Marcato

We’ve been doing (and will be doing) some talks around the Twin Cities on Microsoft’s “AppFabric Caching” technology – a distributed cache not unlike the open-source memcached library.

What has this to do with agile?  Well, we’ve found that it can address a core component of agile teams – the desirability of being self-sufficient and avoiding the bottleneck of scarce experts. 

Let me explain: For projects/products that reach sufficient scale, when a database is present, it becomes the bottleneck for performance.  This often leads to the need for That Guy With Awesome SQL Chops to come to the rescue.  After sufficient time/energy/dollars, your application is performant – but you may have had some intermediate pain, and the data layer (including stored procedures) may have taken a hit in terms of maintainability.  You may now need an expert to touch that code in the future.

For some classes of data, however, a good caching solution can eliminate all that.  A cache that can scale to large volumes of data (by being distributed across a cluster of machines) and yet presents itself as a unified cache logically (so that you have a single place to invalidate cache items) will allow you to tackle a broader class of problems than simple per-node caches.

So, if your team has a good caching solution in the tool belt, you’ll be more agile – able to reach greater scalability while staying self-sufficient.

You can catch this talk on May 20th (Microsoft’s Connected System User Group) or June 8th (Twin Cities Developer Guild.)  Slides are available here.