Re: [infinispan-dev] rethinking ISPN transactions