| Author |
Message
|
| PeterPotkay |
 Posted: Fri Jul 15, 2011 9:55 am Post subject: |
 |
|
Poobah
Joined: 15 May 2001
Posts: 7722
|
| lancelotlinc wrote: |
I am a big proponent of open source, however, comparing ServiceMix to WMB is comparing a Cessna to the Space Shuttle.
|
If all you need to do is fly a couple of hundred miles on a calm, spring day....
_________________
Peter Potkay
Keep Calm and MQ On |
|
| Back to top |
|
 |
| lancelotlinc |
| Posted: Fri Jul 15, 2011 10:32 am Post subject: |
|
|
Jedi Knight
Joined: 22 Mar 2010
Posts: 4941
Location: Bloomington, IL USA
|
| PeterPotkay wrote: |
| lancelotlinc wrote: |
I am a big proponent of open source, however, comparing ServiceMix to WMB is comparing a Cessna to the Space Shuttle.
|
If all you need to do is fly a couple of hundred miles on a calm, spring day.... |
Yes, very true. I used to own a Cessna 172, four seats, nice plane. Mine was a 1978 model, tail number N3475E . My first annual inspection cost me $8,700 . I fell over. The second annual inspection was no better. Then my mechanic was using my plane without my consent, and didn't refill the fuel. I sold it on the third year.
To fly from Houston to San Antonio, with a Cessna 172 costs 10 gallons per hour. An automobile (2010 Honda Civic EX-L) burns two gallons per hour. Either mode has limitation, neither will get you into orbit to deliver supplies to the space station.
_________________
http://leanpub.com/IIB_Tips_and_Tricks
Save $20: Coupon Code: MQSERIES_READER |
|
| Back to top |
|
|
| Vitor |
| Posted: Fri Jul 15, 2011 10:39 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| lancelotlinc wrote: |
| To fly from Houston to San Antonio, with a Cessna 172 costs 10 gallons per hour. An automobile (2010 Honda Civic EX-L) burns two gallons per hour. Either mode has limitation, neither will get you into orbit to deliver supplies to the space station. |
So you need to factor into the decision how fast you need to get from Houston to San Antonio, and the probability you'll one day visit the space station on a regular basis (which means you need to buy a space shuttle or update the maps on your GPS if you thought you were still going to San Antonio)
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| bprasana |
| Posted: Mon Jul 18, 2011 8:19 am Post subject: |
|
|
Disciple
Joined: 18 Apr 2005
Posts: 179
|
Thanks for all the reply guys.
This was basically triggered by a passing comment from one of the clients here. who was saying " Biztalk is faster for in memory processing than WMB"
There is requirement to validate employee code with the list of codes that gets returned when we query on his department.
A user enters a userid and password, we determine the exact department code from LDAP.(its funny - his emp code is not present in LDAP!!).
In WMB, we are supposed to make a http call based on his department code and validate his code against the set of returned codes.(maximum 20).
The http service itself queries a database which has about 8000 departments and 40000 employees in total. But it takes about 2s to return the call.The data is highly volatile. people keep changing departments every week.
We want to reduce this time.
One of the suggestions was to load the whole data "in memory" every day for processing.
Are there any other suggestions
regards,
bprasana |
|
| Back to top |
|
|
| anveshita |
| Posted: Mon Jul 18, 2011 8:53 am Post subject: |
|
|
Master
Joined: 27 Sep 2004
Posts: 254
Location: Jambudweepam
|
| bprasana wrote: |
Thanks for all the reply guys.
This was basically triggered by a passing comment from one of the clients here. who was saying " Biztalk is faster for in memory processing than WMB"
There is requirement to validate employee code with the list of codes that gets returned when we query on his department.
A user enters a userid and password, we determine the exact department code from LDAP.(its funny - his emp code is not present in LDAP!!).
In WMB, we are supposed to make a http call based on his department code and validate his code against the set of returned codes.(maximum 20).
The http service itself queries a database which has about 8000 departments and 40000 employees in total. But it takes about 2s to return the call.The data is highly volatile. people keep changing departments every week.
We want to reduce this time.
One of the suggestions was to load the whole data "in memory" every day for processing.
Are there any other suggestions
regards,
bprasana |
R u hijacking the thread?  |
|
| Back to top |
|
|
| exerk |
| Posted: Mon Jul 18, 2011 8:55 am Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
| anveshita wrote: |
| R u hijacking the thread? |
As it's a 6 year-old thread, I'd say not 
_________________
It's puzzling, I don't think I've ever seen anything quite like this before...and it's hard to soar like an eagle when you're surrounded by turkeys. |
|
| Back to top |
|
|
| Vitor |
| Posted: Mon Jul 18, 2011 9:01 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| exerk wrote: |
| anveshita wrote: |
| R u hijacking the thread? |
As it's a 6 year-old thread, I'd say not |
Especially given who re-opened it after 6 years (how many threads in here can legitimately be reopened after that length of time? This is one for the diary.)
