Last week was AWS Re:Invent in Las Vegas, their big annual event. Trek10 was lucky enough to have a contingent of three at the conference. Here are a few quick thoughts & notes:
Docker has massive momentum right now. AWS released the EC2 Container Service to manage Docker… but even more so, it was just the hot topic. Every Docker-related talk was jam packed, and everyone was discussing their Docker support. I personally spoke to several DevOps folks who had embraced Docker wholeheartedly with great success. Now seems like the right time for forward-looking dev & ops teams to seriously consider Docker adoption.
Aurora & Lambda are both in the running for the single biggest announcement, though we’ll have to see how each plays out they become generally available and start to add more features.
Aurora is a new relational database, built from the ground up by AWS in secret over the past 3 years. They claim it will have the robustness of the high-end commercial solutions at open-source pricing. It’s a big leap for AWS but just might pay off, as the relational DB is undoubtedly the least modern piece of most applications. Lambda is true “cloud computing”… you just upload a function and define event triggers… no provisioning compute or managing a virtual machine, OS, or middleware. Some others like Parse have already blazed this trail, but Lambda has the potential to be huge. They’ll need to come out with support for more languages (only node.js at the moment) and start supporting a lot of event trigger types (currently only a few).
On these two, my general advice is never be the first adopter for a new AWS service. Not because it is not robust… they are usually rock-solid… but because they release features so incrementally, the early versions are often missing some key ingredients to really make them viable for your use case. That is likely to be the case with both of these services initially. But check back in a few months.
Big data… It’s interesting that even though AWS didn’t release any new big data products, I came away from the conference with a strong impression of the things going on with Big Data with AWS. Several existing products… Redshift, Elastic MapReduce, Kinesis, and S3… are all maturing and being more widely used. The use cases out there and success stories are getting pretty compelling. AWS has definitely differentiated itself in big data, in my opinion… you can do things on AWS that you can do on no other platform, for a price and effort level that was completely impossible just a few years ago.
Enterprise features… AWS released a few new features aimed at enterprises… Config, Service Catalog, and Key Management. Lost in that is a key point, though… Config is actually an awesome new feature for just about every AWS user. Being able to easily view a complete history is a pretty fundamentally powerful feature. Every AWS user should enable it on their account… never know when you’ll need it!