[Int-area] Multi-MTU subnet draft

[Iljitsch van Beijnum <iljitsch at muada.com>]

Hi,

Here is an update of my multi-MTU subnet draft. Based on the  
discussion in the summer I've pruned the CRC text (although most of  
the points are still there) and I've spent a lot of time coming up  
with a way to make all of this simpler. I think I found a good tradeoff.

The basic idea is that routers advertise three parameters in IPv6  
router advertisements. (These parameters apply equally to IPv4 and  
IPv6 operation, though.) They are:

- An absolute maximum packet size that is allowed on a link. This can be
   used make sure larger packets aren't used if they're known to have  
bad
   effects (MAXMTU)

- A maximum packet size for nodes operating at 10 or 100 Mbps (the  
cutoff
   is 600 Mbps) in case these are connected to switches that don't  
support
   jumboframes or it's undesirable that slow nodes send large packets  
for
   QoS reasons (SLOWMTU)

- A packet size cutoff for not probing / probing for a jumbo-clean path
   (SAFEMTU)

The neighbor discovery options and jumbo ARP for communicating per- 
neighbor MTUs and probing whether large packets work are still there,  
but they are optional. Hosts may use RFC 4821 as a probing mechanism  
instead. (Although this means that some transport protocols can use  
larger packets and others can't and routers can't use RFC 4821.)

Alternatively, nodes may forego probing altogether if they stick to a  
maximum packet size of SAFEMTU. The disadvantage of this is that  
SAFEMTU must be set administratively because only the administrator  
can know what a safe MTU is in the absense of probing. But the good  
part is that it is compatible with current manually configured jumbo  
frame use.

There needs to be a bit more text about operational issues (or maybe  
those should go in an applicability document), and I'm not sure yet if  
I want to keep the CRC discussion even though it's a lot shorter now.  
Depending on the discussion on the list and in Philadelphia I want to  
see if this can be published as experimental.

Begin forwarded message:

> A New Internet-Draft is available from the on-line Internet-Drafts  
> directories.

> 	Title           : Extensions for Multi-MTU Subnets
> 	Author(s)       : I. van Beijnum
> 	Filename        : draft-van-beijnum-multi-mtu-02.txt
> 	Pages           : 17
> 	Date            : 2008-02-24

> In the early days of the internet, many different link types with
> many different maximum packet sizes were in use.  For point-to-point
> or point-to-multipoint links, there are still some other link types
> (PPP, ATM, Packet over SONET), but shared subnets are now almost
> exclusively implemented as ethernets.  Even though the relevant
> standards mandate a 1500 byte maximum packet size for ethernet, more
> and more ethernet equipment is capable of handling packets bigger
> than 1500 bytes.  However, since this capability isn't standardized,
> it's seldom used today, despite the potential performance benefits of
> using larger packets.  This document specifies a mechanism for
> advertising a non-standard maximum packet size on a subnet.  It also
> specifies optional mechanisms to negotiate per-neighbor maximum
> packet sizes so that nodes on a shared subnet may use the maximum
> mutually supported packet size between them without being limited by
> nodes with smaller maximum sizes on the same subnet.

> A URL for this Internet-Draft is:
> http://www.ietf.org/internet-drafts/draft-van-beijnum-multi-mtu-02.txt

_______________________________________________
Int-area mailing list
Int-area at ietf.org
http://www.ietf.org/mailman/listinfo/int-area