Loading user information from Channel 9

Something went wrong getting user information from Channel 9

Latest Achievement:

Loading user information from MSDN

Something went wrong getting user information from MSDN

Visual Studio Achievements

Latest Achievement:

Loading Visual Studio Achievements

Something went wrong getting the Visual Studio Achievements

ARCast.TV - Scott Hanselman on scaling websites with caching

15 minutes, 1 second


Right click “Save as…”

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.


Follow the discussion

  • Oops, something didn't work.

    Getting subscription
    Subscribe to this conversation
  • When this thing is gonna release, I've been hearing about it for years, and the last CTP was out last April, I though the whole thing is dead already Perplexed

  • What are the plans to bring Velocity to Azure?

  • Just use memcached like the rest of the world.

  • Scott HanselmanGlucose Scott Hanselman

    Memcached is cool, but Velocity has some features like failover and regions that memcached doesn't. With memcached if you lose a node, it's gone. With Velocity your data is replicated.

  • To me that sounds like you're trying to add reliability where there shouldn't be any; The cache layer. A cache is volatile; If stuff is gone, or it expires, you repopulate it. Leave reliability to the database, it's very good at it.

  • Scott HanselmanGlucose Scott Hanselman

    Yes and no. It depends on how easy it is to repopulate. If the date is write-through and hasn't been written to a persistent store, like activity date, then it's gone forever. If you're in a situation like a bank, talking to a mainframe, you may get charged for each hit to the backend and losing a cache could mean a chargeback.

  • That's true, admittedly I hadn't considered such a case, only thinking about in scenarios where e.g. memcached would work. Isn't caching a bit of a wrong term for such use however?

  • Scott HanselmanGlucose Scott Hanselman

    Valid points...perhaps "distributed memory server"?

  • Hehe.. Something like that.


    Bonus points if you can make Velocity into an abbreviation describing it. Wink

Remove this comment

Remove this thread


Comments closed

Comments have been closed since this content was published more than 30 days ago, but if you'd like to continue the conversation, please create a new thread in our Forums, or Contact Us and let us know.