Rabid Wombat wrote:
>
> On Fri, 1 Mar 1996, Paul Ferguson wrote:
>
> > The fragmentation and, more importantly, reassembly should happen in this
> > case transparently long before it reaches your router/firewall/whatever.
> >
> > - paul
>
> Yes - segment and re-assembly should occur at the edge devices.
>
> >
> > At 10:28 PM 2/29/96 -0500, Charles B. Kaplan wrote:
> >
> > >>The only time you're ever likely to see a packet with FO=1 is if a bad guy is
> > >>knocking at your door.
> > >
> > >IE, my east coast LAN wants to connect to my west coast LAN, which will
> > >involve traversing (substitute your favorate backbone providers) ATM link.
> > >Therefor my 68byte header + data get dumped into larger (I forget frame size
> > >at the moment) ATM cell, which could POSSIBLY ?? cause one byte to cross a
> > >cell boundry, and thuse appear fragmented to the remote site ?
> > >
> ATM uses 53 byte cells, 48 bytes of payload, 5 bytes header. Much smaller
> than your IP packtes. SAR should occur before reaching your firewall,
> however.
>
> - r.w.Right Lan Emulation makes the 48 byte payload transparent to IP.
References:
|
|
