Scalable Internet Architectures

We will dive deep into traditional web architectures and discuss what makes them go. Looking are real-life architectures, we’ll analyze weak points and discuss what will happen to them under the intense pressure of a prolonged exposure on Digg, the NYTimes or even MSN (or when traffic jumps from 15Mbits/second to over 1Gbit/second inside 30 seconds).

In a friendly atmosphere, you will enjoy a mix of anecdotes, war stories, good and bad design principles and more tips and tricks than you could use in a year.

Topics will include:

• the importance of visualizing data
• web protocols (HTTP/HTTPS) and browser behavior
• firewalls and load balancers
• HTTP requests, sizes, pipelining and what that means in packets
• remedial queue theory and why apps just “stop working”
• solid caching techniques for correctness and performance
• scary caching techniques for emergencies
• techniques for service isolation (quality of service protection)
• databases, caching and their torrid love affair
• service queues, how they work and why they usually don’t