otmfaqForumsBlogsRegister
FAQMembers ListCalendarToday's PostsSearch


 Subscribe Blogs:RSS
 Subscribe Forums:RSS
 Follow New Posts:Twitter
OTMFAQ Home
OTMFAQ Blogs
OTMFAQ Forums
OTM Wiki

OTM SIG
OTM Wiki
MavenWire


Performance, Scalability and HA Optimizing the performance of OTM / G-Log, configuring Scalability (SCA) and maintaining High Availability.

Reply
 
Submit Tools LinkBack Thread Tools Display Modes
  #1 (permalink)  
Old September 24th, 2008, 21:57
Junior Member
 
Join Date: Aug 2007
Posts: 9
Groans: 0
Groaned at 0 Times in 0 Posts
Thanks: 5
Thanked 4 Times in 3 Posts
Rep Power: 0
jlgottlieb is on a distinguished road
Using OutXmlProfiles to help OTM Performance

Hi all, I just wanted to present some of our results with OutXmlProfiles to boost performance in case any finds this helpful.

I attended the OTM SIG User Conference in Philadelphia last month and Eric Rosenbloom did a presentation on OTM Performance. This is a big concern of ours and I'm always looking for ways to boost performance. In one case study, he talked about a situation where all normal tuning efforts did not find any problem areas. After much investigation it was found that during "slow" times, a great deal of integration was happening. The fix was to implement OutXmlProfiles for the integration. Since I am getting reports of "slow" times as I implemented OutXmlProfiles on two of our ExternalSystems.

Both of these ExternalSystems use the PlannedShipment element and the profile was set to MAX_ELEMENTS. This is good when you are doing development because you want to see all the available data. What we found was that we do not actually use most of the data in the GLOG XML.

We started out developing a OutXmlProfile using the MIN elements and adding the XPaths, but I ran into some problems with this where the XPATH that I added was not in the generated XML. We found that the easiest method was to use an XML_TEMPLATE, which is basically a skeleton XML with only the elements that you want. Then define your OutXmlProfile based on the XML_TEMPLATE.

Here are our results:
# of sql calls
338 74 86
size of Xml 95 - 110K 34-40K 44-50K
avg # pushes/day 1100 (*) 3800 (*)

With a little testing, I was able to determine that the full Glog XML for a PlannedShipment element was making 338 sql calls to generate. After implementing the OutXmlProfiles, the number of sql calls was reduced to 63 for one external system and to 78 for the other. The size of the XMLs were reduced from 95 - 110K to
Digg this Post!Add Post to del.icio.usBookmark Post in TechnoratiFurl this Post!
Reply With Quote
The Following 2 Users Say Thank You to jlgottlieb For This Useful Post:
acuartero (September 29th, 2008), chrisplough (September 25th, 2008)
  #2 (permalink)  
Old September 24th, 2008, 22:05
Junior Member
 
Join Date: Aug 2007
Posts: 9
Groans: 0
Groaned at 0 Times in 0 Posts
Thanks: 5
Thanked 4 Times in 3 Posts
Rep Power: 0
jlgottlieb is on a distinguished road
Using OutXmlProfiles to help OTM Performance

Hi all, I just wanted to present some of our results with OutXmlProfiles to boost performance in case anyone finds this helpful.

I attended the OTM SIG User Conference in Philadelphia last month and Eric Rosenbloom did a presentation on OTM Performance. This is a big concern of ours and I'm always looking for ways to boost performance. In one case study, he talked about a situation where all normal tuning efforts did not find any problem areas. After much investigation it was found that during "slow" times, a great deal of integration was happening. The fix was to implement OutXmlProfiles for the integration. Since I am getting reports of "slow" times as well, I implemented OutXmlProfiles on two of our ExternalSystems.

Both of these ExternalSystems use the PlannedShipment element and the profile was set to MAX_ELEMENTS. This is good when you are doing development because you want to see all the available data. What we found was that we do not actually use most of the data in the GLOG XML.

We started out developing a OutXmlProfile using the MIN elements and adding the XPaths, but I ran into some problems with this where the XPATH that I added was not in the generated XML. We found that the easiest method was to use an XML_TEMPLATE, which is basically a skeleton XML with only the elements that you want. Then define your OutXmlProfile based on the XML_TEMPLATE.

Here are our results:

No. of sql calls: 338 (Full Xml) to 74 (ExtSys1), 86 (ExtSys2)
Size of XML 95-110K (Full Xml) to 34-40K (ExtSys1), 44-50K(ExtSys2)

I'm not sure how to quantify this properly, but we have seen %CPU on our database server drop by about 10-15% with these OutXmlProfiles implemented.

I am now looking at our RateInquiry and GenericQuery, both of which have an optional OutXmlProfile element in the request. Sadly the ShipmentQuery does not have an OutXmlProfile option. I've been trying to use the GenericQuery to get a Shipment Element returned, but no luck so far. If anyone has a suggestion, let me know.

I thought our results with OutXmlProfiles might be useful to some of you out there.
Digg this Post!Add Post to del.icio.usBookmark Post in TechnoratiFurl this Post!
Reply With Quote
The Following User Says Thank You to jlgottlieb For This Useful Post:
acuartero (September 29th, 2008)
Reply


Thread Tools
Display Modes

Posting Rules

Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are On

Similar Threads
Thread Thread Starter Forum Replies Last Post
VPD Performance anirban1982 Database Administration 2 July 14th, 2009 09:23
Performance testing nikesh mohanlal Performance, Scalability and HA 1 December 16th, 2008 18:18
Performance Degradation Senthil Nathan M Performance, Scalability and HA 1 November 24th, 2008 17:05
Reports Performance smroe@pella.com Reports / BI / FTI and Other Analytics 8 October 2nd, 2008 15:22
VPD performance cool_brat Security 2 August 30th, 2008 12:22



All times are GMT. The time now is 17:15.
Copyright © 2006-2009, Open Book Solutions LLC. All rights reserved.


Inactive Reminders By Icora Web Design

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40