Large scale NAT - problem resolved

[Julian Elischer]

On Wed, 28 Jan 2004 veedee at c7.campus.utcluj.ro wrote:

> On Wed, Jan 28, 2004 at 10:41:20PM +0200, Ruslan Ermilov wrote:
> > On Wed, Jan 28, 2004 at 12:15:56AM -0800, Julian Elischer wrote:
> > > 
> > > On Wed, 28 Jan 2004, Andriy Korud wrote:
> > > 
> > > > 
> > > > Hi,
> > > > At last I've managed to build stable NAT on FreeBSD box for 34Mbit link and
> > > > ~2000 clients (cable modem network).
> > > > At full speed (34Mbit) CPU usage is 0% and system load is 0.0 :-)
> > > > 
> > > 
> > > It'd be really interesting to see how natd would handle such a load....
> > > 
> > You must be kidding.  ;)
> 
> Agreed. NATd "crashes" with 400 clients on AMD Athlon 900Mhz. :( ipnat
> works fine.
> 
> This raises a question... is there any point in still having natd? (don't
> throw rocks at me please, I'm just asking). Or maybe it's still being used
> for servers with less clients to nat?

Well for people using ipfw.. 
if_nat requires ipfilter

If it 'crashes' that sugests that a bug exists..
anyone know what 'crashes' means? gets slow?
if so then probably using a hash table somehwere would fix it..