Terracotta Discussion Forums (LEGACY READ-ONLY ARCHIVE)

Hello !

I need to write an application under some design constraints and I wonder if I can do that with Terracotta (with EhCache or without).

I have two applications that need to share a HashTable/Cache "C". Application A will actually access and modify only a subset of C let's say "AC", and application B will access also access a subset of C, let's say "BC".

So not all elements will actually be used by both side, but AC and BC will intersect (we don't want to define in advance which elements of the table will be shared or not).

My problem : A and B don't have a lot of memory allocated. I don't want the whole table to be replicated on A and B, but only the respective subset they will actually use.

When A is startup up and inserting all elements it needs, and I start B afterward, B is loading the whole content of the table on his side, since it will not actually use the whole table.

In other words, my question is : is it possible to design an architecture so when a new client connects to the cluster, he doesn't load all the shared object, but only when it needs it (actually it will need a subpart of it) ?

Thank you so much !

Actually, with EhCache it seems to work pretty well. However I'd like to use a solution without EhCache to have more control on simple objects.

If I use an HashTable, a client connecting to the cluster get his memory filled up as soon as he connects...

If anyone knows a correct design or correct options in tc-config to achieve this, I would be grateful

I think ehcache is the best option for this.

Im agree with Rajoshi

dentist washington dc