[infinispan-dev] Infinispan embedded off-heap cache
cotton-ben
ben.cotton at ALUMNI.RUTGERS.EDU
Wed Jan 15 12:51:06 EST 2014
Thanks very much Tristan and Jaromir for these responses.
Interesting that Netty's off-heap allocation management (jemalloc) may
deliver us a ' Save a 1xCOPY back to the heap!' advantage that Unsafe
malloc/free does not.
Peter Lawrey has commented that he will research Netty's potential to be
used in parts of OpenHFT HugeCollections .... TBD: what is the full
advantage/dis-advantage accounting of replacing direct Unsafe malloc/free
with Netty?
But these are nityt/gritty details. The big-picture news is that API
bridges (i.e. DataContainer) are available (and intended) to empower the
community to build their own "ISPN pluggable" off-heap impls of Cache<K,V>.
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
).
We are excited about this potential to bring us real remedy.
--
View this message in context: http://infinispan-developer-list.980875.n3.nabble.com/infinispan-dev-Infinispan-embedded-off-heap-cache-tp4026102p4028649.html
Sent from the Infinispan Developer List mailing list archive at Nabble.com.
More information about the infinispan-dev
mailing list