This year’s AWS New York City Summit took place on a sweltering day in Manhattan, but that didn’t scare away more than 6,000 cloud enthusiasts. The day kicked off with keynotes from Werner Vogels and customers, with strong themes of transitioning to serverless architectures and containers, peppered with a few minor product announcements. The keynotes, sessions, and expo hall were all packed — If this is any indication of what’s coming at re:Invent, I’m excited.
First, the news. Over the past few years, the NYC Summit served as the launching ground for new products and services, and was the venue where Werner announced some key price drops. This year was a little lighter (holding back for re:Invent?), but some key new announcements included:
- EBS Updates. Long overdue, in my opinion, the price of EBS Snapshots has come down by 47% to a reasonable $.05 per GB per month (retroactive to the beginning of August!). EBS Snapshots are key for most organization’s continuity plans for cloud-based infrastructure, so this will help companies keep costs down for their backups. Even though the price is now significantly lower, it’s still key to periodically go through and clean up old snapshots (SHAMELESS PLUG: You can write a policy to let CloudHealth do this for you). Additionally, for SSD EBS volumes (io1), there is a significant increase in performance-- up to 66% more IOPS per GB.
- Application Load Balancer. Imagine taking the capabilities of Elastic Load Balancer (ELB), but being able to scale down to the individual services level, and provide visibility into the health of these services. That's what Application Load Balancer for ELB can do. It also can support content based routing (up to 10 URL-based rules for now, more methods coming later), and container-based apps. Also, they didn’t mention this in the keynote, but a partner called my attention to the fact that Application Load Balancer is actually slightly less expensive than the Classic Load Balancer. Nice one!
- Kinesis Analytics is Generally Available. I’m not going to lie, I didn’t realize Kinesis was still in beta. With all the buzz about the managed service for real-time streaming data analytics -- and the fact that it was introduced almost 3 years ago -- it’s easy to forget that it was not GA yet. But now it is, yay! We have many large CloudHealth customers using Kinesis for various use-cases, it’s an incredibly powerful service.
“Test and dev is where agility lives”
Werner Vogels, CTO of Amazon Web Services, played host for the keynotes, which had a heavy focus on development and testing. “Test and dev is where agility lives,” Werner proclaimed, and went on to discuss how in the cloud you can test at any level of fidelity. Without the delay waiting for resources, developers can be as productive as possible. Werner also dove into the idea of reducing cost in test/dev environments running on AWS, a subject that is near and dear to my heart.
"If you can drive the cost of experiments to zero, you will do more of them" - Werner Vogels
Werner also called out that one of the best ways to reduce the cost of experiments to zero is to turn off test/dev assets when you go home. He called out that you can save 75% by doing this. The timing couldn’t have been better since we announced the same day that CloudHealth can now automate this process for you.
“There are no cattle, there is only the herd”
There is a common saying that in a cloud environment, you stop treating servers like pets and move towards treating servers like cattle. Werner took it one step further saying that in a microservices environment, there are no cattle, there is only the herd. With more and more serverless architecture options available from cloud providers like AWS, this concept can become a reality. But what’s next, there is no herd, only the shepherd? I’m fascinated as to where this is all going and what it means. Which is why the stories of AWS customers and how they are using microservices was the best part of the day for me.
For example, Chris Lambert, the CTO of Lyft took the stage to discuss how they scaled the ride sharing service to 24 new cities in less than 24 hours last year. They couldn’t have done this without services like DynamoDB, Redshift (which they used for launch their popular Lyft Line service), and Kinesis. In March of 2016, Lyft added their 100th microservice and serve more than 1 million requests per second. Not bad for a company that just turned four.
Another great story came from Abby Fuller, a Dev Ops Engineer at Airtime, who spoke about the company’s challenges with traffic spikes-- they experience ten times more traffic at peak times versus low times. The company originally had a monolithic server design, which created bottlenecks that prevented them from scaling when needed. So, they rebuilt the entire backend, moving over to a “cattle not pets” mentality and focusing on developer efficiency.
The last speaker was Sree Kotay, CTO and EVP of Comcast Cable… did you know that the X1 by XFINITY streaming service is also based on microservices, including DynamoDB? They are currently hosting 6,800 hours of olympics content online, which I intend to personally watch all of. I’ll see you in 283.3 days….
What did you think?
I’m just scratching the surface here of all the great content that was covered in the keynotes, and I haven’t even hit the breakouts yet! I’m curious what are your big takeaways from the NYC Summit? What are you looking forward to at re:Invent? Hit me up on Twitter to get the conversation going.