I work with a medium/large organisation which has around 450 sites to 15000 workstations. We use MS Exchange and Outlook 2003 with Cache Mode enabled by default on all workstations. We have a centralised Exchange environment and to date we have been beating our heads against a brick wall trying to solve a serious WAN link performance issue. Whilst Outlook in Cached mode works well for most reasons it is absolutely crippling our environment for the following reason,
About 60-70% of our users are stationary but the other 30% or so travel to other sites or hot-desk within the same office. This being the case means that the Exchange Mailbox is downloaded/synchronized to every workstation that the user logs onto to. The average size mailbox = 1GB with some as high as 5-8GB.
To date, my login scripts are setup to detect if the user is logging onto their "Default" site. So having gotten this far I am then able to manipulate end user behaviour (and for most cases this is true)........except for MS Outlook. I was hoping that I could control Outlook Cache mode by manipulating the registry but it seems I have come undone here.
My login scripts create 2 variables
If the value of USERLOC and WKSLOC are the same I want to be able to turn Mailbox Caching ON
If the value of USERLOC and WKSLOC are NOT the same I want to be able to turn Mailbox Caching OFF
Here is something to think about, If I logon to a PC that I have never used before Outlook will create a brand new profile (backs up any old profile) in registery and starts caching the mailbox over the WAN. Remember, my login scripts have already run....Too late
OK..How about PRF files in append mode. Can I intruct Outlook via PRF files on how it is going to behave for that user's session without scrambling the user's existing setup for example signature and address history?
There are infact 2 settings that must be changed in order to disable Outlook Cache Mode,
1. Use Cached Exchange Mode
2. Send/Receive - Make this folder available offline
Disabling setting 1 does not turn off setting 2...check your packet scans
QUESTION Time: Has anybody been successful in putting together a solution for this type scenario?