Re: [Int-area] Discussion about Section 6.1 in draft-ietf-intarea-frag-fragile

["Templin (US), Fred L" <[e-mail]>]

Brian,

> -----Original Message-----
> From: Brian E Carpenter [mailto:[e-mail]]
> Sent: Wednesday, September 11, 2019 4:14 PM
> To: Bob Hinden <[e-mail]>; Templin (US), Fred L <[e-mail]>
> Cc: [e-mail]; Suresh Krishnan <[e-mail]>
> Subject: Re: [Int-area] Discussion about Section 6.1 in draft-ietf-intarea-frag-fragile
> 
> On 12-Sep-19 10:59, Bob Hinden wrote:
> > Fred,
> >
> >> On Sep 11, 2019, at 7:48 AM, Templin (US), Fred L <[e-mail]> wrote:
> >>
> >> Geoff, the 1280 MTU came from Steve Deering's November 13, 1997 proposal to
> >> the ipngwg. The exact message from the ipng archives is reproduced below.
> >>
> >> 1280 isn't just a recommendation - it's *the law*. Any link that cannot do 1280
> >> (tunnels included) is not an IPv6 link.
> >
> > Yes from IPv6’s view, but you can make a link that can’t do 1280 work if it has its own local L2 fragmentation / reassembly as noted in
> Steve’s email.  ATM with is 53 byte cells comes to mind.
> 
> IPv4 with a small PMTU also comes to mind, as discussed in Section 3.2.2 of RFC 4213:
> 
>    In this case, the IPv6 layer has to "see" a link
>    layer with an MTU of 1280 bytes and the encapsulator has to use IPv4
>    fragmentation in order to forward the 1280 byte IPv6 packets.

Yes, IP fragmentation - exactly.

Fred

>       Brian
> 
> >
> > Bob
> >
> >
> >>
> >> Fred
> >>
> >> ---
> >> From [e-mail]  Thu Nov 13 16:41:01 1997
> >> Received: (from majordomo@localhost)
> >> 	by sunroof.eng.sun.com (8.8.8+Sun.Beta.0/8.8.8) id QAA14339
> >> 	for ipng-dist; Thu, 13 Nov 1997 16:38:00 -0800 (PST)
> >> Received: from Eng.Sun.COM (engmail1 [129.146.1.13])
> >> 	by sunroof.eng.sun.com (8.8.8+Sun.Beta.0/8.8.8) with SMTP id QAA14332
> >> 	for <ipng@sunroof>; Thu, 13 Nov 1997 16:37:51 -0800 (PST)
> >> Received: from saturn.sun.com (saturn.EBay.Sun.COM [129.150.69.2])
> >> 	by Eng.Sun.COM (SMI-8.6/SMI-5.3) with SMTP id QAA28654
> >> 	for <[e-mail]>; Thu, 13 Nov 1997 16:37:48 -0800
> >> Received: from postoffice.cisco.com (postoffice.cisco.com [171.69.200.88])
> >> 	by saturn.sun.com (8.8.8/8.8.8) with ESMTP id QAA28706
> >> 	for <[e-mail]>; Thu, 13 Nov 1997 16:37:49 -0800 (PST)
> >> Received: from [171.69.199.124] (deering-mac.cisco.com [171.69.199.124]) by postoffice.cisco.com (8.8.5-Cisco.1/8.6.5) with ESMTP
> id QAA20862; Thu, 13 Nov 1997 16:37:48 -0800 (PST)
> >> X-Sender: [e-mail]
> >> Message-Id: <v03110702b0598e80008d@[171.69.199.124]>
> >> Mime-Version: 1.0
> >> Content-Type: text/plain; charset="us-ascii"
> >> Date: Thu, 13 Nov 1997 16:37:00 -0800
> >> To: IPng Working Group <[e-mail]>
> >> From: Steve Deering <[e-mail]>
> >> Subject: (IPng 4802) increasing the IPv6 minimum MTU
> >> Cc: [e-mail]
> >> Sender: [e-mail]
> >> Precedence: bulk
> >>
> >> In the ipngwg meeting in Munich, I proposed increasing the IPv6 minimum MTU
> >> from 576 bytes to something closer to the Ethernet MTU of 1500 bytes, (i.e.,
> >> 1500 minus room for a couple layers of encapsulating headers, so that min-
> >> MTU-size packets that are tunneled across 1500-byte-MTU paths won't be
> >> subject to fragmentation/reassembly on ingress/egress from the tunnels,
> >> in most cases).
