The term grid has been a buzzword for quite a while. I used to think that it was merely about re-branding of the usual clustering/group communications concepts (e.g. think JGroups) and carried little sense from the technical point of view. Well, it turns out that was my misunderstanding probably resulting from all the marketing garbage.
Recently I barely scratched the surface of this new paradigm by looking at a few products such as Hadoop and GridGain. Computational grid frameworks seem to follow a common pattern and definitely introduce a new way (I believe for many of us) to structure extremely scalable algorithms. There are also complimentary data grid technologies such as GigaSpaces enabling distributed caching of enormous amounts of data.
Most people expect that multi-threading will be in more demand soon as a result of the evolution of multi-core CPUs but this area is understood well enough. Those of us interested in middleware and distributed systems will also need to familiarize ourselves with the nascent grid technologies in order to capitalize on the potentially juicy employment opportunities.
Recently I barely scratched the surface of this new paradigm by looking at a few products such as Hadoop and GridGain. Computational grid frameworks seem to follow a common pattern and definitely introduce a new way (I believe for many of us) to structure extremely scalable algorithms. There are also complimentary data grid technologies such as GigaSpaces enabling distributed caching of enormous amounts of data.
Most people expect that multi-threading will be in more demand soon as a result of the evolution of multi-core CPUs but this area is understood well enough. Those of us interested in middleware and distributed systems will also need to familiarize ourselves with the nascent grid technologies in order to capitalize on the potentially juicy employment opportunities.
Posts
