Author |
Message
|
brokerguy |
Posted: Fri Apr 05, 2013 8:44 am Post subject: SCAInput for Atomic service calls |
|
|
 Apprentice
Joined: 18 Mar 2013 Posts: 26 Location: Cosmos
|
Hi,
In process server, we can make SCA invocations to call atomic services (as part of service composition).
Can we do this in Message broker as well ? I mean rather than making an HTTP roundtrip using SOAPRequest node, can I use SCARequest (in composite service flow) and SCAInput node (in atomic service flow) ? No process server is involved in this mix. If this is not possible, is HTTP round trip the only option ? |
|
Back to top |
|
 |
lancelotlinc |
Posted: Fri Apr 05, 2013 9:22 am Post subject: |
|
|
 Jedi Knight
Joined: 22 Mar 2010 Posts: 4941 Location: Bloomington, IL USA
|
|
Back to top |
|
 |
mqjeff |
Posted: Fri Apr 05, 2013 11:25 am Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008 Posts: 17447
|
I surely don't understand the question.
How can you make an SCA connection to anything when there is no SCA container?
Surely this question is not simply "how do I use HTTP or SOAP asynchronously"? That question has a *directly obvious* answer. |
|
Back to top |
|
 |
brokerguy |
Posted: Fri Apr 05, 2013 11:44 am Post subject: |
|
|
 Apprentice
Joined: 18 Mar 2013 Posts: 26 Location: Cosmos
|
lancelotlinc wrote: |
Maybe I did not understand your question? |
Let me elucidate my question
First, there is no WPS in the architecture
Now, if I have to implement a SOAP/HTTP composite service in WMB, I may have to invoke some atomic SOAP based services (these atomic services are also deployed in WMB). One way of doing this is using SOAPRequest node (in case of synchronous call) to make a roundtrip and come back to WMB to invoke the atomic services. The atomic service has SOAPInput in this case.
Alternatively, can I design my atomic services to have SCAInput node and use SCARequest node from the composite service to invoke the atomic service. |
|
Back to top |
|
 |
lancelotlinc |
Posted: Fri Apr 05, 2013 11:52 am Post subject: |
|
|
 Jedi Knight
Joined: 22 Mar 2010 Posts: 4941 Location: Bloomington, IL USA
|
|
Back to top |
|
 |
mqjeff |
Posted: Fri Apr 05, 2013 12:30 pm Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008 Posts: 17447
|
brokerguy wrote: |
can I design my atomic services to have SCAInput node and use SCARequest node from the composite service to invoke the atomic service. |
To what value?
You keep using "atomic".
What meaning are you ascribing to that word in this context? What "atomic" properties are you expecting to achieve?
What properties of "SCA" transport or definition are you attempting to make use of?
You keep complaining about "round trip". Again, are you really trying to make asynchronous calls rather than synchronous calls? Is that what you mean by "atomic"? |
|
Back to top |
|
 |
brokerguy |
Posted: Fri Apr 05, 2013 1:21 pm Post subject: |
|
|
 Apprentice
Joined: 18 Mar 2013 Posts: 26 Location: Cosmos
|
mqjeff wrote: |
You keep using "atomic".
What meaning are you ascribing to that word in this context? What "atomic" properties are you expecting to achieve?
|
Let me state with examples:
Service A (synchronous req/resp with SOAPInput node)
Service B (synchronous req/resp with SOAPInput node)
Service C (synchronous req/resp with SOAPInput node)
All the services are deployed in WMB
Now, Service A to fulfill its functional requirement has to invoke Service B & Service C.
I am referring to Service A as 'composite service' and Service B & C as 'atomic service'
In the above scenario, Service A would have to make SOAP/HTTP calls to Service B & C. Service A uses SOAPRequest node to invoke B&C
I was thinking whether there is any alternative. Like, for Service B & C having SCAInput and Service A calling them using SCARequest node. |
|
Back to top |
|
 |
goffinf |
Posted: Fri Apr 05, 2013 1:35 pm Post subject: |
|
|
Chevalier
Joined: 05 Nov 2005 Posts: 401
|
brokerguy wrote: |
mqjeff wrote: |
You keep using "atomic".
What meaning are you ascribing to that word in this context? What "atomic" properties are you expecting to achieve?
|
Let me state with examples:
Service A (synchronous req/resp with SOAPInput node)
Service B (synchronous req/resp with SOAPInput node)
Service C (synchronous req/resp with SOAPInput node)
All the services are deployed in WMB
Now, Service A to fulfill its functional requirement has to invoke Service B & Service C.
I am referring to Service A as 'composite service' and Service B & C as 'atomic service'
In the above scenario, Service A would have to make SOAP/HTTP calls to Service B & C. Service A uses SOAPRequest node to invoke B&C
I was thinking whether there is any alternative. Like, for Service B & C having SCAInput and Service A calling them using SCARequest node. |
Ok, that makes your use of the word atomic clear, but since there are other more obvious interpretations, perhaps using 'composition member service' would remove the ambiguity. I'm sure you understand the power of shared vocabulary when trying to explain requirement... I'm just saying |
|
Back to top |
|
 |
lancelotlinc |
Posted: Mon Apr 08, 2013 4:06 am Post subject: |
|
|
 Jedi Knight
Joined: 22 Mar 2010 Posts: 4941 Location: Bloomington, IL USA
|
brokerguy wrote: |
Like, for Service B & C having SCAInput and Service A calling them using SCARequest node. |
You can use any node you want, as long as it works. What advantage do you see using SCA? Since it is being called by Svc A, whats the point of using something other than SOAP node to call it?
What disadvantage are you trying to overcome by selecting a different node than SOAPRequest ? _________________ http://leanpub.com/IIB_Tips_and_Tricks
Save $20: Coupon Code: MQSERIES_READER |
|
Back to top |
|
 |
vmcgloin |
Posted: Mon Apr 08, 2013 4:28 am Post subject: |
|
|
Knight
Joined: 04 Apr 2002 Posts: 560 Location: Scotland
|
To add my tuppence worth:
- You still have to choose between an MQ binding or SOAP/HTTP binding if you use the SCA nodes, so you gain no particular advantage as far as I can see.
- Are you proposing this just to allow reuse between WMB and WPS components without multiple input nodes? Or just as a theoretical question?
- Have you tried it yet? |
|
Back to top |
|
 |
mqjeff |
Posted: Mon Apr 08, 2013 5:20 am Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008 Posts: 17447
|
There's a lovely sample that comes with the product for performing webservice aggregation.
It even manages to call each service asynchronously, so the total time to process a single aggregated request is Tn+C, where Tn is the longest time for any single service to response, and C is some constant to represent the time to process the aggregated responses.
This contrasts nicely with the naive first approach which takes T0+T1+T2+T...+Tn + C. |
|
Back to top |
|
 |
|