Håkon, Using a queue in de middle to persist messages while throttling is indeed the way to go. To control the number of requests coming out of the queue, you can either: - use the new JMS endpoints (http://servicemix.apache.org/servicemix-jms-new-endpoints.html) and configure synchronous exchanges so you can keep track of the amount of messages currently being processed - use a Camel JMS endpoint to start the route, which allows you to set the number of threads for that particular route (contrary to the JBI paradigm which uses thread pools) -- camel also has a throttler pattern that allows throttling the number of requests per time slot (e.g. 10 per 30 seconds)
Regards, Gert Vanthienen ------------------------ Open Source SOA: http://fusesource.com Blog: http://gertvanthienen.blogspot.com/ 2009/10/6 Håkon Sagehaug <[email protected]>: > Hi all, > > Any tips or hint if/how this can be resolved in servicemix? > > cheers, Håkon > > 2009/10/2 Håkon Sagehaug <[email protected]> > >> Hi all, >> >> I wanted a tip/guidance to see if I'm on the right track. We have a web >> service(python), the server that runs the server has a queue inside it to >> control that no more then lets say 10 request is forwarded to the ws. The >> reason for this is that the pyhton ws created a new process, so one request >> == one new process. >> >> But then we decided to look closer to servicemix and see if we could throw >> the queue out from the ws server. My initial thoughts on the configuration >> is like this: >> >> incoming soap over http request -> cxf-bc (consumer) -> jms-bc(provider) -> >> queue(activemq) -> jms-bc(consumer) -> cxf-bc(provider) -> external service >> >> Is this the correct picture of what components that needs to be involved? >> Also not sure how i can configure that property we have on the server that >> lets say in the queue the 11'th request is sent to the ws, now before we >> send a new request, at least one response should come back so there is a >> free process from the process pool on the server ready to be started. Any >> tips on how to realize this? >> >> hope my explanation was clear >> >> cheers, Håkon >> >> -- >> Håkon Sagehaug, Scientific Programmer >> Parallab, Bergen Center for Computational Science (BCCS) >> UNIFOB AS (University of Bergen Research Company) >> > > > > -- > Håkon Sagehaug, Scientific Programmer > Parallab, Bergen Center for Computational Science (BCCS) > UNIFOB AS (University of Bergen Research Company) >
