[dhcwg] Re: Last call for <draft-ietf-dhc-csr-05.txt>

> I write a DHCP client, and would like it to request both options, but I 
> expect server administrators will usually configure their servers to 
> return only one or the other, as appropriate for the particular site.
> 
> Ted writes a DHCP server, and would like it to be configured with data 
> for both options, but expects client administrators to configure their 
> clients to only request one or the other, as appropriate for the 
> particular site.

There is actually a good reason to put complexity in the client - it
distributes the work more fairly.  If you do the work on the server,
the workload increases by O(n), where n is the number of clients.
If you do it on the client, each client's workload increases by O(1).

I think in this case, with your proposed changes, there is no
additional workload for the server, so I'm happy to leave it that
way.   But as a general rule, when we are making decisions like this,
I think if we *can* push the work to the client, we should.

BTW, I am both a client and a server implementor - I am very fond of
the ISC DHCP client, which I use constantly as I roam from network to
network.  :')

			       _MelloN_

_______________________________________________
dhcwg mailing list
dhcwg@ietf.org
http://www1.ietf.org/mailman/listinfo/dhcwg