AWS seems to be experimenting with this summit to try and find the right combination of location and duration—after last year’s summit in Santa Clara, this year the summit was back in the city (yay!) and went for two days instead of one.
The staple of any AWS summit is the Werner Vogels’ keynote—always chock full of customer stories and new features. Werner’s talk echoed the theme Andy Jassy’s keynote had at re:Invent: superpowers. Putting his own spin on it, Werner said “AWS customers feel like they are James Bond and AWS is their Q.” James Bond isn’t actually a superhero, but I’ll let that one go.
The announcements
During his keynote, Werner began with the typical spiel highlighting growth (always impressive), logos (I spotted a few new ones in there!) and overall market dominance (yes, AWS is still the fastest growing tech company with more than $1B in revenue). After that, he launched into a stream of announcements, punctuated with some nice customer stories, which I outline below. At one point Werner launched into an overview of how containers enabled the rise of microservices, and highlighted some cool customer use cases. The crowd clearly felt this was building to a big announcement about microservices or containers… which unfortunately never came. Was there something planned that was pulled at the last minute? We’ll have to keep an eye out for container or microservices news in the coming weeks. That being said, an exciting array of new tools and features were announced at the Summit, including these highlights:
AWS CodeStar
This new service makes it easier to develop, build, and deploy applications with the project templates and resources you need to get your latest software project up and running. CodeStar offers a unified user experience for the entire team, including a project management dashboard to keep everyone up to speed on the various tasks underway. Here’s the best part: CodeStar is free. You only pay for the AWS resources you use in CodeStar projects. This could be hugely disruptive.
Amazon RedShift Spectrum
This new service extends the analytical capabilities of Amazon Redshift SQL queries to the unstructured data in your Amazon S3 “data lake.” This removes a huge barrier to entry for customers who want to do analytics on their data, but could not overcome the challenges and expenses associated with moving the data. Redshift Spectrum gives you easy access to the data you need—regardless of where it’s stored or what format it’s in. Redshift Spectrum is priced based on the volume of data you are analyzing ($5 per TB), and they do allow you to compress your data before querying, which can significantly drive down that price. This announcement continues the theme of “analyze your data wherever it lives” which is a pretty big deal, in my opinion. This goes hand in hand with the Athena announcement from re:Invent last year.
SaaS contracts in AWS Marketplace
According to Forrester Research, the average number of different SaaS solutions used by an organization is 66. That means you are likely dealing with 66 different bills every month. If you suffer from SaaS bill exhaustion, AWS just made things a lot easier with the introduction of one-, two-, and three-year SaaS contracts in AWS, and you have the flexibility to upgrade the size or duration of your contract at any time. When you purchase SaaS contracts on the AWS marketplace, your SaaS bill is rolled into your AWS bill as a line item, how easy is that? Twenty SaaS vendors are already featured on AWS Marketplace—including CloudHealth!
NextDoor use case
Not an announcement, but an interesting use case was NextDoor. Prakash Janakiraman, Co-Founder and Chief Architect of NextDoor, the first social network designed to connect users with their neighbors, joined in as a guest speaker to discuss how his company uses ECS. Founded 7 years ago, NextDoor is now used in 70% of neighborhoods in the U.S., and has run on AWS since day 1. Today, NextDoor uses nearly 30 AWS services. By moving to ECS, they reduced their build time went from 25 min to 2-3 minutes and their deploy time from 30 minutes to 4-5 minutes. They now have dozens of deployments a day using continuous deployment and have overall reduced the time from build and deployment by nearly 10x.
