2019 Distributed SQL Summit Recap and Highlights
Well, that’s a wrap! Yugabyte would like to extend a special thanks to JD and Amanda from the Postgreconf.org team, and to all the speakers from Facebook, Google, Amazon, Pivotal, Salesforce, Narvar, Plume Design and others that presented at the first-ever Distributed SQL Summit on Sept 20, 2019.
If you couldn’t make it out to this year’s event, have no fear! In this post we’ll recap some of the highlights from the Summit, as well as provide you with links to playbacks of the presentations. Ok, on to what you missed…
Keynote: James Watters, SVP Products – Pivotal
James walked us through the evolution, benefits and challenges that enterprise IT faces in transforming how it delivers applications, security, automation and data in a cloud-native first world.
View James’ complete presentation: “The Past, Present and Future of Enterprise IT in a Cloud-Native World.” Presentation slides can be viewed on SlideShare.
Panel: Facebook’s Distributed Database Evolution
In this panel discussion we dove into Facebook’s distributed database journey covering technologies like MySQL UDB, RocksDB and ZippyDB. Panelists included:
- Jeff Rothschild, Entrepreneur & Engineer, Ex-Facebook VP of Infrastructure
- Victoria Dudin, Engineering Manager, Facebook
- Vishal Kathuria, Engineering Manager, Facebook
- Harrison Fisk, Director Production Engineering, Facebook
- Dhruba Borthakur, Co-Founder & CTO, Rockset
- Karthik Ranganathan, Co-Founder & CTO, YugaByte (Ex-Facebook and Moderator)
There were plenty of insights, humor and war stories. Here’s a few highlights from the panel:
- Jeff talks about what it was like in 2005 to support one database for each of the 2000 schools registered with Facebook, the acrobatics required to migrate users, complex cache invalidation schemes and the lessons learned along the way to ultimately build an infrastructure that could support billions of users.
- Harrison describes how he accidentally dropped a production database and had 30 mins to restore it before the cache started to invalidate and users started to notice.
- How in the pre-cloud era, Facebook was often on the brink of running out of physical storage, often receiving deliveries to the datacenter, just in the nick of time.
- The origin stories for why projects like RocksDB, Tao and ZippyDB were created.
Want to bring down Facebook? There’s a helper routine for that! Jeff tells a cautionary tale about making sure you understand your abstractions in the snippet below.
View the complete panel discussion: “Facebook’s Distributed Database Evolution.”
Ram Ravichandran, CTO – Narvar
70% of U.S. adults have experienced better retail through Narvar’s customer engagement platform. In this talk, Ram Ravichandran – CTO, walked us through some of the highlights on how Narvar scaled their microservices architecture and data-tier onto multiple clouds. He also discussed how they simplified their operations and maintained GDPR compliance throughout the entire process.
A few highlights and lessons learned from Ram’s talk:
- Migrated from PostgreSQL, ElastiCache and DynamoDB to YugabyteDB
- In production for over a year
- Achieving 10k writes per sec and 4k reads per sec
- Less than 10 ms latency for writes and less than 2ms latency for reads
- 4x cost savings over previous data infrastructure
- Seamless database upgrades from version to version in production
View the complete presentation “Scaling Microservices and Distributed Databases in a Multi-Cloud World.”
Muru Guruswamy, Principal Data Architect – Plume Design
Plume Design, is an intelligent Wi-Fi services delivery platform that powers millions of smart homes with the help of distributed SQL in the cloud. In this presentation Muru Guruswamy, Data Architect at Plume walks us through yearly engineering retros to learn what the growing pains were along the way to eventually be able to support 27 Billion ops per day and 35 TBs of data using 60 nodes of Yugabyte DB.
A few highlights and lessons learned from Muru’s talk:
- MongoDB could not handle high volume event data
- Maintaining MongoDB at scale was a headache
- Running Cassandra is expensive and requires a lot of the Ops team’s attention
- Yugabyte DB was selected after being evaluated against Google Spanner, InfluxDB, TimescaleDB and ScyllaDB
View Muru’s complete presentation: “Powering millions of smart homes with near real-time cloud-agnostic, distributed SQL.”
- Building planet-scale cloud-native apps with GraphQL and distributed SQL: Tanmai Gopal, CEO/Co-Founder, Hasura (Slides)
- Building Microservices in a Cloud-Native World with Distributed SQL: Ryan Scheidter, Lead Software Engineer, Cerner
- Google Spanner’s SQL Evolution: Campbell Fraser, Software Development Lead – Google
- How-To: Zero Downtime Migrations from Oracle to a Cloud-Native PostgreSQL: Rajkumar Sen, Chief Technical Architect, Founder, BlitzzIO (Slides)
- Spring Forward to a Cloud-Native PostgreSQL: Guillermo Tantachuco, Principal Platform Architect and Nikhil Chandrappa, Senior Platform Architect, Pivotal
- Distributed MySQL Architectures, Past, Present and Future: Peter Zaitsev, Founder & CEO, Percona (Slides)
- Strongly Consistent Global Indexes for Apache Phoenix: Kadir Ozdemir, Principal Architect – Salesforce (Slides)
- Benchmarking Distributed SQL Databases: Robert Winslow, Performance Engineer &, Data Scientist
Distributed SQL Summit 2020
We are looking forward to working again with the folks at PostgresConf to bring the Distributed SQL Summit with a new line up speakers to Postgres Conference 2020 in March next year. Stay tuned for a future announcement regarding the event.