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| bprasana |
| Posted: Mon Jul 18, 2011 10:15 am Post subject: |
|
|
Disciple
Joined: 18 Apr 2005
Posts: 179
|
No I am not hijacking.
someone earlier mentioned that 'it depends on your requirements'. So i provided more details. that's all! |
|
| Back to top |
|
|
| fjb_saper |
| Posted: Mon Jul 18, 2011 12:34 pm Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20763
Location: LI,NY
|
| bprasana wrote: |
There is requirement to validate employee code with the list of codes that gets returned when we query on his department.
A user enters a userid and password, we determine the exact department code from LDAP.(its funny - his emp code is not present in LDAP!!).
In WMB, we are supposed to make a http call based on his department code and validate his code against the set of returned codes.(maximum 20).
The http service itself queries a database which has about 8000 departments and 40000 employees in total. But it takes about 2s to return the call.The data is highly volatile. people keep changing departments every week.
We want to reduce this time.
One of the suggestions was to load the whole data "in memory" every day for processing.
Are there any other suggestions
regards,
bprasana |
Well there may be several ways. If the LDAP is up to date you might be able to access it, either by name or Id. The other question would be to look into DB efficiencies. Using an http call may not be the fastest way (MQ may be faster as a transport) but my guess is the crux is in the DB and the way the data is accessed. The right index with a prepared query can work wonders. 
_________________
MQ & Broker admin |
|
| Back to top |
|
|
| bprasana |
| Posted: Mon Jul 18, 2011 4:27 pm Post subject: |
|
|
Disciple
Joined: 18 Apr 2005
Posts: 179
|
@fjb_saper : since its about 40K rows with 2-3 columns, can we not cache it in the memory?
Does WMB support that? |
|
| Back to top |
|
|
| bprasana |
| Posted: Mon Jul 18, 2011 4:55 pm Post subject: |
|
|
Disciple
Joined: 18 Apr 2005
Posts: 179
|
|
| Back to top |
|
|
| fjb_saper |
| Posted: Mon Jul 18, 2011 7:48 pm Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20763
Location: LI,NY
|
| bprasana wrote: |
ftp://public.dhe.ibm.com/software/integration/support/supportpacs/individual/ia91.wbimb.pdf
found this on google. |
Well caching is certainly one way to go. However be aware that the cache is mainly designed for static objects. From what you said earlier your change rate is quite high and your objects are more dynamic than static.
This would call more for a DB cache type than a broker cache type. If an outside source needs to modify the objects it will still write them to the DB and hence to the DB cache. As such a request made against the objects would seldom hit the disk, and be handled in DB Cache memory...
The speed difference between disk retrieval and DB Cache in memory retrieval should be significant enough to satisfy your needs. Note that between those 2 options only DB settings change. No need to change any source code.
Have fun
_________________
MQ & Broker admin |
|
| Back to top |
|
|
| zpat |
| Posted: Mon Jul 18, 2011 10:33 pm Post subject: |
|
|
Jedi Council
Joined: 19 May 2001
Posts: 5866
Location: UK
|
Biztalk is slower than WMB, because everything goes through the SQL server database on the way in and out.
Biztalk is more like a process server for workflow. |
|
| Back to top |
|
|
| lancelotlinc |
| Posted: Tue Jul 19, 2011 4:38 am Post subject: |
|
|
Jedi Knight
Joined: 22 Mar 2010
Posts: 4941
Location: Bloomington, IL USA
|
| bprasana wrote: |
ftp://public.dhe.ibm.com/software/integration/support/supportpacs/individual/ia91.wbimb.pdf
found this on google. |
Support pac IA91 is the last resort for caching WMB infos. You will fair better using Singletons and HashMaps.
Vitor said: Singletons....! 
_________________
http://leanpub.com/IIB_Tips_and_Tricks
Save $20: Coupon Code: MQSERIES_READER |
|
| Back to top |
|
|
| Vitor |
| Posted: Tue Jul 19, 2011 4:50 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| lancelotlinc wrote: |
| Support pac IA91 is the last resort for caching WMB infos. You will fair better using Singletons and HashMaps. |
There you are! Got worried when this didn't evoke a reaction...
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
|
|
