AWS re:Invent 2017 Recap: A Successful Day One
It is one of the most anticipated tech conferences there is. Tens of thousands of cloud nerds scurry across the busy Las Vegas Boulevard. Celebrities like Shaq walk through the halls of the Venetian. Everybody and their sister can are at The Aria trying to get into a serverless session.
This is AWS re:Invent 2017 and today is the close of day one. This is only my second year at the conference. But, I learned so much last year that I want to make sure I share what I learn this year with all you.
If you are not familiar with re:Invent, here is a brief run down.
At re:Invent you will find a variety of talks. There is breakout sessions that focus on how to use AWS services to solve a spectrum of problems. There are hands on labs where you can convert your learning into real practice. Certifications are a hot thing here. Many want to complete them while they are soaking in all the knowledge.
Then of course there is the key notes. Attendees line up hours in advance tomorrow and Wednesday to get a seat. Why you might ask? Announcements that’s why. Everyone wants to know what the new hot thing is going to be. O and there is the re:Play party as well. Which is pretty much a rave packed with 30,000 cloud nerds. Free food and enough free alcohol to fill any data center in the world.
It is a fun conference to attend. You focus on AWS for an entire week and go back to office exploding with ideas. There is a lot of networking opportunities as well. I know of a few people who have formed long lasting bonds by having a beer together at re:Invent.
Enough introduction. Here is my recap of Day One at re:Invent 2017.
Day One: Landed, Lined Up, and Learned
Last year the conference had 30K+ attendees. Today it seemed like a lot fewer people, but then again I spent my entire day at the Venetian. I heard if you were at the Aria where the Serverless sessions were happening, it was a different story.
There was also rumors of 45+ minute shuttle rides to get from the Venetian to the Aria. Hearing that, I opted to stick out day one at the Venetian and take in as many talks as I could there. I managed to sit in on three breakout sessions in between talking with other folks in the expo area.
1 Million bids in 100ms using AWS to power your Real Time Bidder
This session from Beeswax focused on how they created their Real Time Bidder in AWS.
It was interesting to see two former Googlers giving a talk at an Amazon conference. More interesting was the amount of data they had to process on a given day to allow quick decisions.
To enable the performance they needed to leverage Network Load Balancers (NLB). NLB’s allowed them to optimize their network latency. Seeing as they have to provide real time bids, the difference between milliseconds and microseconds is important.
The take away I had here was actually around SSL termination. They also chose NLB’s because it allows SSL termination at the instance rather than the load balancer. I found this interesting as their reasoning for doing this was again performance.
The last thing I found interesting from the talk performed by Beeswax was their use of Scheduled Actions. Bidding windows are predictable at Beeswax. As such, they pre-warm instances in the Auto Scaling group via Scheduled Actions. They launch instances before their morning load comes in to enable quicker response.
Adopting Microservices in Healthcare: Building a Compliant DevOps Pipeline on Amazon ECS
This was one of those sessions that you thought was going to be something. But then was more of a sales pitch by an AWS Partner.
The general topic was that Healthcare has massive compliance obligations (i.e. HIPAA). Blue Cross Blue Shield of New Jersey talked a lot about how they were going from an on premise monolith to microservices in AWS. This was a journey to say the least. There was no lifting and shifting to make that happen for Blue Cross Blue Shield.
The biggest takeaway I had from this session was how different of a world a lot of companies live in. Blue Cross Blue Shield had to make very precise decisions when deciding what work loads to move to AWS. There were things they wanted to move, but couldn’t do to a lack of HIPAA compliant services.
What I found most interesting was how they wrapped existing on premise services. They implemented a strangler type pattern to keep some services contained in their own data centers. Exposing those services via a clean interface. This allowed them to keep HIPAA compliance workloads that were not ready for AWS in their data center. At the same time they didn’t paint themselves into a corner and can move those work loads out piece by piece.
Data Ingestion at Seismic Scale: Best practices for processing petabyte scale HPC workloads in the Cloud
This was an interesting session from Hess Corporation. Hess needed to process, store, and archive massive amounts of seismic readings.
Not just recent readings either. No they want/are migrating terabytes of existing data out of their data centers and into S3 for storage. They then leverage lifecycle policies to automatically archive that data to Glacier for cost savings.
The interesting takeaway from this session was how they move data from on premise locations into AWS. They use a two prong approach:
- DirectConnect: For low volumes of data, which is hundreds of gigabytes for them, they use a direct connect link. This gives them the dedicated bandwidth they need to get the data from on premise to S3.
- Snowball: For larger volumes of data they leverage Import/Export Snowball devices. They send a snowball to where the data is located, offload it onto the device, ship it back and load it into S3.
I found it interesting that they didn’t want to be stuck with just Snowball. They wanted to have the flexibility to upload reasonable data volumes quickly via Direct Connect. It was a use case for Direct Connect I had never really considered up to this point.
Wrapping It Up
All in all it was a successful start to re:Invent 2017. The crowds are big, the lines are long, but the sessions are always thought provoking. If you intended any sessions today that are worth noting, feel free to add a comment. If you have more notes to add to these sessions, add a comment as well. I will be posting my recaps after each day and I always available to answer questions on Twitter.
Hungry To Learn More?
If you enjoyed this post and are hungry to start learning about Amazon Web Service, I have started putting together a small book! You can learn about AWS services and techniques through the practical example of hosting, delivering, and securing static websites using Amazon Web Services. Stay updated on the progress of the book here.