Re: [tcpm] On TCP option codepoints

[Wesley Eddy <[e-mail]>]

On 10/8/2013 8:49 AM, Scharf, Michael (Michael) wrote:
> Hi all,
> 
> I've performed today a small and very nonscientific experiment:
> 
> For all TCP option codepoints N marked as "Reserved" on
> http://www.iana.org/assignments/tcp-parameters/tcp-parameters.xhtml,
> I entered the term "tcp option N" in a major Internet search engine
> and applied a special, human ranking by post-processing the first 5
> results for the exact match, if available. To ensure that only the
> data analysis tool is proprietary and biased, but not in the data
> source, I also verified the key result in a second major search
> engine, and I used hex values for N as well ;)


Nice idea; thanks for doing this!


> - For N= 38 (0x26), there are several links to official product
> documentation
> 


In at least this case, where there is evidence that deployed products
have stolen the codepoint, there should be an indication on the IANA
webpage, similar to the existing ones.  I think it's best if the
responsible AD for TCPM handles this (sorry Martin!) though.

In my personal opinion, the number of codepoint thefts is significant
enough (and showing no signs of stopping) that instead of the generic
statement about unauthorized use on all of them, we should specifically
mention the vendors and products known to be abusing them.  That way,
if at some point those products go away and become obsolete, then the
codepoints can be reaped back for legitimate use.


-- 
Wes Eddy
MTI Systems