Re: Is Fragmentation at IP layer even needed ?

[joel jaeggli <[e-mail]>]

On 2/8/16 10:56 AM, David Borman wrote:
> 
>> On Feb 8, 2016, at 12:23 PM, Warren Kumari <[e-mail]>
>> wrote:
> ...
>> On Mon, Feb 8, 2016 at 9:05 AM David Borman <[e-mail]>
>> wrote:
> ...
>> So if you are writing an application that needs >1500 octets, use
>> an IPv6 implementation that supports >1500 octet fragmentation and
>> reassembly.
>> 
>> ... but as an application writer (or, basically anyone else), I
>> have no control over the "IPv6 implementation". Even if I'm in an
>> environment where I do control the OS / model of all devices, and I
>> know they support >1500 octet, it seems like a bad idea to *rely*
>> on that. Sometime I'm going to want to change OS / add some other
>> device, be able to interact with some other system. This sounds
>> like vendor lock at its worst…
> 
> If you wind up in a scenario where you get locked to a particular OS
> vendor because it’s the only one that supports IPv6 fragmentation
> >1500 octets, then that is probably the least of your worries.  I’d
> be much more worried about IPv6 fragmentation in light of Ron
> Bonica’s comment that intermediary nodes drop packets with extension
> headers, which is bad news even for fragmented packets in the
> 1280-1500 range.

For those of us with ecmp load balancing the challenge of associating a
fragement with the rest of the flow are also a problem. In my own case I
can engineer circumstances where I should never receive such a fragment,
so I can safely drop them anyway but I doubt everyone has that luxury.

> -David
> 
> 
>