Re: [tcpm] On allocating reserved bits in the TCP header

["Scharf, Michael" <[e-mail]>]

> As author of draft-grimes-tcpm-tcpsce, which is also requesting use of a TCP bit
> for experimental purpose, ask what the next steps are?

The usual dos and don'ts apply: Socialize the idea, get feedback, consider the feedback, update the document, convince the working group, etc.

We have 4 reserved bits left in the TCP header. There has been a past and completed experiments with bit 7. The past use could be considered a precedent when publishing a standards track document that allows new experiments with reserved bits. The precedent only exist for bit 7.

I am aware of at least 4 different ideas that have been posted recently to this list regarding use of a reserved bit. If the TCPM working group (well, actually the IETF) reached consensus that more than one experiment was useful, we may have to think about a way of negotiating what experiment is in use on a connection. Or find another solution.

As an individual contributor to TCPM, my current thinking is that one bit (i.e., bit 7) would be enough for experimentation - for the time being.

This is my personal view. The last change to the main TCP header was done in 2001. It is hard to predict what the IETF consensus would be these days.

Michael