> 1) Can we use Storm on No SQL or do we need to mix them with Pig/Hive/Spark > or any other for any reason? > 3) Do Storm supports only Hadoop, HBase?. We may use > Cassandra/MongoDb/CouchBase as well. > 4) Do Storm supports RDBMS too?. We can have a single interface to pull out > data from multiple data sources?
I'm not 100% sure what you mean by "supports". I'm assuming you mean can you write data to any of those data stores from Storm bolts. If you can access the data store using the programming language you've chosen to implement your Storm topologies (likely Java), you can write to it from Storm bolts. There are helper utilities / classes / bolts for some of the data stores listed above (some of them mentioned in Parth's response) which make writing data to those data stores easier, but there's nothing forcing or requiring you to use them to write to a data store of any kind. If, by chance, by "supports" you mean can you feed data from those data stores to a Storm topology, the answer is more or less the same, except that writing your own implementation of a Storm spout doing so is a little bit more complicated than writing Storm bolts. IMHO. Personally I've found the data store specific connectors a little too generic for my needs, but YMMV. -TPP
