, it creates a Log Archive aggregation account for you and is locked-in with no access for anyone to which all other accounts send logs from CloudTrail so that whenever you need it you can audit them. We learned from one of the sessions that hackers mining for Bitcoin, in order not to get caught until the AWS bill arrived, were able to hijack an account and remove logs with CloudTrail. It would not have been possible if the logs account had been in place with proper security guardrails.
AWS Lake Formation
Lake Formation is a service to set up a Data Lake on AWS in a couple of days. It might be an excellent product for a Data Lake proof-of-concept. Afterwards, a customer can decide to further develop on top of the Lake Formation or go for a custom solution tailored to his needs. Upon evaluating this solution in more depth, we can decide whether it goes towards our toolbox.
As you can see machine learning and AI were one of the most prominent themes highlighted during Andy Jassy’s (CEO of Amazon Web Services) keynote. Those topics are bread and butter for us here at Semantive. Therefore, we can’t wait to use the new services in our projects!
We are incredibly excited about new AWS offering, how about you? We would love to hear your thoughts!
Written by Anna Stępień, Big Data Engineer at Semantive.