On May 2, 2018, at 6:14 AM, Dave O'Reilly <[email protected]> wrote: > No, what you said was that “it’s not clear that the work is in scope of the > working group anyway”. And I was asking you, if it isn’t, where would be a > more appropriate place to have the discussion?
It's possible that the IETF would be the best place to do this work if we had the energy to do it. But I don't think we do. Stephen actually said he'd be willing to work on something, but I don't think you took that as a genuine offer, or else you concluded that what he'd come up with wouldn't satisfy your needs. There isn't some other working group in the IETF of which I'm aware that would be a better place to work on this. That's why I've mentioned doing it in government standards bodies. E.g., since I think you're in Ireland, maybe ETSI. > Well what it seems to me that you’re saying is that you don’t want to work on > the draft, and that’s entirely up to you. So we have you on one side of the > opinion and me on the other. I can already see which way (at least the vocal > part of the) consensus is - you’d need to be intellectually blind not to - > but for me this conversation is as important as the eventual outcome of the > document per se. What I was exactly saying is that if you want people here to work on it, you need to convince us that it's worth working on, and you haven't done that. > I could go to law enforcement agencies, regulators etc. and talk about this > problem - and I have - but they already know it’s a problem and recognise > that something needs to be done. It’s only when you present an opinion to a > group who have a totally different perspective that you get new arguments > presented and get make any sort of intellectual progress. In that regard, > none of this discussion has made me less convinced that source port logging > is the best solution to crime attribution problems arising from CGN (IPv6 > migration comes with its own can of worms that I discuss elsewhere). In fact > I think I recall that even you agreed that if IP address is being logged, > source port should be logged as well. Yes. I don't think anybody here is disagreeing with you on a technical level—that's why that's not what we need to talk about. > The only alternatives that I can see for closing the crime attribution > information gap that has been opened by CGN are (a) severely limit the number > of endpoints behind a given CGN IP address (a la what the Belgian ISPs are > doing) or (b) require ISPs to perform connection logging (which is a privacy > disaster) or (c) force migration to IPv6 (which only moves the problem around > as per my other document). If there are any other options, I’d love to hear > them. I can tell you all out there who work in ISPs that if/when regulators > act on this, it’s going to be one of those options for you. Regulators have > only power to regulate ISPs so they will find ISP-centric solutions, which > are the list of options above - none of which I consider to be very > attractive (apart from migration to IPv6). What I was proposing was an > “Internet” solution of a sort, where if source port logging is done routinely > then there is no need for regulatory intervention to close the information > gap - I would like to think that what I was trying to propose was the “least > worst” option for dealing with the problem. That's a much better approach to take if you want to convince the working group that this is worth doing.
_______________________________________________ Int-area mailing list [email protected] https://www.ietf.org/mailman/listinfo/int-area
