| Would this enhancement request be reasonable? |
| Yes, and it should be given high priority |
|
0% |
[ 0 ] |
| It would be nice but not necessary |
|
66% |
[ 6 ] |
| I don't know or care |
|
22% |
[ 2 ] |
| I don't understand |
|
0% |
[ 0 ] |
| What a horrible and dangerous idea |
|
0% |
[ 0 ] |
| I don't like it |
|
11% |
[ 1 ] |
|
| Total Votes : 9 |
|
| Author |
Message
|
| Esa |
 Posted: Fri Jan 27, 2012 12:09 am Post subject: MQHeader enhancement |
 |
|
Grand Master
Joined: 22 May 2008
Posts: 1387
Location: Finland
|
Hi all,
I'm planning to send an enhancement request on MQHeader node (applies to HttpHeader and JMSHeader, too)
Header nodes deep copy the message root before making changes to the headers. This has a certain performance impact: it allocates memory for the new copy of the message and forces the message body to be fully parsed.
I would like to make it possible to optimaze the node performance by adding an 'optimize' checkbox to the node(s). It could be placed in 'Advanced' tab that would also need to be added. Feel free to suggest a better name or location.
Selecting 'optimize' checkbox would authorize the node to 'do the unspeakable': modify the headers directly in InputRoot.
Improperly used this option may cause the message that is propagated to the input nodes catch path to have the modified header. That's why I would like to make it a design question with this switch rather than suggesting some automatic intelligent optimization.
Should this property be overridable? By LocalEnvironment overrides, certainly. But if you allow overriding it in the bar file, it may cease to be a design choice. I can imagine an administrator systematically optimizing the flows...
But there can be things that I have overlooked. This is why I'm asking your opinion with this poll. |
|
| Back to top |
|
 |
| lancelotlinc |
| Posted: Fri Jan 27, 2012 5:46 am Post subject: |
|
|
Jedi Knight
Joined: 22 Mar 2010
Posts: 4941
Location: Bloomington, IL USA
|
I don't see where WMB performs poorly. While your point about optimizing the way some activities occur in the runtime is worthwhile, before you submit the enhancement request, you may like to consider demonstrating the problem with a use case.
Therefore, create a message flow that demonstrates the current issue, and measure its performance on two or three platforms. For example, one million messages processed on z/OS takes x seconds, and on AIX takes y seconds. Then suggest what the target is supposed to be and how to project that performance impact improvement.
IMHO, without these data, your enhancement request would be imcomplete since no goal has been defined so the enhancement can be pronounced successful or unsuccessful.
_________________
http://leanpub.com/IIB_Tips_and_Tricks
Save $20: Coupon Code: MQSERIES_READER |
|
| Back to top |
|
|
| Vitor |
| Posted: Fri Jan 27, 2012 5:52 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
While parts of broker could be optimized as suggested here, it's also worth remembering that the most common cause of poor broker performance is poorly written user code.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| Esa |
| Posted: Fri Jan 27, 2012 6:06 am Post subject: |
|
|
Grand Master
Joined: 22 May 2008
Posts: 1387
Location: Finland
|
| Vitor wrote: |
| While parts of broker could be optimized as suggested here, it's also worth remembering that the most common cause of poor broker performance is poorly written user code. |
One of the main goals of this initiative is to facilitate development of patterns that would perform well regardless of message size.
And pattern development is a wonderful way to minimize the possibility of poorly written user code. |
|
| Back to top |
|
|
| Vitor |
| Posted: Fri Jan 27, 2012 6:15 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| Esa wrote: |
| pattern development is a wonderful way to minimize the possibility of poorly written user code. |
It's a laudible goal of which I'm fully in favour.
My point was simply that no matter how performant the components (or the pattern) it's not the typical bottleneck. That's in the business tranformation & routing logic you can't provide in a pattern. You can provide the most wonderful pattern, with fully optimized components, and a ham fisted developer can eliminate all those advantages in a single module.
This doesn't make what you're trying to do a bad thing. I'm just pointing out it's not the entire solution & more for the beneifit of future readers who've found this thread by typing "optimize message broker" into Google.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| Esa |
| Posted: Fri Jan 27, 2012 6:39 am Post subject: |
|
|
Grand Master
Joined: 22 May 2008
Posts: 1387
Location: Finland
|
| Vitor wrote: |
My point was simply that no matter how performant the components (or the pattern) it's not the typical bottleneck. That's in the business tranformation & routing logic you can't provide in a pattern. You can provide the most wonderful pattern, with fully optimized components, and a ham fisted developer can eliminate all those advantages in a single module.
|
Vitor, the answer to your question is 42.
