December 7, 2017
Time To Read
Estimated reading time: 4 minutes

Updates and Videos from re:Invent 2017

By in Cloud, Events on December 7, 2017 | Responses (1)

As Mark Okerstorm announced during last week’s re:Invent in Las Vegas, teams at Expedia are in the midst of migrating our travel platforms to AWS. Our teams had a busy time at re:Invent showcasing some of our work and lessons learned. In this post, I want to share highlights and videos of these sessions.

Optimizing Costs as You Scale on AWS (ENT302)

In this joint session with Gulshan Kharbanda from AWS, Abiade Adedoyin from Expedia shares our approaches to cost optimization, and how we incorporate cost awareness in to our engineering culture. In particular, listen from 34:00 minutes for Abiade explain why, as a “finance person”, she “loves working with cool technologists” at Expedia to drive cost-awareness culture, and some of our best practices to manage cloud costs. Just like fences between dev and ops create a barrier for automation, agility and reliability, listen to this session to know why a fence between finance and dev can lead to spiraling cloud costs.

Also read the blog post Cloud and Finance – Lessons learned to learn more the power of partnership between tech and finance at Expedia.

Use EC2 Systems Manager to Perform Automated Resiliency Testing in Your CI/CD Pipeline (DEV338)

In this session, Willie Wheeler and Dave Bakshani share our foray into chaos engineering and show how to use EC2 Systems Manager for automated introduction of failures as part of continuous integration and delivery pipelines.

Building a Mission Critical, Serverless Photo Pipeline for 100 Million Photos (SRV315)

HomeAway’s image processing pipeline is one of the few cloud native architectures that Werner Vogel shared in his keynote at this year’s re:Invent. Mike Broadway’s session dives deep into this architecture, how this pipeline uses serverless patterns on AWS, and some of the best practices for using serverless patterns.

In case you missed Verner’s reference, click on the image below, and video from about 1 hour 55 minutes in the keynote (yes, it was a really long keynote).HomeAway image processing pipeline

Automating Zero-Downtime Production Cluster Upgrades for Amazon ECS (CON314)

In this chalk talk, Matt Callanan explains how we manage the EC2 infrastructure for large-scale ECS clusters to perform day-2 tasks like rolling upgrades. As an early adopter of ECS, Expedia has learned some of the operational lessons the hard way, and Matt talks about how it all works. This session was not recorded. The slides are below.

Going Big with Containers: Customer Case Studies of Large-Scale Deployments (ENT209)

In his second session Matt Callanan walks through the Primer platform and how our teams use containers for CI/CD. My favorite parts of this session include the discussion about feedback loops, and stats like 20 new micro-services created every day, and new apps deployed to production like environments in under 10 minutes. Currently Primer enables over 1100 applications on AWS ECS across 13 ECS clusters in different regions and environments.

Log Analytics at Expedia Using Amazon Elasticsearch Service (ABD331)

In this session with Carl Meadows of AWS, Kuldeep Chowhan from Expedia explains five different usage patterns of ElasticSearch at Expedia.

Getting Started with Amazon Aurora (DAT202)

This is a joint session with Debanjan Saha, GM of Aurora at AWS, and Brandon O’Brien and Gurmit Ghatore of Expedia. Debanjan gives an overview of Aurora, and the benefits over databases like MySQL, followed by Gurmit and Brandon sharing one of our use cases for adopting Aurora. Expedia is one of the early adopters of Aurora.

Migrating Microsoft SQL Server Databases to AWS – Best Practices and Patterns (WIN301)

As part of our cloud migration, we are also in the process of migrating our large scale production databases, which includes Microsoft SQL Server. In this joint session with Prahlad Rao of AWS, Gerry Manioci of Expedia walks through a case study of our SQL Server migration.

Expedia flies with DynamoDB: lightning fast stream processing for travel analytics (DAT324)

Brandon O’Brien Works on real-time streaming analytics systems at Expedia. In this session, Brandon shares how we use DynamoDB and DAX for fast access to reference data for streaming analytics, and the rationale behind choosing DynamoDB with DAX over other choices.

Building High-Performing Cloud Teams (ENT217)

Cloud migration does not come with a manual. Matt Parzybok, a Director of program management, is one of the few responsible for managing the cloud migration program at Expedia. In this Chalk Talk, Matt shares his stories of managing such a complex endeavor as migrating our travel platforms to the cloud, and making sense of the migration with tens of teams as it happens.

Unfortunately, this session is not recorded. Regardless, the pictures below summarize the human and social nature of this otherwise technology journey. The one on the left is of Abiade Adedoyin before the Chalk Talk and Matt Parzybok, and the one the right is during the talk.

Cloud helper, Vice President of Technology at Expedia

1 Comment

  • Gurmit - December 7, 2017

    Nicely summarized Subbu.

    Cloud is never ending journey. New products and features are being introduced at a faster rate. Teams should also provision some bandwidth in keeping pace and evolve contineously.