Hi,
Can we send a Http Request directly to wM Broker i.e without IS intervention. If yes, Please let me know how can we achieve that.
Thanking in advance,
nshvr
No its not possible,IS intervention is definetely required.ie you should invoke an IS service and this will internally submits to Broker (for pub/sub).What is the reason for this requirement and trying to do??
HTH,
RMG
IMO, you should be looking to bypass/eliminate the use of Broker, not IS.
If you want to use http, you need a IS.
But you can also write your own Broker Client.
See the guides:
–webMethods Broker Client Java API Programmers Guide.pdf
–webMethods Broker Client C API Programmers Guide.pdf
Maybe this helps
-Rob
Good suggestions by Rob. I would offer that if you’re going to connect directly to Broker that you should probably use the JMS API. That way you at least avoid the proprietary nature of the Broker Java and C APIs.
Hi All,
Thanks for the response.
My intention of asking above question is,
If http request is sent to IS and IS is down after receiving data. How can we retain the Http request data without re-sending the Http request ?
Regards,
nshvr
If you need this functionality on IS, you’ll need to persist the data to disk in some way. Using TN for this is common. You receive the request, write the data, then return the status–which might just be a simple http 200 status.
The key part of this is writing the data before returning “success” to the client. The client must wait for a positive status return before considering the message to be sent.
If IS fails, or the network fails, or the client fails after the data has been sent from the client to IS but before IS says “got it”, the client has no alternative than to resend. If you want a client-side “fire and forget” facility (client-side queueing) that is not susceptible to server or network outages, then you’ll need to implement such a thing or use an off-the-shelf component (MQ is often used for this).
Using Broker doesn’t really simplify this logic. If you send to Broker directly you won’t have to implement server-side persistence but your client still needs to make sure it is robust enough to assure data transfer in the face of server and network outages.
Using IS, you can have IS use a couple of different approaches to persist the data before returning “got it”. One is to have IS publish the doc to Broker right away, and then return control to the client. Another, is to submit the data/document to TN and if a bizdoc is returned from TN, success can be returned to the client.
HTH
All the Responses helped me a lot. Thanks for the Response.