> >>
> >> After the short discussion in the Munich meeting, I called for a show of
> >> hands, and of those who raised their hands (about half the attendees, if
> >> I recall correctly), the vast majority were in favor of this change --
> >> there were only two or three people opposed.  However, we recognized that
> >> a fundamental change of this nature requires thoughtful discussion and
> >> analysis on the mailing list, to allow those who were not at the meeting
> >> and those who were there but who have since had second thoughts, to express
> >> their opinions.  A couple of people have already, in private conversation,
> >> raised some concerns that were not identified in the discussion at the
> >> meeting, which I report below.  We would like to get this issue settled as
> >> soon as possible, since this is the only thing holding up the publication
> >> of the updated Proposed Standard IPv6 spec (the version we expect to advance
> >> to Draft Standard), so let's see if we can come to a decision before the ID
> >> deadline at the end of next week (hoping there isn't any conflict between
> >> "thoughtful analysis" and "let's decide quickly" :-).
> >>
> >> The reason I would like to increase the minimum MTU is that there are some
> >> applications for which Path MTU Discovery just won't work very well, and
> >> which will therefore limit themselves to sending packets no larger than
> >> the minimum MTU.  Increasing the minimum MTU would improve the bandwidth
> >> efficiency, i.e., reduce the header overhead (ratio of header bytes to
> >> payload bytes), for those applications.  Some examples of such applications
> >> are:
> >>
> >>    (1) Large-fanout, high-volume multicast apps, such as multicast video
> >> 	("Internet TV"), multicast netnews, and multicast software
> >> 	distribution.  I believe these applications will end up limiting
> >> 	themselves to packets no large than the min MTU in order to avoid
> >> 	the danger of incurring  an "implosion" of ICMP Packet-Too-Big
> >> 	messages in response.  Even though we have specified that router
> >> 	implementations must carefully rate-limit the emission of ICMP
> >> 	error messages, I am nervous about how well this will work in
> >> 	practice, especially once there is a lot of high-speed, bulk
> >> 	multicasting happening.  An appropriate choice of rate or
> >> 	probability of emission of Packet-Too-Big responses to multicasts
> >> 	really depends on the fan-out of the multicast trees and the MTUs of
> >> 	all the branches in that tree, which is unknown and unknowable to
> >> 	the routers.  Being sensibly conservative by choosing a very low
> >> 	rate could, in many cases, significantly increase the delay before
> >> 	the multicast source learns the right MTU for the tree and, hence,
> >> 	before receivers on smaller-MTU branches can start receiving the
> >> 	data.
> >>
> >>    (2) DNS servers, or other similar apps that have the requirement of
> >> 	sending a small amount of data (a few packets at most) to a very
> >> 	large and transient set of clients.  Such servers often reside on
> >> 	links, such as Ethernet, that have an MTU bigger than the links on
> >> 	which many of their clients may reside, such as dial-up links.  If
> >> 	those servers were to send many reply messages of the size of their
> >> 	own links (as required by PMTU Discovery), they could incur very
> >> 	many ICMP packet-too-big messages and consequent retransmissions of
> >> 	the replies -- in the worse case, multiplying the total bandwidth
> >> 	consumption (and delivery delay) by 2 or 3 times that of the
> >> 	alternative approach of just using the min MTU always.  Furthermore,
> >> 	the use of PMTU Discovery could result in such servers filling up
> >> 	lots of memory withed cached PMTU information that will never be
> >> 	used again (at least, not before it gets garbage-collected).
> >>
> >> The number I propose for the new minimum MTU is 1280 bytes (1024 + 256,
> >> as compared to the classic 576 value which is 512 + 64).  That would
> >> leave generous room for encapsulating/tunnel headers within the Ethernet
> >> MTU of 1500, e.g., enough for two layers of secure tunneling including
> >> both ESP and AUTH headers.
> >>
> >> For medium-to-high speed links, this change would reduce the IPv6 header
> >> overhead for min MTU packets from 7% to 3% (a little less than the IPv4
> >> header overhead for 576-byte IPv4 packets).  For low-speed links such as
> >> analog dial-up or low-speed wireless, I assume that header compression will
> >> be employed, which compresses out the IPv6 header completely, so the IPv6
> >> header overhead on such links is effectively zero in any case.
> >>
> >> Here is a list of *disadvantages* to increasing the IPv6 minimum MTU that
> >> have been raised, either publically or privately:
> >>
> >>    (1) This change would require the specification of link-specific
> >> 	fragmentation and reassembly protocols for those link-layers
> >> 	that can support 576-byte packets but not 1280-byte packets,
> >> 	e.g., AppleTalk.  I think such a protocol could be very simple,
> >> 	and I briefly sketch such a protocol in Appendix I of this
> >> 	message, as an example.