Lancelot, do we have to build a universe of test environments to get the same answer?
Small is beautiful. My modest wish is just to help to make this little node a little bit more beautiful. |
|
| Back to top |
|
|
| lancelotlinc |
| Posted: Fri Jan 27, 2012 6:42 am Post subject: |
|
|
Jedi Knight
Joined: 22 Mar 2010
Posts: 4941
Location: Bloomington, IL USA
|
No, but it helps to have at least two different readings from two different platforms. If nothing else, Windows and Linux.
_________________
http://leanpub.com/IIB_Tips_and_Tricks
Save $20: Coupon Code: MQSERIES_READER |
|
| Back to top |
|
|
| Vitor |
| Posted: Fri Jan 27, 2012 6:59 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| Esa wrote: |
Vitor, the answer to your question is 42.
|
No, that's the answer to a far more ultimate question.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| smdavies99 |
| Posted: Fri Jan 27, 2012 7:16 am Post subject: |
|
|
Jedi Council
Joined: 10 Feb 2003
Posts: 6076
Location: Somewhere over the Rainbow this side of Never-never land.
|
| Vitor wrote: |
| Esa wrote: |
Vitor, the answer to your question is 42.
|
No, that's the answer to a far more ultimate question. |
The mice might have something to say about that....
Happy bunny here. I've just got tickets to see the original Radio Cast of HH2TG performing it at Basingstoke next month.
_________________
WMQ User since 1999
MQSI/WBI/WMB/'Thingy' User since 2002
Linux user since 1995
Every time you reinvent the wheel the more square it gets (anon). If in doubt think and investigate before you ask silly questions. |
|
| Back to top |
|
|
| Vitor |
| Posted: Fri Jan 27, 2012 7:41 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| smdavies99 wrote: |
| I've just got tickets to see the original Radio Cast of HH2TG performing it at Basingstoke next month. |
The Atlantic ocean has never seems so inconviently large. The original cast? I'm Hulk-like with envy!
Admitedly it's Basingstoke but that would be a sacrifice worth making.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| mqjeff |
| Posted: Fri Jan 27, 2012 8:24 am Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008
Posts: 17447
|
| Vitor wrote: |
| smdavies99 wrote: |
| I've just got tickets to see the original Radio Cast of HH2TG performing it at Basingstoke next month. |
The Atlantic ocean has never seems so inconviently large. |
Surely that's someone else's problem. |
|
| Back to top |
|
|
| Esa |
| Posted: Fri Jan 27, 2012 10:51 am Post subject: |
|
|
Grand Master
Joined: 22 May 2008
Posts: 1387
Location: Finland
|
| smdavies99 wrote: |
| The mice might have something to say about that.... |
The mice said they like my idea. |
|
| Back to top |
|
|
| zpat |
| Posted: Fri Jan 27, 2012 11:54 am Post subject: |
|
|
Jedi Council
Joined: 19 May 2001
Posts: 5866
Location: UK
|
Doesn't seem to be an option for Yes, but not top priority.
There's no harm having options, as long as the effect is made clear and if the default behaviour is as today, no-one loses. |
|
| Back to top |
|
|
| lancelotlinc |
| Posted: Fri Jan 27, 2012 12:59 pm Post subject: |
|
|
Jedi Knight
Joined: 22 Mar 2010
Posts: 4941
Location: Bloomington, IL USA
|
| zpat wrote: |
Doesn't seem to be an option for Yes, but not top priority.
There's no harm having options, as long as the effect is made clear and if the default behaviour is as today, no-one loses. |
It's a balance of cost-benefit. It's going to cost money to put any feature in a product. So the loser is the owner of the product. This feature will not generate any new revenue and therefore, if it is decided to move ahead with the idea, the cost for doing so will be at a loss.
_________________
http://leanpub.com/IIB_Tips_and_Tricks
Save $20: Coupon Code: MQSERIES_READER |
|
| Back to top |
|
|
| Esa |
| Posted: Sat Jan 28, 2012 1:56 am Post subject: |
|
|
Grand Master
Joined: 22 May 2008
Posts: 1387
Location: Finland
|
| lancelotlinc wrote: |
It's a balance of cost-benefit. It's going to cost money to put any feature in a product. So the loser is the owner of the product. This feature will not generate any new revenue and therefore, if it is decided to move ahead with the idea, the cost for doing so will be at a loss. |
Yes, that's the common strategy with products nearing the end of their lifetime. I think I can read between your lines. |
|
| Back to top |
|
|
|
|
