Performance and MDEs

[Leslie Isaacs]

Hello All

I am running a complex, split FE/BE A97 db on a peer-to-peer network of 4
W2K PCs, with the BE on a W2K Server machine. I have read a little about
MDEs, and was wondering what would be the performance effect (if any) of
deploying MDEs on the 4 workstations instead of MDBs. Is this necessarily a
good idea - and is there a downside?

I would be grateful for any advice.

Many thanks
Leslie Isaacs

 

[SusanV]

I only give users MDEs - so they can't modify forms or reports or poke
around in the modules or any of that. Or open the VBE and remove or add
references. Or whatever.... A lot less problems since doing that (I have a
few users who know just enough to be dangerous - once someone deleting a
module! Thankfully I backup nightly!)

The mdb is for development only, users cannot access it.

DB is split, and performance seems the same with either mdb or mde.

 

[Douglas J Steele]

I'm assuming you're talking about the front-end only. There's no reason to
make the back-end an MDE.

 

[SusanV]

Yes of course - sorry if that wasn't clear

;-)

I'm assuming you're talking about the front-end only. There's no reason to
make the back-end an MDE.

 

[Albert D.Kallal]

Hello All

I am running a complex, split FE/BE A97 db on a peer-to-peer network of 4
W2K PCs, with the BE on a W2K Server machine. I have read a little about
MDEs, and was wondering what would be the performance effect (if any) of
deploying MDEs on the 4 workstations instead of MDBs. Is this necessarily
a good idea - and is there a downside?

I would be grateful for any advice.

The compiled mde does strip out the source code, and eliminates any runtime
compiling of code So, it does run faster. However, it been about 8+ years
since ms-access has been processing bound. If you double the speed of the
processor, you have not changed the network speed, or the disk drove
speed...which is where ms-access spends all it s time waiting.. It is always
waiting for the OS (disk drive, network etc), so more processing will
generally not fix much.

However, if your mdb does become un-compiled, then some noticeable delays
can occur as the mdb re-compiles....(and bloats along at the same time). So,
some nasty delays can be eliminated. And, by using a mde, you eliminate the
track-auto name correct issues, and another of other got-ya's that can
impact performance. So, using a mde just forces to ensure you are
distribution an application in a nice clean state....

Don't forget to keep a persistent connection open...it will make your
application fly.....


if you mdb is compiled, then raw processing speed is VERY close to that of a
mde. however, why take a chance of your application having compile errors,
and un-compiled code?