Marshall Eubanks wrote: > Note, too, that many simple web pages have in them pieces from > other servers and other IP addresses (banner ads are frequently > done this way) and so this could put a real performance hit on > web page load times in the real world.
I have been thinking of this, but this is the first mention of it on the list. This is a typical Internet usage where people have to wait as it is, and would be waiting longer if any of the web servers relied on LISP-ALT ITRs which did not yet have the mapping for the particular EID in which the client is located. I was also thinking about BitTorrent, but that is not much of a problem, since it is a non-real-time operation and takes a while anyway. > If all of this is coming from one server farm, all is well and > good. But, what if (like Joost) these streams are being provided > by P2P ? I hadn't heard of real-time video delivery, but it doesn't surprise me that it exists, and I hadn't assumed it does not or would not exist: http://en.wikipedia.org/wiki/Joost http://en.wikipedia.org/wiki/P2PTV Minutes later I am running TVUPlayer and Joost. Neither has an obvious stop or pause button, so many users will leave the thing running when they aren't actively watching - just like real TV. I guess these real-time P2P applications are going to be very widely used, sucking data across the world even when no-one is watching - burdening ITRs all around the world and therefore any pull-based mapping database system. Perhaps when we think of adding delays to the start of Internet communications, we think along well known lines: "TCP means two mapping lookups, plus perhaps one or more for the DNS which preceded it." This is textbook stuff, but what is going on in the wild? Delays to a single TCP session, for instance for a website, is enough of a concern, but real-time P2P involves arbitrary numbers of streams of packets, to and and from hundreds of millions of hosts (later billions), without any DNS and with the expectation that the packet is, in general, delivered immediately. Adding any delays to initial packets will impact these systems in some way. In the one "session" the packets could be coming from any number of hosts all over the world, and that set of hosts could be changing all the the time. While UDP-based P2P networks don't assume reliable delivery, any delays in "initial" packets would affect packets in the middle of a session, and so degrade the service's quality, in addition to its start-up time. - Robin -- to unsubscribe send a message to [EMAIL PROTECTED] with the word 'unsubscribe' in a single line as the message text body. archive: <http://psg.com/lists/rrg/> & ftp://psg.com/pub/lists/rrg
