On Sun, 2008-12-07 at 22:59 -0800, prasha8 wrote: > Have been working on a pilot project to use ActiveMQ as core message provider > for a legacy system (which is C++ based) > > One of the critical requirement is to figure out the best approach to solve > following problem > 1. I have ESB converting the WS/SOAP callouts into JMS endpoint and pushing > the messages into ActiveMQ container. > 2. Have CMS handlers on the C++ side to consume/pick the message and reply > back. > > What is not clear to me is, > 1. Can i have 2 different operations of a service push the messages to same > queue? (e.g. Authentication service has login and logout operation). What is > the best approach for these scenarios? How would the scalability issue play > into this, assuming there would be 100 users using the system > simultaneously? > > 2. Do i need to create CMS handlers as .exe files. Coz, that way, i would > end up having n number of CMS handlers ( .exe files) as consumers for each > of the services? Is there a best approach for solving this?
This really depends more on your own needs. You could embed CMS client code in existing apps or write new apps to link the legacy apps to the ESB through you own proprietary messaging subsystem. > 3. How is multi-threading handled by the ActiveMQ CMS apis? CMS follows the basic rules setup by the JMS API, we have gone to much effort to make things work in a thread safe manner. > > thanks in advance.. > -- Tim Bish http://fusesource.com http://timbish.blogspot.com/
