Ratios & peering [was: Level 3 Communications Issues Statement Concerning Comcast's Actions]
mpetach at netflight.com
Tue Nov 30 13:46:27 CST 2010
On Mon, Nov 29, 2010 at 8:47 PM, Patrick W. Gilmore <patrick at ianai.net> wrote:
> On Nov 29, 2010, at 6:34 PM, Seth Mattinen wrote:
>> My take on this is that settlement free peering only remains free as
>> long as it is beneficial to both sides, i.e. equal amounts of traffic
>> exchanged. If it becomes wildly lopsided in one direction, then it
>> becomes more like paying for transit.
> [*] 10 second explanation for those who do not understand: I hand you a small HTTP GET request, you carry it across the country. You had me a 1500 byte web page, I carry it across the country. My costs are much higher than yours, you need to compensate me for the additional costs.
Clearly, to balance out the traffic ratios, content providers should set their
server MTUs to 64 bytes. That way, small HTTP request packets will be
nicely balanced out by small HTTP reply packets. If the content providers
also turn off SACK, and force ACKs for each packet, they can achieve
nearly the perfect traffic ratios the eyeball networks seem to desire.
Small packet one way, equivalent small packet the other way, and
everyone is happy.
Obviously those recent infidels pushing for the so-called "Jumbo Frames"
here on NANOG were nothing more than shills for the eyeball networks,
seeking to get more and more networks out of ratio, in an effort to get
them to cough up money. Fie on them, I say--instead of JumboFrames,
we need MicroFrames! Exchange points should start enforcing a maximum
frame size of 64 bytes, to truly bring the internet into perfectly-balanced
(*in search of forceps to extract a tongue planted far too forcefully
into the cheek*)
More information about the NANOG