On Thu, Aug 03, 2006 at 02:28:21PM -0400, James Carlson wrote: > Nicolas Williams writes: > > In any case, it seems to me that if forwarding is enabled then routing > > should be enabled and vice-versa, > > I strongly disagree with that. > > Forwarding is the act of receiving packets on one interface and > transferring them to another interface. Routing is the process of > computing feasible routes and in this case, "enabling routing" means > turning on one or more dynamic routing protocols.
I was including static routing in my conception of what the "routing" service would manage. > > so the split into transient (right?) > > forwarding services and non-transient (right? what if only static > > routes are configured) routing services seems artificial. > > If it's not "enabling routing protocols," how do you name the > distinction between running a dynamic protocol and using only static > routes? I think I'd rather configure both, dynamic and static routing in one service, rather than in two, what, mutually exclusive services. Nor is it clear that one couldn't have a routing daemon even when all it's configured to do is install some static routes and then wait (for the sysadmin to change the configuration, essentially). In any case, the post I was replying to didn't, IIRC, mention static routes, whereas I did mention them. And if we'd not have any daemons lying around if only static routes were configured then I'd not like to end up with more services than necessary due to the need to choose transient vs. non-transient-ness for services upfront. Nico --
