ADSML (aka Kelly),

>Nothing like seeing a IBMer blame Microsoft....

I don't believe there was any intent to "blame" anyone...it was a simple
statement of fact.
The backup/recovery APIs for Exchange do not support that level of granularity.
Vendors who
provide mailbox or folder level granularity must be using the Messaging API
(MAPI) which is not
intended for backup  and recovery purposes.

>Nope, the "dopeheads" that implemented it says one must
>restore the whole thing.

If you consider us dopeheads for choosing to stay within the Exchange
backup/recovery architecture,
I guess you are entitled to your opinion.  However, I believe we have made the
right choice given the
various issues related to the use of MAPI for backup and recovery at a more
granular level. In addition
to the performance and scalability issue, it is really important to understand
the impact of breaking the
single instance store architecture in Exchange.   To better understand this,
consider the following example:

When a mail message is sent to a distribution list of 20 people, that mail item
is stored in exchange only
once and each of the 20 mailboxes has a pointer to the single instance of the
item.  When MAPI interfaces
are used to extract the mail items for mailbox level backup, then the full mail
item is extracted repeatedly
for each mailbox.  This means that the size of the backup data when stored per
mailbox could be much
larger than the actual size of the IS.   Further, when restoring by individual
mailbox, the mail object will be
stored multiple times so that the resulting IS will be larger than the
original...and may even cause issues
with being able to fit on available storage.

This is probably the main reason why Microsoft recommends against use of the
MAPI for backup and
recovery and also why they are providing alternate mechanisms to address the
issue of individual item
recovery that do not compromise this architecture.

Chris Zaremba
TDP CLient (formerly ADSM Agent) Development
internet  zaremba AT us.ibm DOT com