<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
<font color="#666666">> [...] , but I don't understand why the
fact that it's
<br>
running on a different process is limiting in any form.</font><br>
<br>
Y<font color="#3333ff">ou are correct, it is not limiting. XA is
process locality independent. Individual TXN participants
(heterogeneous or homogeneous) register as XA resources (via JTA)
with a central TM (which, if JTS compliant, supports XA). The TM
then basically implements a classic 2PC to bound the scope of the
TXN processing and outcome. An example of an XA transaction with
heterogeneous participants could be a database (viewed by the TM
as a resource=javax.sql.XADataSource) and a Queue (viewed by the
TM as resource=javax.jms.XAQueueConnection). <br>
<br>
BTW, though not done in the JCache 1.0 draft, an ideal Java API in
support of Cache as XA participant will be realized when JSR-107
provides something like a javax.cache.XACacheSource.</font><br>
        
        
        
<br/><hr align="left" width="300" />
View this message in context: <a href="http://infinispan-developer-list.980875.n3.nabble.com/infinispan-dev-CacheStore-redesign-no-XA-cache-stores-tp4027770p4027772.html">Re: [infinispan-dev] CacheStore redesign: no XA cache stores</a><br/>
Sent from the <a href="http://infinispan-developer-list.980875.n3.nabble.com/">Infinispan Developer List mailing list archive</a> at Nabble.com.<br/>