Archive for October, 2010

muSOAing for 10/22/10 – Liquid

October 22, 2010

An interesting week with a lot of interesting readings and observations to muSOA on. First of all, a city in the Hurricane belt that moved it’s data center from a bldg that can withstand only Cat 1 Hurricanes to a Water Tank that is practically indestructible.

Another surprise (or was it), the number one or two (depending on who you ask) or maybe one and two items on every CIO’s list is Cloud Adoption. BigData management and Storage Virtualization is also high up there somewhere. With data volumes growing exponentially, Data management has become a top priority, with techniques like deduplication, virtualization, thin provisioning etc.

The weather seems to be cooperating too so the forecast for the next few weeks and months are Cloudy with chances of rain and thunderstorms.


muSOAing for 10/10/10 – Colossus

October 10, 2010

With Google announcing it’s move to supplant it’s hugely popular Map/Reduce search algorithm with Colossus, what does it portend for this whole NoSQL/BigData space. Will existing users defect en masse to the new camp? I say no. Google’s case of needing a real-time or near real-time infrastructure for indexing results may not be the case for other users. A lot of folks are still happy to maintain their BigData stores as historical data warehousing repositories eventough in some cases this history may be just a few minutes old.

The news trickling down the grapevine is that there will not be a mad rush to adopt this new kid around the BigData block. Map/Reduce is too entrenched and folks using this are quite happy with it though everyone has implemented their own homegrown analytical frameworks around it using tools like Pig and Hive. In fact, there is nothing cast in stone here, almost every map/reduce job is unique in itself.