Techniques to Scale Up Mission-Critical Real-Time Infrastructure
In this presentation, Brian will explain the operational techniques used to scale AppNexus, the NY-based ad-tech cloud platform that has the 2nd-widest reach on the open web (Google is first.)
Brian will discuss how AppNexus scales its infrastructure without any downtime while a substantial portion of the world's Internet advertising is constantly flowing through it.
The focus of this session will be on the core component of the AppNexus infrastructure, "Cookie Monster", the server-side cookie database. The capacity of Cookie Monster has now reached tens of billions of objects with millions of read/write operations per second. The system currently runs in half-a-dozen geographically distributed data centers.
You will learn from Brian's accumulated experience while expanding the system over the past several years, always ensuring that there is no interruption of service at any time.
Brian Bulkowski, CTO & Co-Founder of Aerospike, has a background of 25 years in the Silicon Valley as a team leader, architect, and senior manager entirely in the area of large-scale server infrastructures (commodity routers at Novell, video servers at Starlight Networks, large-scale broadband at Liberate, and distributed computation at Aggregate Knowledge).
Brian founded Aerospike after learning first hand, the scaling limitations of sharded MySQL systems at Aggregate Knowledge. As director of performance at this media intelligence SaaS company, Brian led the team in building and operating a clustered recommendation engine. Prior to Aggregate Knowledge, Brian was a founding member of the digital TV team at Navio Communications and chief architect of Cable Solutions at Liberate Technologies where he built the high-performance embedded networking stack and the Internet-scale broadcast server infrastructure. Before Liberate, Brian was a lead engineer at Novell, where he was responsible for the AppleTalk stack for Netware 3 and 4.