On 15 Jan 2014, at 18:51, cotton-ben <ben.cotton@ALUMNI.RUTGERS.EDU> wrote:

Nice!  The consequences of our *necessarily* staying on-heap has been a
"monstrous" experience for us (see
http://4.bp.blogspot.com/-upwza0_lLn4/TmXB4lKkPKI/AAAAAAAAAHY/9lA7VYCmSkI/s1600/heap_0001.jpg
).

As Tristan hinted, can share (worse case privately) the reasons that lead to that horrific experience? This would be very useful to the infinispan team to better shape and explain the off-heap approach. Off-heap does come with non trivial drawbacks around manual garbage collection and memory fragmentation (at least when the data is not homogeneous).

Emmanuel

_______________________________________________
infinispan-dev mailing list
infinispan-dev@lists.jboss.org
https://lists.jboss.org/mailman/listinfo/infinispan-dev