[e2e] What should e2e protocols know about lower layers?

Joe Touch touch at ISI.EDU
Wed Oct 10 13:59:59 PDT 2001


Bob Hinden wrote:

> Joe,
> 
>> As Ran observed on the TSVWG mailing list, IPv6 lacks broadcast, one 
>> could argue for reasons related to this. However, replacing the 
>> well-known term broadcast with the magical 'anycast' doesn't replace 
>> the need.
> 
> 
> IPv6 anycast does not replace broadcast.  Anycast is the delivery of a 
> packet to one of a set of destinations.  IPv6 uses multicast instead of 
> broadcast.  See RFC2373.  Also I never though of anycast as "magical"....


Calling it "multicast" doesn't make it any less magical or less 
dependent, in a shared link environment, on broadcast. RFC2373 refers to 
an ID (yikes) on the matter, notably the ID which later became RFC2464, 
which maps IPv6 multicast onto ethernet multicast.

At some point you need a MAC address. How do you divine these on shared 
media, except using broadcast?

FYI, RFC2461 says:

  6.3.3.  Interface Initialization

    The host joins the all-nodes multicast address on all multicast-
    capable interfaces.

Perhaps you'd like to share how "all-nodes multicast" is substantially 
different from broadcast, except that it reaches beyond the local link?
(i.e., it's MORE than broadcast, not less).

>> The LAN is an important component of the current architecture, and 
>> differentiates between local and nonlocal for the use of broadcast to 
>> avoid centralization of MAC address registration, and the alternative 
>> of explicit dispersed registration.
> 
> I don't think it is correct to assume that "local" means a LAN. 


I was asserting the reverse, that LAN means 'local'. Of course, there 
are pt-pt links for which it would be useful to consider the locality, 
if, as you mention later, it could be deduced.

Joe




More information about the end2end-interest mailing list