> >>
> >> 	Often, those links that have a small native MTU are also the ones
> >> 	that have low bandwidth.  On low-bandwidth links, it is often
> >> 	desirable to locally fragment and reassemble IPv6 packets anyway
> >> 	(even 576-byte ones) in order to avoid having small, interactive
> >> 	packets (e.g., keystrokes, character echoes, or voice samples)
> >> 	be delayed excessively behind bigger packets (e.g., file transfers);
> >> 	the small packets can be interleaved with the fragments of the
> >> 	big packets.  Someone mentioned in the meeting in Munich that the
> >> 	ISSLL WG was working on a PPP-specific fragmentation and
> >> 	reassembly protocol for precisely this reason, so maybe the job
> >> 	of specifying such a protocol is already being taken care of.
> >>
> >>    (2) Someone raised the concern that, if we make the minimum MTU close
> >> 	to Ethernet size, implementors might never bother to implement PMTU
> >> 	Discovery.  That would be regrettable, especially if the Internet
> >> 	evolves to much more widespread use of links with MTUs bigger
> >> 	than Ethernet's, since IPv6 would then fail to take advantage of
> >> 	the bandwidth efficiencies possible on larger MTU paths.
> >>
> >>    (3) Peter Curran pointed out to me that using a larger minimum MTU for
> >> 	IPv6 may result in much greater reliance on *IPv4* fragmentation and
> >> 	reassembly during the transition phase while much of the IPv6
> >> 	traffic is being tunneled over IPv4.  This could incur unfortunate
> >> 	performance penalties for tunneled IPv6 traffic (disasterous
> >> 	penalties if there is non-negligible loss of IPv4 fragments).
> >> 	I have included Peter's message, describing his concern in more
> >> 	detail, in Appendix II of this message.
> >>
> >>    (4) Someone expressed the opinion that the requirement for link-layer
> >> 	fragmentation and reassembly of IPv6 over low-cost, low-MTU links
> >> 	like Firewire, would doom the potential use of IPv6 in cheap
> >> 	consumer devices in which minimizing code size is important --
> >> 	implementors of cheap Firewire devices would choose IPv4 instead,
> >> 	since it would not need a fragmenting "shim" layer.  This may well
> >> 	be true, though I suspect the code required for local frag/reasm
> >> 	would be negligible compared to the code required for Neighbor
> >> 	Discovery.
> >>
> >> Personally, I am not convinced by the above concerns that increasing the
> >> minimum MTU would be a mistake, but I'd like to hear what the rest of the
> >> WG thinks.  Are there other problems that anyone can think of?  As I
> >> mentioned earlier, the clear consensus of the Munich attendees was to
> >> increase the minimum MTU, so we need to find out if these newly-identified
> >> problems are enough to swing the consensus in the other direction.  Your
> >> feedback is heartily requested.
> >>
> >> Steve
> >>
> >> ----------
> >>
> >> Appendix I
> >>
> >> Here is a sketch of a fragmentation and reassembly protocol (call it FRP)
> >> to be employed between the IP layer and the link layer of a link with native
> >> (or configured) MTU less than 1280 bytes.
> >>
> >> Identify a Block Size, B, which is the lesser of (a) the native MTU of the
> >> link or (b) a value related to the bandwidth of the link, chosen to bound
> >> the latency that one block can impose on a subsequent block.  For example,
> >> to stay within a latency of 200 ms on a 9600 bps link, choose a block size
> >> of .2 * 9600 = 2400 bits = 240 bytes.
> >>
> >> IPv6 packets of length <= B are transmitted directly on the link.
> >> IPv6 packets of length > B are fragmented into blocks of size B
> >> (the last block possibly being shorter than B), and those fragments
> >> are transmitted on the link with an FRP header containing the following
> >> fields:
> >>
> >> 	[packet ID, block number, end flag]
> >>
> >> where:
> >>
> >> 	packet ID is the same for all fragments of the same packet,
> >> 	and is incremented for each new fragmented packet.  The size of
> >> 	the packet ID field limits how many packets can be in flight or
> >> 	interleaved on the link at any one time.
> >>
> >> 	block number identifies the blocks within a packet, starting at
> >> 	block zero.  The block number field must be large enough to
> >> 	identify 1280/B blocks.
> >>
> >> 	end flag is a one-bit flag which is used to mark the last block
> >> 	of a packet.
> >>
> >> For example, on a 9600 bps serial link, one might use a block size of
> >> 240 bytes and an 8-bit FRP header of the following format:
> >>
> >> 	4-bit packet ID, which allows interleaving of up to 16 packets.
> >> 	3-bit block number, to identify blocks numbered 0 through 5.
