Another reInvent has come and gone in Las Vegas, and with it came a range of new products for Amazon Web Services. Here’s what you might have missed.
AWS Reinvent main announcements
AWS Lightsail - Lightsail is an new automated way to setup a Virtual Private Server (VPS). This compute, storage, and networking, all wrapped together via one 2-step wizard. This is simplifying the process of setting up a functional server in three easy steps. Just need a server that runs WordPress? A server that already has Node preconfigured so you can just install your application and get it live? When AWS was originally introduced, it was built towards DIY engineers, but now AWS is looking to simplify and remove extra steps for people that don’t want them. When every company is a tech company now, every user isn’t interested in getting too technical. Think about it this way: some people want to go to IKEA and get their furniture in pieces, and some just want it delivered prebuilt. Lightsail is bent towards web developers and designers that just need a platform they can build on without worrying about infrastructure.
AWS Batch - An automated way for scientists to run hundreds of thousands of batch programs jobs.
5 New Instance types were announced along with attachable Elastic GPUs - an easy (relatively) way to boost graphics on an instance when you need it.
- The C5 are powerful Compute Optimized instances that are the most affordable for their level of performance in the EC2 ecosystem.
- The F1 instance type enables users to access FPGA programmable hardware for application acceleration. FPGA, by the way, just means that circuits that are designed to be manipulated after manufacturing.
- I3 instances are built for heavy workloads, like transaction processing, noSQL queries, and analytics.
- R4 instances are Memory Optimized, 20% more than R3 instances. Which is neat.
- Lastly, there’s two new large T2 instances which are consistent performers that are also built for burstable performance.
AWS Snowball Edge - Massive data transfer service at the petabyte scale.If you’ve used Amazon Snowball, this is a step up.
AWS Greengrass - Compute, data, storage for IoT connected devices. Greengrass gives you the ability to run IoT applications across the cloud. Essentially all of the core AWS products and services built for IoT devices. If you’re a company that builds connected devices (like Juno with ovens, or thermostats with Nest), this gives you the ability to manage products across the board. On a smaller scale, if you’re a retailer using Beacons across dozens of stores, this is another way for developers to keep track of all IoT devices.
AWS Athena - A simpler way to run queries on data hosted on S3. It’s serverless, so you’re only charged for the queries you run.
AWS Glue - Still a work in progress, but Glue makes it easier to transfer information between data stores.
VMware on AWS Cloud - VMware workloads on AWS clouds. It’s still in Technology Preview, but here’s the rundown: most enterprises, big or small, usually have on-premises infrastructure running VMware. As much as the cloud is growing, there’s typically always been a host of on-site servers running VMs.
This is designed to seamlessly integrate between those on premise infrastructures and AWS. The biggest benefit? You can deploy native VMware environments on AWS, and use all of Amazon’s services like leveraging better latency on databases with DynamoDB or Aurora. This deal has been directly brokered with VMware, so if your organization is deeply nested with VMware, all of your licenses and investments won’t be lost. Plus, if you’re deciding to completely switch over instead of co-existing, this works seamlessly with vSphere vMotion. Your VMs keep network identity and connections, so the migration is straightforward.There’s a lot to unpack here and we’ll be going in further detail when it’s widely released. Be on the lookout for an in-depth analysis blog post this month.
AWS Snowmobile - The data truck, the biggest and most visual thing everyone’s talking about. It’s an exabyte-scale data transfer service that transports your data via a massive, secure truck. The reasoning is that even if you have the fastest data transfer service over the internet, when you’re sending over thousands on thousands of terabytes, it’s just faster to download it locally and drive it across the country to a data center.
AWS Personal Health - Personalized view of health for all of your AWS servers. Users and companies can receive alerts and guidance (ala Azure Advisor). You can observe database failures, get billing and services notifications, view scheduled changes, and get a broad view of issues and events happening in your AWS ecosystem.
AWS OpsWorks for Chef Automate - Fully managed Chef master server along with automated testing, compliance, and automation for your servers in a user interface on the AWS Console. Chef’s web UI isn’t built for complex tasks, so AWS is taking on the task in a bigger way. Chef updates, configurations, and backups are all handled via AWS. And you can do all of this through the AWS Console instead of slogging it through Chef documents.
The benefit here is, as with most AWS products, that you don’t have to go through the complicated process of setting up and configuring a Chef server: this master server is preconfigured and ready to go. If you’ve never used Chef here’s the sales pitch: configuration automation for the infrastructure on all your servers, dictated from one central server.
