On Fri, Aug 7, 2009 at 3:42 PM, Marshall Eubanks<t...@americafree.tv> wrote: > > On Aug 7, 2009, at 2:59 PM, Christopher Morrow wrote: > >> On Wed, Aug 5, 2009 at 5:07 PM, Margaret Wasserman<m...@sandstorm.net> >> wrote: >>> >>> On Aug 5, 2009, at 3:55 PM, Christopher Morrow wrote: >>>>> >>>> >>>> This I don't recall at all... I think part of my question is we (as a >>>> group) are assuming that the reasons for requiring ipv6 udp checksums >>>> as stated +10 years ago are still valid, I don't see data supporting >>>> that fact. >>> >>> There are some classic papers on this topic. The most recent I could >>> find >>> is from 2000: >> >> If by 'classic' you mean 'not relevant to today's networking >> technologies', sure. >> >>> http://conferences.sigcomm.org/sigcomm/2000/conf/abstract/9-1.htm >> >> This study captured data on: >> >> 1) a web crawl server on a 10mb hub >> 2) a dorm on a broadcast 10base2 network >> 3) 2 locations that don't give enough information about the >> connections/tech used >> >> The dataset analyzed is not relevant to today's networking >> connectivity or technologies. Looking very quickly at a small set of >> data I have access to (servers serving web content to the internet >> users): >> >> 32,945,810,591 packets received, 0 dropped due to bad checksum (ip >> header checksum) >> >> 1,004,728,008 datagrams received, 0 bad checksum, 15886 with no >> checksum (udp datagram stats) > > Just polling the routers here, I see a similar scenario, e.g., > > 4,166,900,871 packets 0 dropped due to bad checksum
neat! (I'm also going to see if I can get some stats from a wider set of hosts, but....) > However, this is over good clean fiber links. What I would worry about are > RF links, such as 802.11 or P2P microwaves. These generally have link layer > redundancy / checksums, etc., which I think are pretty good at detecting > corrupted packets, > > it might not be wise to rely on that. > > I wonder if we could dig up similar numbers for the IETF 802.11 network. I > will make inquiries. > > Regards > Marshall > > >> >> >> (collected from some unix hosts, via netstat -s or netstat -s -p udp >> output) >> >> Given this set of data I don't think that having a checksum matters >> for UDP or IP-header on today's internet, since there are zero errors >> out of ~34B packets. >> >> -chris >> >>> In a quick re-read of this paper, I didn't see anything that is obviously >>> dated about it. So, I'd assume its error rates and analysis are still >>> pertinent, unless there is a more recent study that says otherwise. >>> >>> Margaret >>> >>> >>> >>> >> -------------------------------------------------------------------- >> IETF IPv6 working group mailing list >> ipv6@ietf.org >> Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 >> -------------------------------------------------------------------- >> > > -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