> >> 	1-bit end flag.
> >>
> >> On a 256 kpbs AppleTalk link, one might use the AppleTalk-imposed block
> >> size of ~580 bytes and an 8-bit FRP header of the following format:
> >>
> >> 	5-bit packet ID, which allows for up to 32 fragmented packets in
> >> 		   flight from each source across the AppleTalk internet.
> >> 	2-bit block number, to identify blocks numbered 0 through 2.
> >> 	1-bit end flag.
> >>
> >> On a multi-access link, like AppleTalk, the receiver uses the link-level
> >> source address as well as the packet ID to identify blocks belonging to
> >> the same packet.
> >>
> >> If a receiver fails to receive all of the blocks of a packet by the time
> >> the packet number wraps around, it discards the incompletely-reassembled
> >> packet.  Taking this approach, no timers should be needed at the receiver
> >> to detect fragment loss.  We expect the transport layer (e.g., TCP) checksum
> >> at the final IPv6 destination to detect mis-assembly that might be caused by
> >> extreme misordering/delay during transit across the link.
> >>
> >> On links on which IPv6 header compression is being used, compression is
> >> performed before fragmentation, and reassembly is done before decompression.
> >>
> >> ----------
> >>
> >> Appendix II
> >>
> >> From: Peter Curran <[e-mail]>
> >> Subject: Re: IPv6 MTU issue
> >> To: [e-mail] (Steve Deering)
> >> Date: Mon, 22 Sep 1997 11:50:34 +0100 (BST)
> >>
> >> Steve
> >>
> >> My problem was that moving the MTU close to 1500 would have an adverse
> >> effect on the transition strategy.  The current strategy assumes that the
> >> typical Internet MTU is >576, and that sending an IPv6 packet close to the
> >> minimum MTU will not require any IPv4 fragmentation to support the tunnel
> >> transparently.  The PMTU discovery mechanism will 'tune' IPv6 to use a
> >> suitable MTU.
> >>
> >> If the IPv4 MTU is <= 576 then IPv4 fragmentation will be required to
> >> provide a tunnel with a minimum MTU of 576 for IPv6.  This clearly places
> >> a significant strain on the tunnelling nodes - as these will normally be
> >> routers then there will be a demand for memory (for reassembly buffers)
> >> as well as CPU (for the frag/reassembly process) that will have an overall
> >> impact on performance.
> >>
> >> This is an acceptable risk, as Internet MTU's of <= 576 are not too common.
> >>
> >> However, if the minimum MTU of IPv6 is increased to something of the order
> >> of 1200-1500 octets then the likelihood of finding an IPv4 path with an
> >> MTU lower than this value increases (I think significantly) and this will
> >> have a performance impact on these devices.
> >>
> >> During the brief discussion of this matter in the IPNG session at Munich
> >> you stated that MTU's less than 1500 where rare.  I don't agree with this
> >> completely - it seems to be pretty common practise for smaller 2nd and 3rd
> >> tier ISP's in the UK to use an MTU of 576 for connection to their transit
> >> provider.  Their objective, I believe, is to 'normalize' the packet sizes
> >> on relatively low bandwidth circuits (typically <1Mbps) to provide better
> >> performance for interactive sessions compared to bulk-file transfer users.
> >>
> >> I think that before we go ahead and make a decision on an increased minimum
> >> MTU for IPv6 then we should discuss the issues a little more.
> >>
> >> Incidentally, I am not convinced of the benefits of doing this anyway
> >> (ignoring the issue raised above).  With a properly setup stack the PMTU
> >> discovery mechanism seems to be able to select a good MTU for use on the
> >> path - at least that is my experience on our test network and the 6Bone.
> >>
> >> I appreciate that you are trying to address the issues of PMTU for multi-
> >> casting but I don't see how raising the minumum MTU is going to help much.
> >> PMTU discovery will still be required irrespective of the minimum MTU
> >> adopted, unless we adopt a value that can be used on all link-layer technolo-
> >> gies.
> >>
> >> I would welcome wider discussion of these issues before pressing ahead
> >> with a change.
> >>
> >> Best regards
> >>
> >> Peter Curran
> >> TICL
> >>
> >>
> >> --------------------------------------------------------------------
> >> IETF IPng Working Group Mailing List
> >> IPng Home Page:                      http://playground.sun.com/ipng
> >> FTP archive:                      ftp://playground.sun.com/pub/ipng
> >> Direct all administrative requests to [e-mail]
> >> --------------------------------------------------------------------
> >
> >
> > _______________________________________________
> > Int-area mailing list
> > [e-mail]
> > https://www.ietf.org/mailman/listinfo/int-area
> >