Re: Alvaro Retana's Discuss on draft-ietf-6man-rfc2460bis-09: (with DISCUSS and COMMENT)

["Alvaro Retana (aretana)" <[e-mail]>]

Ole:

Hi!

Thanks for the clarification!  It would be nice if the document also said this – from what I could read, the text doesn’t match what you said – or did I miss it?

Alvaro.

On 4/11/17, 5:17 PM, "[e-mail]" <[e-mail]> wrote:

>
>The initial sentence mentions 4 actions: examine, process, insert, or delete.
>The text already says that any node can examine the headers “for any reason”, according to [rfc7045].  So maybe take that out as part of the original 4.
>I do have an additional clarity question.  What does “process” mean?  Does it include a “change en-route”?    I’m asking because Section 4.2. (Options) says that “the Hop-by-Hop Options header and the Destination Options header” has an option bit that indicates “whether or not the Option Data of that option can change en-route to the packet's final destination”.  I’m assuming that to change the data the transit node has to not just “examine” (which I think means “look at”), but also (at least) “process” the EH as well.   If to “process” includes the ability to “change en-route”, then it looks like the Destination Option may also be an exception…

Both options inside of an HBH option header and a Destination Options header can have the change en-route flag set.
The Destinations Options header are only processed at the destination host, so an option can then only change en-route when it is part of a source route path (and the Destination options header is before the routing header).

The "feature" here is that one can specify options that can be processed only by the routers listed in a routing header or one can specify options that can change en-route for every hop along the path.