AWS Organizations - Enables users to configure multiple AWS accounts. This is management at scale for enterprises and even development firms that run servers for multiple clients. Organizations gives users the ability to automate new account creation, centralize billing (no more juggling emails, accounts, and passwords just for different consoles), and control IAM policies and user permissions at scale.
EC2 Systems Manager - Systems Manager lets you apply OS patches, automates image creation, and helps you update applications at scale. While Systems Manager doesn't change the self-service end-user experience or the reactive nature of AWS policies, it offers some great automation options for admins.
Blox - Blox is an open source project that enables users to track clusters of containers (be they Docker or otherwise). As a third party scheduler, one of its primary features is the ability to track the state of these clusters via an event stream, and respond via API, logging data to your AWS Console. Blox also enables you to run tasks in each container in a cluster.
AWS CodeBuild - You can test and build code in the cloud. Why is this important? As every developer has experienced, the code on your local machine doesn’t always work on staging, much less production environments. In large enterprises, it isn’t always convenient to take up resources to experiment with new builds. CodeBuild is attempting to fix that with cheap resources you can use to experiment with that don’t involve configuring fleets of instances. While that isn’t necessarily enough motivation to try it, it does integrate with Elastic Beanstalk, so if you already have staging and production environments configured, this is worth playing with.
X-Ray - Debugging for production servers. If you have distributed applications or use microservices, X-Ray is useful. By adding an message handler to your code, you can use X-Ray to track all incoming and outgoing requests from your servers. Through the X-Ray console, you can visualize when and which requests happen, and on what servers. One example: if you use Stripe for payments, and your servers are having trouble processing payments, this is one way to quickly identify where those requests are failing.
Amazon Shield - Protects against DDoS attacks using Elastic Load Balancing (ELB), Amazon CloudFront, and Amazon Route 53.
Amazon Polly - Amazon Polly turns text into natural sounding speech. When you send the Polly API a request, it’ll convert the text to speech and relay it back to your application. If you’d like it’ll also save a copy of the audio stream in an S3 bucket. Polly processes the text as a stream, meaning that you get a response faster - it doesn’t need to synthesize an entire paragraph at once and then send it back to you, cutting out awkward delays. Polly makes use of the AI Amazon has been working on over the last few years to deliver natural sounding language in over 47 languages. If you’ve always wanted an application that can speak in English, Icelandic, Brazilian Portuguese, and Danish, you’re in the right place. Why would you need text to speech? Call centers, for one, can use Polly to express complex, on-the-fly information that a stilted auto-responder can’t handle. Or if you’re building IoT products, you can enable more natural interactions with users and your products. Imagine your oven announcing that the potatoes on tray one are done, but the ribs need another twenty minutes.
Amazon Lex - This is a conversational user interface and SDK for developers that processes natural language. This is the other side to Polly - speech to text powered by the same deep learning language processing that Alexa uses. If you haven’t used Alexa, it’s a heavy step in the next direction of tech - the end of the visual user interface and the rise of the conversational one. Alexa has thousands of skills - from asking it to play music on Spotify, ordering Ubers, and ordering new products for your kitchen, right from your kitchen. Amazon is opening the gates to other companies to build their own conversational UIs. While they own the home, imagine a streamlined DMV or a voice operated McDonald's drive-through. The possibilities are endless. Now Mark Zuckerberg’s 2016 plan to have a conversational UI running his entire house doesn’t seem so crazy.
Amazon Rekognition - Over the last few years, we’ve been seeing a massive push by Google in image analysis: the ability to identify separate elements of a photo (like the sky in the background, a car in the foreground, different people’s faces). This is Amazon’s response, with the ability to detect objects, scenes and faces powered by Amazon’s deep learning neural network. The benefit here is a form of visual search and discovery. Rekognition can determine whether the different elements of a person’s face mean if they are feeling a certain emotion. i.e. are they smiling, are their eyes open, are their eyebrows raised? It can also determine facial comparisons. It serves all of this data as metadata on each photo, and gives you the ability to process thousands of photos and mine information on large sets of photos. Imagine - scanning thousands and photos and videos at a press conference and determining who was there, and what they didn’t like when it was said. Dictatorships powered by technology are in our close future.
Amazon Pinpoint - Marketing campaigns for push notifications for mobile devices (both Android & iOS). You can define target segments, change messages, and view results. You can build those target segments through data from RedShift or S3, or Salesforce if you have the integration configured. This is only ideal if you’re already using a host of Amazon services to host the backend for your mobile applications.
And that’s it for everything that happened during reInvent this year! Once again, Amazon is positioning itself as the platform and the tools for a new era of tech companies. Coming later this month on the Scalr Blog, we’ll have the latest cloud news, and a in-depth look at the VMware/AWS Cloud offering.