Hi Moster,
the job rate figures are seriously confusing with time to load/unload tapes and write catalogs included.
That said, there is a correlation between what is going on (systems busy/out of hours), the type of backup (full/incremental) and the type of thing backed up. Also the speeds of the various boxes running the agents and Bex. Also processing logs, etc for Exchange and SQL. Also snapshot ptocessing for AOFO.
On a W2003 64 box hosting Bex with local SAS storage and DELL PV124T 2 x 8 LTO2 autochanger with local and remote backups over a private gigabit network I get these mb/min figures.
Local disk: Full: 2,216
Remote SQL 2005 + Files: Full 1,200
Remote incremental multi-gig .bkf files: 950
Remote full multi-gb .bkf files: 1,413
Remote incremental User Profiles: 86
Remote full User Profiles: 311
10gb Exchange 2k remote PFolders 2 disk: 1,053
10gb Exchange 2k remote PFolders to tape: 2,080
Other backup and Exchange traffic uses the gigabit network.
Disk fragmentation can easily cause 30% slower backup
A second Bex server shows similar results.
10-30% slower all round in daytime
The 'hotfixes' do seem to affect speed.
Cheers,
PC