Re: HIP, the Host Identification Protocol
Garrett.Wollman@uvm.edu Thu, 17 June 1993 17:48 UTC
Received: from ietf.nri.reston.va.us by IETF.CNRI.Reston.VA.US id aa10274; 17 Jun 93 13:48 EDT
Received: from CNRI.RESTON.VA.US by IETF.CNRI.Reston.VA.US id aa10270; 17 Jun 93 13:48 EDT
Received: from thumper.bellcore.com by CNRI.Reston.VA.US id aa22038; 17 Jun 93 13:48 EDT
Received: by thumper.bellcore.com (4.1/4.7) id <AA18580> for ietf-archive@nri.reston.va.us; Thu, 17 Jun 93 13:47:09 EDT
Received: from griffin.emba.uvm.edu by thumper.bellcore.com (4.1/4.7) id <AA18559> for /usr/lib/sendmail -oi -fowner-pip X-pip; Thu, 17 Jun 93 13:47:07 EDT
Received: by griffin.emba.uvm.edu id AA11650 (5.65/1.11); Thu, 17 Jun 93 13:47:01 -0400
Date: Thu, 17 Jun 1993 13:47:01 -0400
Sender: ietf-archive-request@IETF.CNRI.Reston.VA.US
From: Garrett.Wollman@uvm.edu
Message-Id: <9306171747.AA11650@griffin.emba.uvm.edu>
To: Paul Francis--formerly Tsuchiya <francis@thumper.bellcore.com>
Cc: pip@thumper.bellcore.com
Subject: Re: HIP, the Host Identification Protocol
In-Reply-To: <9306171503.AA02391@tsuchiya.bellcore.com>
References: <9306171503.AA02391@tsuchiya.bellcore.com>
(Is there anyone else out there on this list who has the slightest interest in this?) <<On Thu, 17 Jun 93 11:03:10 EDT, francis@thumper.bellcore.com (Paul Francis--formerly Tsuchiya) said: >> If the host has no router entry, then it sends the packet as a >> multicast; the routers then hear this multicast and send back proxy >> HELLOs to the sending host (if it is not on the same LAN segment); > SHouldn't be necessary. When a host comes up, it sends out the > router discovery solicitation message, and learns of routers. > There should be no case where a host has a packet to send and > doesn't know of a router. Aha! I see the confusion! Remember that, in my model, there are /no/ intra-area redirects. The host does not want to send the packet to /a/ router, it wants to send the packet to /the/ router (i.e., the router which will actually do the forwarding). >> Remember that /all/ routers which do intra-area routing are considered >> responsible. Here's an example of how we would do load-balancing... > Then, what is an example of a not-responsible router? After all, > all routers are running intra-area routing..... No. For example, the entry router for a domain should know about inter-area routing, but if it is only connected to a single physical subnetwork of an area, then it is not responsible in that area and so is completely oblivious to intra-area routing (or rather, it knows no more about it than a host on the same wire does). In other words, it participates in the level-[1/2] routing (that is its function), but not the level-0 routing. Note the interaction with router discovery here. Because router discovery is on a per-LAN basis, the hosts on the same LAN as our ``irresponsible'' router are able to discover it in the normal way; however, hosts on a different LAN in the same area won't see it directly, but rather their intra-area router (which /does/ have to participate in level-1 routing) will advertise itself on behalf of the original router. > Huh? The preferences are on a per-LAN basis. No, they're not, that's the whole point! (As I said in the original proposal: it bears some resemblance to ESIS, but it most emphatically is NOT the same!) > How can a router give certain preferences to some hosts and not > others....the router is only advertising one hello message, and it > has a preference that all hosts record..... No. The router doesn't advertise a preference in the router HELLO, only in the proxy HELLO. > I don't think it is good for routers to be sending hellos individually > to hosts, except in the case where a host has just booted and needs > quickly to know which routers are there, in which case it does the > solicitation and the routers respond...... Well, that's one of the major difference between HIP and ESIS. (One of the reasons why I call proxy HELLOs by that name is because of the similarity to proxy ARP in IPv4.) > Of course, but the routing protocol advertises the hosts, ala ISIS. > Routers only need learn of hosts on their attached LANS via hellos. > Routers learn of hosts not on their LAN but in their area via the > routing protocol..... Intra-area routers, yes. Non-intra-area routers, no. -GAWollman PS: I should point out that this is just my proposal; anyone else who wants to see areas in Pip is free to create his/her own proposal and protocols that do the same thing. -- Garrett A. Wollman | Shashish is simple, it's discreet, it's brief. ... wollman@emba.uvm.edu | Shashish is the bonding of hearts in spite of distance. uvm-gen!wollman | It is a bond more powerful than absence. We like people UVM disagrees. | who like Shashish. - Claude McKenzie + Florent Vollant
- Re: HIP, the Host Identification Protocol Paul Francis--formerly Tsuchiya
- Re: HIP, the Host Identification Protocol Garrett.Wollman
- Re: HIP, the Host Identification Protocol Paul Francis--formerly Tsuchiya
- Re: HIP, the Host Identification Protocol Garrett.Wollman