Interesting Snippets from 2012-05-09

• Fundamental Progress Solving Bufferbloat « jg's Ramblings

  Kathie Nichols and Van Jacobson today published an article entitled ”Controlling Queue Delay” in the ACM Queue. which describes a new adaptive active queue management algorithm (AQM), called CoDel (pronounced “coddle”). The article will appear sometime this summer in the Communications of the ACM. Additionally, another independent adaptive AQM algorithm by other authors is also working its way through the academic publication cycle.

• Controlling Queue Delay - ACM Queue

  This article aims to provide part of the bufferbloat solution, proposing an innovative approach to AQM suitable for today’s Internet called CoDel (for Controlled Delay, pronounced like “coddle”). This is a “no-knobs” AQM that adapts to changing link rates and is suitable for deployment and experimentation in Linux-based routers (as well as silicon).

• Gmvault: gmail backup

  Backup and restore your gmail account at will.

  Liberate your emails and never lose that part of your life.

• How Amazon saved Zynga’s butt—and why Zynga built a cloud of its own | Ars Technica

  Zynga’s infrastructure is big. Leinwand reports having 24.5 trillion rows of data in Zynga’s database system, for 1.4 petabytes in total. Zynga has to push massive amounts of data in and out of memory as players make changes to their in-game worlds. Load balancers distribute the traffic across Web servers, while data moves from servers to an in-memory cache and on to a high-capacity, highly available system for longer-term storage. In addition to serving millions of casual gamers, Zynga also provides a platform to help third-party developers build social games. While Zynga doesn’t build its own servers the way Facebook and Google do, it does buy servers from hardware vendors who can customize them for Zynga's infrastructure, with a configuration that's similar to Facebook’s Open Compute technology. Leinwand says a Zynga workload that runs on three servers on Amazon can run on just one server in-house. The solution isn’t running a giant server—they’re about the same size as what Zynga gets from Amazon. The secret is making dozens of tweaks that add up to huge gains in efficiency. Deep analysis into server performance, network traffic trends, and how applications use resources is a time-consuming but crucial process.

[/snippets] permanent link