Re: [Int-area] WG Adoption Call: IP Fragmentation Considered Fragile

[Joe Touch <[e-mail]>]

> On Aug 26, 2018, at 2:31 PM, Toerless Eckert <[e-mail]> wrote:
> 
> On Sun, Aug 26, 2018 at 09:09:54AM -0700, Joe Touch wrote:
>> 
>> 
>>> On Aug 24, 2018, at 8:24 PM, Toerless Eckert <[e-mail]> wrote:
>>> 
>>> Of course. Will take a decade to get ubiquitously deployed, but
>>> neither IPv4 nor IPv6 will go away, only the problems with fragmentation
>>> will become worse and work if we do not have an exit strategy like this.
>>> 
>>> If we don't try an exit strategy like this, we will just get what
>>> Joe said, the complete segmentation of the Internet with more and
>>> more L4 or even higher layer proxies
>> 
>> FWIW, what I said was that *this exit strategy* would lead to the complete segmentation of the Internet and its consequences.
> 
> I can't rmember i saw a good explanation why on the thhread and neither in
> the draft - aka: Why we still need fragementation to keep the internet
> from falling apart". I like 4821 and think its the way to go also for other
> transports.


As I noted before in this thread and in other discussions on this topic, requiring support for fragmentation at another layer merely pushes all of what IP does (endpoint addressing, message multiplexing, etc.) to that other layer, at which point you’re back where you start - that layer then will end up needing to support fragmentation in places and ways you don’t want to support because of its cost.

For IPv4, the ONLY problem with fragmentation is the limited fragment ID space, and that can be mitigated as per RFC 6864.

Joe