AWS continued to rock the technology world by revealing an array of new services at its annual re:Invent conference. With the release of new products and feature add-ons to existing properties, AWS shows no signs of slowing down.
If you’re looking for a list of the most impactful new AWS services, you’ve come to the right place. The following are the 7 coolest product announcements from re:Invent 2016.
- AWS Athena (new service) – Anyone who has used the S3 service knows of its data querying limitations. Prior to Athena, running rich queries was a complex process, having to build your own client-side database and breaking out all the relevant bits (like file type, file name, file content) you’d like to query. With Athena, AWS takes care of this process for you with a fast, hands-off, pay-as-you-go service. Athena uses Presto (one of the most popular tools to query cloud-based data) and works with many different types of formats.
- AWS Rekognition (new service) – Although AWS pretty much owns the public cloud, it’s long been far behind nearly everyone else (IBM, Google and Microsoft) in terms of machine learning. With the release of Rekognition (and three other new services, Lex, Polly and MXNet), AWS is finally putting a stake in the machine learning ground. Rekognition is an image recognition service that analyzes pictures to tell you details—detecting objects and scenes, while using facial recognition and analysis to identify people. But maybe the most exciting thing about the AWS launch into AI is CEO Andy Jassy’s statement that “more will be coming in 2017.” Expect AWS to finally catch up to the machine learning competition this year.
- AWS Shield (new service) – Shield is an AWS service built to holistically protect your AWS assets against a distributed denial-of-service (DDoS), one of the worst types of Internet attacks. Shield has pre-packaged and configurable protection rule sets you can enable across various types of back-end services, including your own custom-built EC2 or Lambda-based solutions. With the service, AWS is also providing access to the 24×7 DDoS Response Team (DRT), which can either assist during a live attack or preemptively help with rule creation.
- Amazon Lightsail (new service) – Need a cheap-as-possible, point-and-click managed web server for your business, kid’s soccer team or bowling club? Do you use GoDaddy, Digital Ocean or a similar provider to easily set up and manage things like DNS, monitoring and key management? Lightsail offers an SSH-accessible, easy-to-manage server—perfect for non-techies—in five different price points ranging from $5 to $80 a month. With Lightsail, you get all the power of EC2 with the ease of a GoDaddy-like interface.
- C# for Lambda (new feature) – Lambda (which offers serverless computing) was one of the hottest new services released at re:Invent in 2014, but with its limited language support (Python, Java, Node.js), the Microsoft crowd was left pretty much out in the cold. With the addition of C# support, Lambda now allows .NET developers access to the same pay-as-you-go, infinitely scalable, functional programming model that others have enjoyed for two years.
- AWS Batch (new service) – The scale of AWS, in terms of both raw number of services (now more than 60) and the features of each service (many have more than 100 available API calls), means you have unlimited options when processing big data or rendering images, video or audio. With unlimited options comes great confusion—which of the 80-plus instance types work best? How many do you need? Where should they output the data to? AWS Batch dynamically provisions the optimal quantity and type (like high-CPU or high-RAM) based on the volume and specific resource requirements of each batch job submitted. Like many other services including VPC, IAM and CloudFormation, Batch is free to use and you’re only charged the normal cost of the underlying resources.
- AWS Glue (new service) – Glue is a fully managed ETL service that simplifies and automates the difficult parts of discovering, transforming and moving data. It can connect to any JDBC-compliant data store (either AWS-based, or even on-premises), automatically logging into and crawling your schemas. It then suggests schemas and transformations (which you can edit, if you like). Once you accept the transformations, Glue goes about running the actual data flow job to move the data out of the source and into the sink. Like AWS Batch, it’s free to use and you’re only charged the cost of the underlying resources.
Other announcements of note: Snowmobile (an Exabyte-scale data transfer service), Elastic GPUs for EC2 (allows you to dynamically attach GPUs to current gen EC2 servers), F1 instances for EC2 (FPGA-based EC2 instances), Amazon Pinpoint (targeted push notifications—useful for personalized marketing campaigns), X-Ray (allows developers to more easily analyze and debug issues in their distributed systems) and Step Functions (provides a visual workflow interface to simplify the coordination of components in a microservices-based architecture). If you’re unfamiliar, it’s worth noting that FPGAs and GPUs are highly useful for machine learning workloads.
A full list of all service announcements from re:Invent can be found on AWS’ “Product Announcements” page. The launch into machine learning this year looks especially promising for those looking to build out ML platforms to support either their own applications or ones centered on Amazon Echo. If 2016 was “the year of IoT” for AWS, they’ve just set up the starting block for 2017’s “year of ML.”