On Thu, 18 Sep 2003, Tim Soderstrom wrote: > Just a couple of concerns: > > So this has been bugging me since I first read technical reports on the > whole thing quite a few years ago: Isn't 128-bits kinda, well, a lot? I > mean that's 4 times bigger than 32-bits, so doesn't that mean it will > incur 4x more overhead?
There is no 4x overhead: The length of the header of a simple IPv4 packet is usually 24 byte while a simplest IPv6 header is usually 40 byte since lots of unnecessary fields become optional. > > For example, I have a 160bps upstream with my DSL provider...right now > is just barely enough to stream mp3's on IPv4. On IPv6, however, I worry > that a much bigger chunk of my bits will be used simply for addressing. > How does IPv6 answer this (as it is really the only thing holding me > back). > > Also, does IPv6 or could IPv6, 7, or 8 :) Employ a type of 'smart > addressing' feature? For example, if all I need to do is communicate > amongst my subnet or my local network, it seems wasteful to send > 128-bits for that. So why not simply send the number of bits that is in > the subnet mask or assume a right justification of the bits recieved so > that the computer can and it to a mask and know where it came from or > something similar? There is a possibility to compress the header see: RFC 2507 Regards, Janos Mohacsi --------------------------------------------------------------------- The IPv6 Users Mailing List Unsubscribe by sending "unsubscribe users" to [EMAIL PROTECTED]