I was in that session as well, and Hynek replied very good on that. That specific insert might run 50% faster (which is 1.5 times faster, not 2 times faster), but you must be careful with as statement like that, and don’t take it out of proportion. If your transaction contains 10% of that specific insert and 90% other stuff, then only the part of the 10% runs faster … .
If you’re talking about 5.0SP1 and indexed views replacing sift, I didn’t get this during Hynek’s session, but I did in “Meet The Experts” from Jesper Falkebo himself. I didn’t get if this was going to be a complete replacement though … this was not “literally” said.
It was not Hyneks session, it was the other one.
I don’t think you did David. But that may be because I am seeing red in front of my eyes.
I am amazed this information did not make it into any of the blogs I read from Convergence from you guys.
Am I alone in thinking this is a huge change? And worried that with MS’s tendency to test things in Cronus this could have some horrible performance effects?
I mean look at the outcry at MS adding Index Hinting which is a functionality that can be easily disabled. And no one reacts to this change?
I have 3 customers with dbs over 100GB and Value entry tables with 7-20 million entries and months put in to make them work at an acceptable speed. I have hard time getting clients to pay for this kind of work (you know Navision costs x$ but you will have to pay y$ to make it work with your volume of transactions). How would we be able to justify another round of SQL tuning?
Thing is as it has been proven by the SQL experts out there Navision can be tuned pretty well with current technology. Yes MS has not incorporated too much of this in the std db for now but we all thought it was a matter of time.
Oh well … whatever …
Sometimes I feel like writing some rants like NavStudent from Mibuso [:@]
Well I must say that Eric does listen to this stuff closer than I do. So may be we will get an option. In fact I even asked “Will we have an option MaintainIndexedViews, just like we now have MaintainSIFTfields” and was told no, but who knows. if we got both, this would be a dream scenario.
Please no, we are a civilized bunch here. [:$]
Don’t worry I won’t … I counted to 10 and then went to watch a little more of this crappy Giants vs. Dolphins game. Damn England and their rain.
Let’s balance also this statement [:)]
Most read-statements should come from the DBcache. So disk IO should MOSTLY be for writing.
Current plan is to deliver the “Indexed Views” to maintain SIFT in 5.00 SP1. These changes should also be on the next major release “6.0”.
There are no plans on keeping the current SIFT triggers together with Indexed Views, but we are in close collaboration with the SQL team to address any possible issue.
Also, on top of lab tests, we are working with partners interested in Beta testing the bits, so that we can see the results on a live environment.
Hi Jorge,
Thanks for the info. How can my company apply for the beta testing? I have 3 customers with high transaction volume (2 with dbs > 150GB). Also the newest one has been live since July 2007 and already has 7 million Value Entries. So I am sure you can understand my concern about performance and ability to sustain them as Navision moves to the new technologies.
Regards,
Cristi Nicola
I should have used my words correctly. Currently 5.00 SP1 does not have a beta program nor a TAP program.
As far as I know, MVP’s, which have signed a NDA will get an alpha release.
Regards,
Jorge
Hi Jorge,
Your posts confer no rights anyway [:)] so don’t worry about choice of words. It is still great to hear from MS.
P.S. If you keep making big strides like this with every release that you give to the MVPs you might consider getting bodyguards for them. You never know when one of them might disappear and get sent away to a country that has no problems with torture [:P]
Where are you getting that information?
Jorge is a Microsoft employee, so my guess he should know… [;)]
We’ll see what happens, it would be a first for the MVP’s to get an alpha release. We didn’t even know there was going to be a SP for 5.0 until the public announcement at Convergence.
A little bit off topic.
But does this service pack change anything in regards to their support of the Native database?
Mind you the way the MVPs were treated at convergence [:'(] , did not give me ANY confidence what so ever that the Dynamics Team care much about our opinions.
Well based on today’s mail from Ilana (for MVP’s only) then I don’t think that they have given up MVP’s.
Just a little side note from an end user, of the native database.
If Microsoft can duplicate the speed and efficiencies of an unmodified Navision 2.0 program in Nav 6 with SQL we might upgrade. But from my experience, each new version has gotten significantly slower, and requires more tuning (read more money to developer ) to get the program running.
Just reading this post I can see moving to SQL is not for the faint of heart. and not for a mid size company that Navision used to target.
the move to 2.6 from 2.0 saw a performance hit, then 3.1 was slower then 2.6, and 4.0 slower then 3.1 - 3.6/3.7
We made the official move from 2.0 to 4.0, and boy was that a performance hit. testing on 2.6, 3.1 and 3.7 all discouraged upgrading because of the slower performance. And the database on the upgrade from 2 to 4.0 doubled in size just with the upgrade, before a single transaction was posted.
So, you guys with a 100gig database, would probably only have 50 gig database on a Native 2.0 system, You would probably get better performance downgrading to 2.0 and custom programming your changes in that database.[;)]