Saturday, November 11, 2006

NETI@Home

Network neutrality nightmare scenarios are largely hypothetical: Bad thing X could happen if a network operator did Y. Activist outrage and some media attention is one reason we’re still in the realm of conjecture. Even if network operators were minded to do Y, they'd rather not draw the spotlight. How to keep the klieg lights shining?

Since discrimination, in the neutral sense of making distinctions, can have both good and bad outcomes [1], regulations will also have to make subtle distinctions. It’s even harder to make law about subtle hypotheticals. I’m therefore inclined against detailed legislation in advance of facts. Legislation, if any, can lower the risk of unintended side effects by simply establishing principles that a regulatory agency can apply to alleged bad behavior. But agencies don’t have the means to gather data. Even companies with an interest in the matter operate in the dark; I’ve heard that Vonage found out about the discrimination against their VOIP service in the Madison River case only by accident. How to find bad behavior?

The answer is millions of volunteer PCs on the net sending test payloads to each other to characterize what happens to traffic on the net. This would yield an inventory of how different kinds of traffic are carried across all the different segments of the net. Software could identify potential problems, and volunteer people could scan them to identify cases that need more attention.

Anybody could download small applications that would run tests when they’re not using their machines. Just as SETI@Home uses spare CPU cycles to search radio data for the signature of extra-terrestrial intelligence, so this code would use space cycles and spare bandwidth to search for signatures of net neutrality exceptions. SETI for the net… NETI@Home: Neutrality Exception Tracking Infrastructure.

NETI@Home would use peer-to-peer technology like bittorrent; there would be no single repository of data, and no central organization directing the work. People could select the kinds of test payload they want to run; to make it easy for the majority, various recognizable organizations might recommend payload sets, or one could find popular sets on sites like digg. Volunteers would write the software, and parse the results. The tests could change quite quickly to meet changing network operator strategies, while the underlying software would change more slowly.

Citizens and regulators need good data if they’re to get the best out of the Internet. Finding the balance Jon Peha describes between allowing discrimination that benefits users, and preventing market power abuses, will be a lot easier with real-time tracking of network operator behavior.

----------

[1] Jon M. Peha, “The Benefits and Risks of Mandating Network Neutrality, and the Quest for a Balanced Policy,” 34th Telecommunications Policy Research Conference, Sept. 2006, at http://web.si.umich.edu/tprc/papers/2006/574/Peha_balanced_net_neutrality_policy.pdf

3 comments:

Anonymous said...

Very intriguing idea. It does raises some questions as to why would somebody install it on his own PC. I mean, SETI@Home gives you the opportunity of be part of something that could potentially change mankind history. I can see a group of people that would love to have that, like everybody on the EFF, but the public at large would have little motivation.

However, if somebody like Google distribute this with their toolbar, even if you are explicitly asked to participate, you could have 1 million users instantenously.

That would also have other benefits, like understand and optimizng the net.

Anonymous said...

There is a NETI@home already. See neti.gatech.edu/ Are you saying you thought this up completely independently? Or could it be that you had heard about it somewhere in the past and it just popped back into your head?

JP (Pierre) de Vries said...

[blush] I should've googled the name before posting... Thanks for the info. I think I thought of it independently, and I don't recall ever having heard of the Georgia work - but who knows? Maybe it just popped back up. More proof, if proof were needed, of the sobering rule that if I can think of something, someone else's already doing it.

I'll look at their stuff - hopefully they already do the kind of neutrality testing that I have in mind.