Hey,
so I've updated my tests for measuring memory consumption so as to work
in a compatibility mode and the results are following (extending latest
Galder's report with my results):
Infinispan memory overhead, used in library mode:
Infinispan 5.2.0.Final: ~151 bytes
Infinispan 5.3.0.Final: ~135 bytes
Infinispan 5.3.0.Final (compat. mode): ~135 bytes
Infinispan memory overhead, for the Hot Rod server:
Infinispan 5.2.0.Final: ~174 bytes
Infinispan 5.3.0.Final: ~151 bytes
Infinispan 5.3.0.Final (compat. mode): ~175 bytes
Infinispan memory overhead, for the REST server:
Infinispan 5.2.0.Final: ~208 bytes
Infinispan 5.3.0.Final: ~172 bytes
Infinispan 5.3.0.Final (compat. mode): ~173 bytes
Infinispan memory overhead, for the Memcached server:
Infinispan 5.2.0.Final: ~184 bytes
Infinispan 5.3.0.Final: ~180 bytes
Infinispan 5.3.0.Final (compat. mode): ~224 bytes
Notable differences are for HotRod and Memcached clients. The tests for
Memcached are (and have always been) using SpyMemcached client to store
the entries into the server. For the compatibility mode, I had to
register a custom marshaller that would be able to marshall/unmarshall
data sent by SpyMemcached. I used this marshaller:
https://github.com/infinispan/infinispan/blob/master/integrationtests/com...
. That might be a partial cause of the bigger footprint.
Martin