Re: [tcpm] On the implementation of TCP urgent data (IETF Internet Draft)

[Andrew Yourtchenko <ayourtch at cisco.com>]

On Thu, 12 Mar 2009, Jerry Leichter wrote:

> In summary:  This is not, and has never been, a TCP problem.  The issues are 
> at the boundary between the TCP stack and the socket API, affecting code on 
> both sides of that boundary.

That's somewhat the problem is that it's "noone's problem". In my opinion, 
the best thing is a strongest possible message to the application 
developers "if you were planning to use this, don't". Whether the 
definition of is one byte left or right does not matter - the ambiguity is 
*already* there. Granted, it can be mitigated within the specific 
administrative domain, but if I talk to you across the internet, there's 
absolutely no guarantee whatsoever (try to interrupt an FTP transfer from 
a randomly selected server and see how many actually do not hang - 
amongst those that keep the tcp/21 connection of course)

> Finally, on the question of whether the socket API should be changed to 
> eliminate OOB data:  That's perhaps an arguable position, exactly because the 
> implementations are so bad; but the implicit message in the way you state it 
> - that TCP is what matters, and the socket API is just some minor thing - is 
> exactly backwards.  There are at most, what, a couple of hundred engineers in 
> the entire world, working on at most a few tens of distinct code bases, who 
> actually deal with *TCP*.  There are hundreds of thousands of engineers 
> dealing with millions of applications that depend on the socket API.

The initial discussion we had with Fernando that gave a rise to this draft 
was of the flavour "TCP Urgent considered harmful for application 
developers".

> Changing TCP is a big job, but one that would be doable if there were 
> sufficient need:  Look at IPv6.  Changing the socket API is impossible - if 
> you think IPv6 is taking a long time to roll out, that's nothing compared to 
> how long it would take a new, incompatible socket API replacement.  Chances 
> are it would never happen.

All depends on the business needs, but the TCP Urgent is indeed not 
among the most widely used ones. A clear and well-structured semantics of Out-Of-Band 
Notifications might be of some use though, IMHO - but then, one could as 
well use the separate UDP channel along - though it is not as "convenient".

thanks,
andrew