Wednesday couldn’t arrive fast enough for our team, as the first wave of new product features and capabilities are typically announced on the third day of AWS re:Invent. This year satisfied the tradition with an array of announcements and exciting new features. Andy Jassy was the keynote speaker, and he certainly lived up to the expectations of the ‘geeked up’ crowd. In case you missed it, the announcements were so big this year that they required an 18-wheeler to be driven onto the stage.
A summary of each new product feature is below.
New Compute Family Options
T2.Xlarge and T2.2Xlarge instances (16 and 32 GB RAM)
We see customers choosing to run T2 family instances for many general-purpose workloads where scalability is possible, however, sometimes engineers must choose to use M4 instance family because the T2 family did not provide enough memory. The addition of these two new T2 instance sizes will allow even more customers to run instances in the cost effective T2 family when applications require 16 or more GB of RAM.
R4 Family (The upgrade of the previous generation R3 family)
This new family of instances is an upgrade to the R3 family of instances, designed for high performance workloads and data processing. Providing high end compute and massive amounts of high performance RAM all at a lower price per GB than the previous generation.
C5 Family (The upgrade of the previous generation C4 family)
A new family of compute-optimized instances that are built on the new Skylake Xeon processors. This upgrade features new hardware acceleration that promises to pack more power for compute intensive workloads.
I3 Family (The upgrade of the previous generation I2 family)
The upgrade to the I2 family of instances used for workloads that require highly intensive input/output workloads. This upgrade features more memory and compute than previous generations, as well as a new lower latency SSD capability through non-volatile memory express.
The F1 family is a brand-new family of instances that provides FPGA hardware acceleration. For more information on FPGA acceleration develop, you can view the Wikipedia page.
AWS made the G family of instances available a few years ago. The ability to run workloads that require graphics accelerated hardware in the cloud. This family helped many customers who needed to stream 3D graphics or encode videos.
This capability will allow any instance to become a node for graphics accelerated workloads. Now customers will be able to use a C5 instance to do graphics intensive work. Maybe even play a graphics intensive game on an AWS EC2 instance.
Quick and Easy Standup of Secure Dev/test Environments
Amazon Lightsail will allow developers to stand up a single, isolated, Virtual Private Server (VPS) to do development and testing of applications, without having to define VPC and address connectivity security concerns. This service will empower rapid development of applications in a truly fail fast environment.
Most data analytics processes put a massive amount of data into S3 before using Elastic MapReduce to transform that data and analyze the data, then stick that data into RedShift so it can be searched and visualized. But, what if you could just search that data on an ad-hoc basis while the data was sitting in S3? Athena delivers this capability. We look forward to seeing how customers leverage Athena to more quickly make business decisions.
Anyone who owns an Amazon Echo is an end user of AWS’ machine learning capabilities. The services that back Amazon Echo are now available for consumption as well as a couple of new services that will provide customers value.
Lex is the service that provides response to audio input. This service will allow customers to trigger Lambda functions from a voice command. Imagine a development pipeline where we can ask Alexa to, “deploy application ‘Web Login’ to production,” or, “add 6 instances to the web application.”
Polly is a new text to speech service. Type a bit of text in a document, submit it to the service API, and you will receive a mp3 with the audio of the document. The initial offering will support 24 languages and 27 voices. Get your favorite books ready for that long road trip.
Rekognition enables image recognition. Some use cases are facial recognition, face matching, and image mapping of items in images for differences.
Massive Data Migration
Last year, AWS announced Snowball. Snowball allows businesses to use an Amazon provided storage device to load up to 100 TB of data and ship it to AWS faculties for import. Snowball has evolved this year in 2 ways:
Snowball Edge will combine the power of the original snowball with the additional capabilities of built in compute and the ability to cluster snowballs for additional density. We really like this addition to the service because it demonstrates how we rarely anticipate all the ways customers will consume an offering. Snowball is now being used to store local data in smaller subsets for immediate analysis, then shipped to AWS and imported when it fills up. The collective data from all the snowballs over time can then be used for long-term analysis.
Do you have more than 100 petabytes of data? Even using Snowball, it would take you approximately 10 to 25 years to get all of your data imported into AWS. Enter a tractor-trailer to the stage, now known as Snowmobile. This service offering will allow customers to migrate up to 100 petabytes of data in a single shipment. For an exabyte data migration, it would take approximately 6 months rather than the 25 years before snowmobile was available.
What are you most excited about from yesterday’s announcements? Stop by Booth 2840 to meet our team and chat about the exciting updates!
Always be in the Know, Subscribe to the Relus Cloud Blog!
- Maturity Models and Metrics: Part Two - December 14, 2016
- Relus Cloud at AWS re:Invent 2016 – Day Four - December 2, 2016
- Relus Cloud at AWS re:Invent 2016 – Day Three - December 1, 2016
- Relus Cloud at AWS re:Invent 2016 – Day Two - November 30, 2016
- Relus Cloud at AWS re:Invent 2016 – Day One - November 29, 2016