RE: [LVS - NAT] alternatives

[Peter Mueller <pmueller@xxxxxxxxxxxx>]

> > NAT is not -bad-. The additional latencies are mostly 
> irrelevant. The only
> > real issue is the sheer number of packets being handled by 
> the network
> > interfaces (as pointed out by Joe already).
> 

one more point I'd like to mention... I have been bit by a few etherexpress
pro / linux network driver issues in the past when there is sustained high
traffic load.  Things like _that_ make me shy away from architectures like
NAT.
 
> Because NAT is handled in the RISC processors. OTOH, try NAT 
> and sticky bit
> set on the Alteon ACEdirector3 and watch huge latencies in 
> server takeout
> and packet dropping. The Cisco had a bug where approximately 
> 10% of all packets
> weren't NAT under high traffic load (I cannot confirm this on 
> the gigabit
> ones).

don't even get me started on local directors (shiver - bad memories).. :P  -
when I was working in the lawson-exodus cage I seem to remember seeing the
hotmail goon squad come in and overclock a bunch of their localdirectors
waaaaay beyond spec.  yeah nice intel PII/PIII's if I can remember correctly
:)