Hi Joe,

> -----Original Message-----
> From: Joe Touch [mailto:[email protected]]
> Sent: Monday, October 10, 2016 10:29 AM
> To: t.petch <[email protected]>; Templin, Fred L <[email protected]>
> Cc: [email protected]
> Subject: Re: [Int-area] Questions to draft-intarea-tunnels-03
> 
> Hi, Tom,
> 
> 
> On 10/9/2016 3:05 AM, t.petch wrote:
> > ----- Original Message -----
> > From: "Joe Touch" <[email protected]>
> > Sent: Saturday, October 08, 2016 12:24 AM
> >
> >
> >> Hi, Tom,
> >>
> >> Thanks for pointing that out. I'll try to revise to make that easier.
> >>
> >> If you have any other specific examples, please let me know (either
> >> on-list or directly).
> > Joe
> >
> > How about
> >
> > '   o  Tunnel internal MTU (TIMTU): the largest message that a tunnel
> >       egress can emit into a tunnel without requiring further
> >       fragmentation to reach the tunnel egress.' ?
> >
> > You may have to read it twice to see my confusion.
> 
> Yes - that's a typo that should read as follows, which should be more clear:
> 
>    o  Tunnel internal MTU (TIMTU): the largest message that a tunnel
>       ingress can emit into a tunnel without requiring further
>       fragmentation to reach the tunnel egress.
> 
> 
> > More generally, I think that Fred's point is well made, that the
> > terminology, particularly the abbreviations thereof - e.g. LMTU, TIMTU,
> > TMTU, PMTU, ERMTU, RMTU - make the I-D harder to follow.
> 
> The challenge is that it's nearly impossible to discuss these issues
> without coining new terms because the existing ones are so poorly
> understood in the context of layering.
> 
> We generally care about many different MTUs:
>     - the largest IP packet that can traverse an IP path
>         this is NOT the path MTU, but rather the receiver reassembly MTU
>         we CANNOT transmit transport segments larger than the payload of
> this IP E2E MTU
> 
>     - the largest IP packet that can traverse an IP path without needing
> on-path fragmentation
>         this is the path MTU
>         it is critical for IP source fragmentation (or the packets won't
> get there at all)
>         it's an optimization for transports that frag/reassemble (to
> avoid source fragmentation)
> 
> Neither of these has anything to do with the link MTU, which has similar
> variants.
> 
> As we keep discussing this issue, we've come up with various notions of
> what these names could be.
> 
> We could call these the:
>     IP transit MTU  (defined as the source-to-dest MTU, but equal to the
> IP receiver reassembly MTU)
>     IP path MTU (min of the IP hop MTUs)
>     IP hop MTU = link/tunnel transit MTU - link/tunnel overhead
> 
>     link (or tunnel) transit MTU (defined as source-to-dest MTU, equal
> to link/tunnel receiver reassembly MTU)
>     link path MTU (min of the link hop MTUs)
>     link hop MTU = MTU of a single physical link
> 
> I'm glad to revise if these are more clear.

IMHO, your draft is getting wrapped up with too many new acronyms for no
good reason when simple English-language text would make things easier to
understand. See Section 3.12 of 'draft-templin-aerolink' for simple and easy
to understand text.

In 'draft-templin-aerolink', there are only five acronyms:

  MTU - the MTU of the tunnel as seen by the IP layer
  MFU - the maximum fragment size within the tunnel
  MRU - the maximum reassembly size supported by the egress
  ENCAPS - the size of the encapsulation headers inserted by the ingress
  PTB - an IP Packet Too Big message

Everything else is just simple English language, and does not need any
other acronyms.

Thanks - Fred
[email protected]

> Joe
> 


_______________________________________________
Int-area mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/int-area

Reply via email to