Entries tagged with scalability - Channel 9

Water Cooler Interview with Jeff Becraft from AT&T

Kirk Evans — Tue, 17 Nov 2009 20:30:00 GMT

Host Kirk Evans talks with Jeff Becraft from AT&T Hosting & Application Services about the hosting services that AT&T provides, why customers should take a serious look at hosting in the cloud, and what features of SharePoint 2010 interest him as a hosting provider. We had a great time at the SharePoint Conference 2009, and @JeffBecraft demonstrated his mad Twitter skillz, using it to take great notes at the sessions.

ARCast.TV - Scott Hanselman on scaling websites with caching

Bob Familiar — Mon, 27 Jul 2009 13:54:00 GMT

Architecting an enterprise or web application for high performance, scale, and availability can be a challenge without an appropriate caching strategy. One approach is to leverage a distributed caching platform to support the needs of performance, scale, and availability. In this ARCast episode, Scott Hanselman and Joe Shirey discuss Microsoft's codename "Velocity" project and how it supports building out these types of applications as well as what architects and developers need to think about when implementing a distributed caching approach.

ARCast.TV - Steve Michelotti of e.imagination on High Performance Web Solutions

Bob Familiar — Tue, 02 Dec 2008 13:30:00 GMT

Dr. Zhiming Xue interviews Steve Michelotti, senior solution architect at e.magination. Steve discusses what his team had to do to build a highly available 64-bit based web solution that supports a demanding service level agreement (SLA): more than 3 billion daily transactions, 99.99% of all transactions under 250 ms response time. One particular challenge they faced was dealing with the garbage collection (GC) process that is normally managed by the CLR of the .NET framework. With potentially millions of .NET objects in memory, a generation 2 GC can take 5 to 8 seconds on a 64-bit server with 16 GB memory, which quickly breaks the SLA. At their first attempt Steve and his team tried to reduce the frequency of GC occurrence. They applied several techniques such as delta refresh, minimization of boxing and un-boxing, use of object pooling, optimization of the large object heap, and an implementation of flow control between the web tier and the application tier. To ultimately address the challenge, they worked closely with the Premier Field Engineer team and the CLR team at Microsoft, which introduced a new feature called GC notification. With the new feature, which will be released in a future .NET service pack, they were able to deterministically take off a server from servicing any web request when a GC was pending on the server and put it back to service rotation when ready