Your AWS re:Invent 2017 Keynote re:Cap
re:Invent attendees know the excitement and joy around the two major Keynotes during the event each year. The 2017 keynotes revealed key services to help AWS users improve the way they collaborate and manage containers as well as how they dive into deep machine learning on the cloud. Let’s walk through a few major themes from the two keynotes around these exciting announcements.
Deploy At-Scale Containerization with AWS
AWS EKS is a new managed service that makes it easy for you to run Kubernetes on AWS without needing to install and operate your own clusters, and you can request a preview of EKS immediately. With AWS Fargate, users can run containers without having to manage servers or clusters. AWS Fargate is available immediately for folks using ECS, with support for EKS coming in 2018. Costs will be a bit higher than if you choose to manage the instances and clusters yourself – starting at $0.0506 per hour or $0.00001406 per second according to AWS docs.
With Great Power Comes a Few More Costs to Wrangle
Our engineering teams are working quickly to support Fargate in our existing reporting engine and expect progress through the end of the year and into Q1. We also anticipate a Beta release of our Kubernetes Cost allocation product in Q1, which will support folks that are using EKS as well, and will give you visibility into cost allocation by Cluster, Namespace, Label, and Service.
If you’re interested more in our Kubernetes cost allocation support, please ask us for a demo of the current Alpha.
Adapt at the Speed of the Cloud with AWS Machine Learning
AWS introduced SageMaker, its fully managed end-to-end machine learning service. It enables data scientists, engineers and machine learning experts to get machine learning models up and running quickly on the cloud. There’s a free tier available for teams wanting to get started. Pricing is broken down by building your model, training it and then hosting it. Please see the official AWS docs for more details.
Along with SageMaker, AWS announced DeepLens, a machine learning tool that uses video recognition. During the keynote, CEO Andy Jassy announced it as “the world’s first deep learning-enabled video camera for developers.” Teams wanting to get started with DeepLens will need to pre-order the camera and then configure it with their SageMaker application.
Expanding the Human Interface with Cloud
Werner Vogels, CTO of Amazon, ended his day two Keynote stating that voice will be a disruptive element of future software development. Supporting this announcement, AWS released Alexa for Business, the same voice command application many already know and love in their homes, but catered to the business or enterprise.
Getting started with Alexa for Business will require teams to build the proper Alexa Skills to support key workflows and day-to-day queries and interactions with software. Pricing is monthly, starting at $7 per shared device and $3 per user.
Along with Alexa for Business, AWS released their own transcription service, Amazon Transcribe, which is based on neutral machine translation and transcription. This allows developers to add speech-to-text capabilities to their applications and even includes an API for the calling of audio files (on S3) transformed into text. For teams wanting to try out Transcribe, there is a free tier available. Beyond that free tier, folks will be charged for Transcribe and also S3 costs based on how much audio is stored.
Enterprises that need to localize their messaging and content know the value of accurate translation. AWS announced their own service called Translate, which they position as fast, high-quality and affordable language translation. Using the same neural machine translation as Transcribe, Translate will help teams with large volumes of text at-scale to support localization of websites and applications globally. AWS has example pricing for translation on their documentation for review.
Improving Computing, Storage and Database Services
Outside of the deep machine learning and container news, we also have new EC2 instances and updates to current families. AWS announced the M5, H1 and T2 Unlimited instances. For teams that want to work with the direct accessibility of bare-metal servers, AWS also announced i3.metal. As we recommend in previous articles on our blog, take a close look at the new sizes, costs per vCPU (or actual hardware in the case of i3.metal) and memory and other hardware perks before migrating. We’ll be investigating these new instances throughout 2017 and into the new year.
NOTE: Cloudability’s True Cost Engine already has support for all of these new instances. Start a Trial today if you’re already running M5, H1 or T2 Unlimited instances to track their costs and usage moment by moment.
New Ways to Find Specific Bytes in S3 and Glacier
AWS announced two new capabilities for S3 and Glacier that allow users to perform simple SQL expressions to identify and pull out only the bytes you need from those objects. This greatly increases the accessibility of specific bytes of data from storage and reduces the need to retrieve entire volumes to search for objects, which can incur certain charges or overages.
Reduce RDS Costs with Aurora Serverless
AWS also announced a preview of Amazon Aurora Serverless. Now your engineering teams can user AWS Aurora without spinning up RDS and instead be charged by the query. According to launch documentation, scaling is rapid, with new resources coming online within five seconds. Pricing for Aurora Serverless is done by Aurora Capacity Units (ACUs) starting at $0.06 cents per hour using the example on the AWS docs.
NOTE: Our engineers anticipate that customers will start saving with Aurora Serverless as it means they’ll spend much less on maintaining the RDS servers to run Aurora.
There are many more updates and announcements, such as AWS Cloud9, IoT breakthroughs, security-focused updates, other machine learning announcements like GreenGrass and more. Get the real lowdown by re-watching the keynotes once available.
Get the Most out of AWS by Knowing Your True Cost
If you were at any of our sessions or visited our booth you learned about our new mission to help customers of all sizes find the True Cost of their cloud. Whether it’s using AWS’s compute, storage or database services and even scaling up machine learning and containerized projects, knowing their True Cost allows your teams to make better technology choices that help your business achieve its goals.
From our team to yours, we hope you all had a great re:Invent. If we didn’t get a chance to catch you in Vegas, feel free to check Cloudability out at any time.