At 10:59 AM 11/30/01, Brent Wrisley wrote: >Although intf's won't combine packets, I usually set higher MTU's so that >larger packets with the DF flag on will be allowed through.
Although applications may set the DF bit (usually as part of the MTU discovery process but for other reasons too), they still don't tend to send very large packets. Even once you add some slop for extra encapsulations, tunneling, VPNs, VLANs, etc. you don't need the MTU to be much larger than the Ethernet MTU because end-user applications and configurations don't tend to send packets that are larger than the Ethernet MTU. This may change with time (and I'm sure there are some applications that do send larger frames), but typical ones don't. In fact, it's a good thing that they don't, as realtime (such as voice) packets get delayed when a (slow) interface takes a long time to serially output bits in a very large frame. I don't want to beat a dead horse, but it seemed like maybe you missed my point.... ;-} Priscilla >Brent > >On 29/11/01 21:16 -0500, Priscilla Oppenheimer wrote: > >: >[...] > A lot of novices think that having a large interface >:MTU is going to make a big difference, but I'm not convinced. The interface >:isn't going to combine packets it receives into larger packets just because >:of the larger MTU. Packets can't grow!? >[...] >: >:I'll have to look into this. Thoughts? Comments? >: >:Priscilla >: >:At 05:26 PM 11/29/01, Howard C. Berkowitz wrote: >:> >Found this in RFC 1661 which documents PPP: >:> > >:> >The maximum length for the Information field, including Padding, but not >:> >including the Protocol field, is termed the Maximum Receive Unit (MRU), >:> >which defaults to 1500 octets. By negotiation, consenting PPP >:> >implementations may use other values for the MRU. >:> > >:> >P. >:> >:>Hmmm...I definitely am aware of providers using 4470 on POS links, >:>and a general trend in the gigabit-plus world to use larger MTUs. Is >:>this simply industry practice, I wonder, or are there some overriding >:>IEEE or IETF documents? Perhaps in the sub-IP area, such as IP over >:>Optical? >:> >:> > >:> >At 03:05 PM 11/29/01, Howard C. Berkowitz wrote: >:> >> >Howard, >:> >> > >:> >> >Correct me if am wrong but, the HDLC advantage over PPP is the MTU >:size. >:> >> >PPP supports 1500 while HDLC 4xxx (can't remember the exact number), >:> >> >this might be helpful in situations where DF bit is set. >:> >> > >:> >> >Nabil >:> >> >:> >>I'd have to research this -- I don't offhand remember PPP (as the >:> >>protocol) having a MTU limit that small. It would surprise me, given >:> >>the interest in POS. >:> >________________________ >:> > >:> >Priscilla Oppenheimer >:> >http://www.priscilla.com >:________________________ >: >:Priscilla Oppenheimer >:http://www.priscilla.com >: >: >: >: >Brent Wrisley >-------------------- >2FB6 85AD 7084 80A0 8381 C116 CDE5 78B5 E959 C536 >PGP Key ID: 0xE959C536 (us.pgp.net) ________________________ Priscilla Oppenheimer http://www.priscilla.com Message Posted at: http://www.groupstudy.com/form/read.php?f=7&i=27863&t=27637 -------------------------------------------------- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
