I have kinda posted this as a reply of a 'Me Too' fashion. But here's my
own...I have an open support case with Veritas and I'm dealing with a
call centre in India, which is not the easiest - outsourced and cheap
doesn't always make better!!!

Anyway,I get relatively good backup speed but the restore is painfully
slow, I also get an issue with Media server and target server CPU solid
at 93% during restore making both boxes unusable. Both boxes on a common
gigabit line card and my workstation in the same card (Cisco Cat4), all
devices are set to 1gig F/Duplex.

The data being restored is from a different target server (a redirect),
but the OS and patch levels are the same, along with compression and
volume type.

At the request of the support centre I have:
set tcp delayed acknowledgement=off
set tcp nagle algorithm=off

I have applied the latest beta TSA5UP17 patch to both media and target
servers at the request of support centre.

No change.

Running NW5.1sp7 / BE 9.1 1152

Oh, and the Java GUI on my XP workstation is sooooo slow it's not funny.