I am right now in Microsoft Project “Velocity” talk in PDC2008 by Muralidhar Krishnaprasad. Microsoft has been promising a distributed (and in-memory) cache system for a long long time. If I remember right it was first talked about in COM/ASP days. After that in every Microsoft event a version of it was shown (by a different team each time) in pre-release stages, but none of them got released. The story from Microsoft on the need for one, how to solve it and roadmap kept changing all the time. As for me, having got tired of this I have been using SQL Server as the distributed cache for few years now.

Notes from the session:

• “Velocity” is Microsoft's Distributed Cache .
• Usage scenarios are: Reference Data, Vendor Catalogs, Activity Data, Resource Data (Flight Seat Inventory and like)
• It is an explicit, in-memory, distributed cache
• Any .NET Objects that can be serialized can be cached
• Scale very easily, add as much memory and add as much machines as you can
• Velocity is going to be free and released in MSDN
• Runs on standard Windows PC. If machines go down, the data is preserved and not lost. High Availability (HA) is ensured
• Velocity releases: CTP2 now in PDC, CTP3 in Mix ’09 and release at Mid ’09 timeframe
• In V1.0 simple Add queries can be done. In later versions LINQ queries will be available.

You can read more on the CTP2 that got released today from the Velocity blog post here .

With what we were shown today of Velocity, especially its high availability, monitoring tools, ease of use and scalability are pretty impressive. I just hope this time they ship this and not go the previous paths.