[e2e] Transparent subnets with Proxy Arp

[Bob Braden]

  *> The text from RFC 1127 below seems to have a specific
  *> phenomenon in mind, but it's not clear to me what that is.
  *> Does it mean that expiry on entry from  the cache will
  *> cause a new ARP exchange and additional latency which
  *> then affects TCP RTO?  Or is there something more "subtle"
  *> than that?
  *> 
  *> Regards,
  *> 
  *> - Eric
  *> 

Eric,

Based on what I wrote, I would assume that the mechanism is as you
describe.  The mechanism is simple, but TCP's reaction to such
perturabations are not.  See PILC.  (Of course, in 1989 we did not
understand TCP performance as completely as we do in 2001).

Bob Braden

  *> ----- Original Message -----
  *> From: "Bob Braden" <braden at ISI.EDU>
  *> To: <end2end-interest at postel.org>; <eric23 at softhome.net>
  *> Sent: Thursday, September 06, 2001 5:46 PM
  *> Subject: Re: [e2e] Transparent subnets with Proxy Arp
  *> 
  *> 
  *> >   *>
  *> >   *> RFC 1127 ("Perspective on host requirements"), section 4, item 2:
  *> >   *>
  *> >   *>     Even without proxy ARP, the management dynamics of the IP route
  *> >   *>     cache interact in subtle ways with transport-layer dynamics;
  *> >   *>     introducing routing via proxy ARP brings a third protocol layer
  *> >   *>     into the problem, complicating the inter-layer dynamics still
  *> further.
  *> >   *>
  *> >   *>
  *> >   *> - Eric Goldman
  *> >   *>
  *> >   *>
  *> >
  *> > Oh, OK.  And your question was... ?
  *> >
  *> > Bob Braden
  *>