I think that Weka is the most famous and used software for data mining in general. If you don't find what you look for in Weka, I suggest to focus more on your problem/aproach in next question here. I think that maybe there is more suitable software for very specific situations. In general, to learn/teach or experiment without wanting to focus on the tool, Weka can probably be used.
Either Rapid miner will work for you or you can make clusters using Hadoop , open source and then run with big data program using map reduce framework.
Not sure what big data means, but SEARCH in MicrOsiris runs very quickly with more than 100,000 records, as well as the other analytic programs in MicrOsiris.
Bossie Awards is a prestigious international awards for the best open source big data tools, in addition to other invaluable sources and information. Bossie Awards 2014 can be found at the following link:
we had very nice results using map-and-reduce build in in MongoDB. Oft course it depends on what you are about to do, but perhaps you might give it a try.
it depends on what actually you want to analyze in big data. I don't know about the tools but i am working on unstructured data so i found hadoop as a better platform. you may found a good review of various big data analytics platforms in attached paper.