[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

length of shim6 control mesages



Iljitsch in his review has raised the following issue:

my comments are inline

I see no discussion of size issues. A single option can be made large enough to push a packet beyond 1280 bytes. More realistically, this will happen when multiple options are present. What happens in this case? What is the largest option size and the largest shim packet size implementations must be prepared to handle?


so far we haven't considered the case where the options wouldn't fit in one packet... 1280 bytes seems big enough.... otoh, CGA parameter data strutucture may be quite big, but not much more than 3000 bits, and i guess that in general there won't be much more than 10 addresses in a locator set, which is 1280 bits more.... so we are still with quite a few extra bits... In addition, it is always possible to do option engineering, meaning that in the case that you need to send many options, you can send them in different messages. In particular, you need to send the CGA paramter data strucutre only once, and then you can send the other options in UPDATe messages...

the other option would be to define a mechanism to fragment shim6 control messages, but i think this i too much complexity for this, that seems to be a corner case...

what do you suggest?