on 05-14-2009 4:24 AM
I referred following links and as many as 25 previous posts before posting this question.
/people/boris.zarske/blog/2007/06/13/sizing-a-system-for-the-system-landscape-directory-of-sap-netweaver
I have some queries related to improving the performance of a scenario in my landscape.
Scenario : IDOC-SOAP (Synchronous with BPM)
The mapping is simple with 3 fld direct mapping.
Works grt with no problem in normal cases for one message transfer.
But initial load ( when we put it in new system, the scenario generates erros if we send more than 23 IDOCs at a time)
I have decided to perform the following to solve it.
/people/william.li/blog/2008/03/07/step-by-step-guide-in-processing-high-volume-messages-using-pi-71s-message-packaging
Will it help me because its a IDOC sync scenario.
What can I do to improve the performance of this scenario?
Please suggest me good suggestions to improve the performance
Please do list me out points that I have to perform (for a sync scenario with BPM) as I am already confused watching lots of blogs posts.
Nikhil.
/people/william.li/blog/2008/03/07/step-by-step-guide-in-processing-high-volume-messages-using-pi-71s-message-packaging - dont think that will work for sync scenarios
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
when you say its a sync scenario i assume its IDOC -> SOAP -> IDOC.
How much time is the WS taking to respond back?
Is this causing the issue?
If its a simple transformation i dont see any reason for the BPM to cause issues unless a quick sizing or performance tuning of the XI server has not been done after post installation.
If I enable packaging for IS, how will it increase the performance?
it will carry the data in equal mesg of the size we specify but the processing at the WS needs to be one by one. Correct me if I am wrong.
I believe this is for improving performance inside PI.
How will it react for synchronous scenarios where it has to wait for the response back from the receiver.
Is there a way in PI that when I send multiple messages the files are processed many at a time.. without chaning the existing MM and whn we send only one msg, it has to process that single msg can it b done?
Thanks,
Nikhil
Is there a way in PI that when I send multiple messages the files are processed many at a time..
without chaning the existing MM and whn we send only one msg, it has to process that single msg can
it b done?
If a scenario is configured to handle many messages at a time then it can also proces a single message at a time.....no special setting is required....but rember vice-versa is not always posssible...
If a scenario is configured to handle many messages at a time then it can also proces a single message at a time.....no special setting is required....but rember vice-versa is not always posssible...
What do you mean by this?
does it mean i have to duplicate IDOC structure and WS structure in my Mapping?
can you explain a little in detail?
thanks
Nikhil.
If I enable packaging for IS, how will it increase the performance?
I believe this is for improving performance inside PI.
check out this section 3.2.9.1....it will tell you how performance will be improved by packaging
[Sizing And Performance Tuning|https://www.sdn.sap.com/irj/scn/go/portal/prtroot/docs/library/uuid/2016a0b1-1780-2b10-97bd-be3ac62214c7]
Dont forget to read the embedded note....
excuse me for not putting it across in right way...
What I meant was...
Suppose I am sending 100 IDOCs from SAP - PI - WS.
PI will make one call to WS each time.
instead, is it possible to enable any settings on WS / PI or change design objects so that WS accept 10 mesg from PI at a time and it should ideally work good even for a single message.
i dont think we can do it unless we change the mapping by duplicating messages or chaning the occurances... still not a good idea but just checking
Nikhil
Please do list me out points that I have to perform (for a sync scenario with BPM)
As far as possible avoid a Transformation step in BPM....instead if possible include the mapping in the Interface Determination....this is one of the methods to improve performance
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
do you think that the performance tuning that I mentioned in the link will hold good with sync scenarios?
i dont think so.....in this scenario the async mesg processing is made to wait
from the blog:
As you can see, for the 1st minute, all the messages are waiting to be processed. After 60 seconds,
the packages will be created and processed. There is no change to the monitoring of each of the
individual messages in SXI_MONITOR.
....but if you make a sync scenario wait....then probably you may run into the risk of blocking the Queues.......
If this is going to be in production then i would have been more careful...becoz firstly it is BPM....then synchronous BPM.....then a processing wait......normally do not try to make the BPM processing wait...my small suggestion
User | Count |
---|---|
84 | |
10 | |
10 | |
10 | |
7 | |
6 | |
6 | |
5 | |
4 | |
4 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.