[omniORB] omniORB v402 anomally ?

Silvia Andreia Rocha silvia.rocha at SE.EFACEC.PT
Tue Sep 2 13:23:46 BST 2003


Hi All!

 

I have been exhaustively testing and comparing several ORBs.

One of the tests gives info on the ORB "performance" in several scenarios of
"m clients" -"n servers". The servers are actually several instances of the
same CORBA Server.

 

This test is very simple: for the time window of 5 seconds a client
continuously calls a method in all existing CORBA servers, the score is the
number of CORBA invocations that the "very busy" client was able to do in
that period.

 

I have detected what I believe to be an anomaly in omniORB v.402. Just to
make sure I've made the test to omniORB v402 twice and the results are
basically the same, differ in very small scores but the anomalous behaviour
persists.

 

The first ORB also behaves strangely but it will be of no importance in the
future, I will however maintain it anonymous ;)

 

Test environment: OS: NT4.0, CPU: PIII 1GHz, RAM: 256 MB.

 

 

Scenario 1-20

ORB X: 104

omniORB v401: 152    

omniORB v402: 197       Excellent performance!

 

Scenario 1-30

ORB X: 113

omniORB v401: 128

omniORB v402: 131       Still good results!

 

Scenario 1-40

ORB X: 104

omniORB v401: 112

omniORB v402: 99         What's going from now on?

 

Scenario 1-50

ORB X: 92

omniORB v401: 100       

omniORB v402: 96

 

Scenario 1-60

ORB X: 66

omniORB v401: 96         

omniORB v402: 92

 

 

Did anyone observe similar behaviour while testing/using omniORB v402? 

 

 

Thanks a lot for the attention,

Sílvia Rocha

 

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.omniorb-support.com/pipermail/omniorb-list/attachments/20030902/a4073305/attachment.htm


More information about the omniORB-list mailing list