On Fri, Mar 21, 2003 at 04:33:51PM +1200, Gareth Williams wrote:
> One more question, if I may - I thought bridges were switched (at the ethernet 
> layer). You imply that a bridge is simply a 2 port hub (and can be switched 
> or unswitched/broadcast, depending on complexity). I got the idea somewhere 
> that a bridge's intended function was different from a hub...
> ie. hub -> join end terminals together
> bridge -> join networks together
> Now I think about it for a couple of seconds, the actual functionality is 
> pretty much identical, no? So... is then "hub" just another word for "bridge" 
> (ie. they are essentially the same device) ?

Yes, they are the same ... a bridge was originally conceived as a
two-port device, but when manufacturers figured out to put multiple
ports on the devide, they started to look for new names. And as they had
effectively created a star-topology network from Ethernet, the name
'hub' worked for them :-)

> ps. still not sure I "get" NAT (isn't it supposed to save address space? how's 
> that, if it's 1-to-1 direct mapped..?) ... but I'll google for that one :-)

Network Address Translation does just what it says on the box -
translates network addresses :-) That's defined as a 1-to-1 exchange.
Useful in places like firewalled environemnts, where you need to ensure
that hosts can only talk to each other via the mediation of the
firewall. It's also useful whjere you are joining two previously
unrelated networks, that have both been configured to use the same
ranges (i.e. all my private networks are 10.n.n.n) - you can put one NAT
box between them, instead of renumbering the whole of one of your
networks.

IP Masquerade (source address NAT) is the space-saving mechanism you're
thinking of ...

-jim

Reply via email to