Day 1 summary here.
Day 2 of re:Invent was awesome. AWS delivered exciting new products. I still think that AWS understands customer pain points better than any other tech vendor on the planet. Day 2 keynote shook up the racket known as tool vendors ( not all are bad, but many just market old tech as new ).
Here is my reaction to new announcements:
- X-Ray : The idea that I can get performance monitoring right from AWS is awesome. I don’t have to configure agents. I won’t be told daily that ‘Your subscription level does not give you access to this feature’. I think using X-ray as step 1 is no brainer for many startups and even for some enterprises. While large enterprises may not move to x-ray right way, it does allow their devs and ops to use it for new workloads.
- CodeBuild: Have you ever woke up at 2 am by a slack message that build server is failing? You get on the phone with your sys admin, who figures out that one of the slave servers is running out of disk space. You check your calendar and realize it’s 2016, not 1992 and wonder why the hell am I hosting my own build server? So, it’s a no-brainer for me to start using Code Build as early as next week. This may not be the case for large enterprises, but is having a team of build engineers optimal use of your investment?
- AWS Shield — Meh. I kinda expected AWS to be doing this already, so its cool, you are late, but thanks for doing this finally.
- Pinpoint — This could be exciting for those building apps for consumers, so I have no reaction to this. I am not yet building a location based coffee exchange app.
- AWS Batch — This is awesome. I remember taking angry calls from customers while at an enterprise vendor on the last day of the month as the software is choking under heavy load. We called this end of the month problem and it was a popular time for support folks to take a sick leave. For this, we often told customers to quadruple their hardware. That was one way to solve it, but expensive. Cloud allowed customers to solve this problem better, but AWS Batch takes this one step further. Because you don’t have to reserve instances ahead of time, it cuts down on work that customers have to do.
- Personal Health Dashboard — this is cool for large teams, I don’t have either positive or negative reaction to this.
- Blox — well, there were people expecting AWS to embrace Kubernetes and play second fiddle to Google. I told folks that the chances of it happening are non-existent. Blox puts a damper on all container orchestration players. It adds one more to the mix of a highly confusing and over populated container orchestration market. I view this as a “neutralize Google” step — this is bad news for Google, as they are now required to update the strategy presentations they have been using since 2014. My sympathies with them, I know the pain of updating powerpoint slides. I am doubtful if any customers would ever actually use Blox though.
- Step Functions — Exciting. I saw this as a next logical step — I called it a dev studio, but Step Functions is a way cooler name. This is a natrual evolution of Lambda, but not enough. A state machine is step 1, there need to be higher level constructs — almost like a new programming language to define this flow logic. I often compare Lambda functions to assembly language coding. It works, but it becomes painful over time. Just as we got to higher level languages, there is a similar need here.
- IPV6 support : I like many others still don’t understand IPv6. Will refrain from saying anything that shows my utter lack of understanding of computers.
It’s great to see AWS continue to evolve to meet customer needs. The innovation machine hasn’t slowed down a bit.