What do you like best?
Great beginner documentation, proven track record, open-source, performant, great community.
Our adoption has been relying mostly on the open-source version of Infinispan. We have not reached out to RedHat for any enterprise support, as we would like to first evaluate the product before opting for the Enterprise features.
What do you dislike?
Lack of documentation on some of the advanced features. Some sections of the documentation are unfinished.
Monitoring of the Infinispan cache is done through RHQ, not included out of the box. RHQ implements an agent/server model, where there's an agent on every Infinispan node, sending metrics to the RHQ server for aggregation and visualization. This means you have yet another component to setup and configure.
Recommendations to others considering the product
Infinispan is quick to get running out of the box, but you will start slowing down when you try to do anything non-standard (or not explained in the documentation). Overall adoption is quick but some of the advanced features are not well documented.
The community is active and I was able to Google answers to most of the questions. We have not reached out for paid support.
What business problems are you solving with the product? What benefits have you realized?
Using Infinispan as a remote, clustered Cache. Allow server restarts without clearing/reloading the cache.