Hi, On Fri, Mar 09, 2018 at 06:48:16PM -0500, Andrew Sullivan wrote:
brought up already more than once, and I see at least one recommendation that is so breathtakingly ignorant of the architecutre of the Internet as to make me roll my eyes.
I've been asked by a couple of people (slightly pointedly in one case) what inspired that remark. It was recommendation 6. It talks about "transborder dataflow". As I said in my email, I'd basically just skimmed the paper, because I doubted there was anything new in it. Worrying about transborder dataflow was the only thing that struck me as at all novel. But it turns out, now that I have actually read the explanatory text, that this recommendation isn't about "flows" at all. It's about data that is transferred from one place to another, and therefore about trans-border _shipment_ of data and where that data comes to rest. This isn't a new issue, so I again wonder what was supposed to be so new in this paper that the WG ought to add it to the mountain of material we already have. The difference between flow and shipment is actually an important distinction, however, because the architecture of Internet routing means that the actual data _flow_ is quite unpredictable. For instance, depending on whom on the Internet I am talking to while in Toronto, my data might go from my house to 151 Front Street to the other person's house, or might go from my house to Chicago or Newark and back to Toronto. I am aware of a pair of businesses in Singapore that were one floor apart in the same building, whose network path met in California (we fixed that one). What made me roll my eyes was the suggestion that ICANN ought to have some sort of policy about what routes data took on the Internet. Since that is not actually what was being suggestion, I have stopped rolling my eyes (but I still don't think there is anything new in the report). Best regards, A -- Andrew Sullivan ajs@anvilwalrusden.com