disrupted the music industry. P2P technology certainly did that. Skype shook up telecoms and Joost may do the same for TV.Word association: P2P = Napster =
P2P (Peer to Peer) networks could be a lot more. They could be the last big disruptive technology online. The online world is increasingly getting divided into a few mega firms that can afford to invest billions in server farms (Google, Amazon, Microsoft, Yahoo, eBay) and the rest of us. Start-ups may love to have that “pay as you go infrastructure”, but it does create a huge barrier between the big league and everybody else.
It almost makes you want to love that old kludgy PC. For all its myriad faults, the PC was/is a great leveler; we all get the same tools at our disposal. I like the freedom and lack of hassle from having it all out there “in the cloud”; but there is a nagging concern that somebody will hold me hostage to all that. We are already seeing what happens when somebody wants to get their data out of Facebook.
It is now accepted wisdom that the search game is over and Google has won. Wikia, one of the well-funded attempts to take on Google, launched to a chorus of boos. I took a look at Powerset, another well-funded and well-hyped attempt, using natural language search and was seriously underwhelmed. Microsoft buying FAST for $1.2bn will make some enterprise $$$ but won’t impact the broad consumer market.
The barriers are not search algorithms. That’s easy. The first barrier is overcoming habit. I think that a genuinely better alternative will change habits. That worked for Google and with search boxes having multiple alternatives it really is simple to switch.
The second barrier is that success requires investing $$$ billions in server farms. Now that VCs have pumped money into lots of attempts and they are not seeming to get much traction we are entering a phase where nobody will put significant new money behind a search venture. This is like the decade or so when nobody would invest to challenge Microsoft.
That is bad news for everybody. It creates a monoculture and that’s always bad.
P2P networks are the only way to overcome that second barrier, avoiding the need for $$$ billions in server farms. P2P uses the infrastructure of the users.
I like the look of Faroo, the Peer to Peer Web Search company. There are some side benefits related to privacy and SEO spam control, but the core advantage of P2P for search is a very simple economic proposition. Because Faroo don’t need to invest in server farms they can return 50% of the search revenue to users. My guess is that models will evolve that give way more than 50% back to users. The payoff from scale without need to invest in scale is so enormous that ventures would be wise to give back whatever it takes to get scale. That is simple “show me the money”.
Project out a few years of Internet growth and the basic economics of a centralized index start to look a bit suspect. How big can those server farms get?
I am not qualified to decide whether P2P networks can scale. There is plenty of well-informed debate online on this subject. I do know that my Skype button shows over 10m users online right now; so I am inclined to believe those who say that P2P networks can scale.
Of course, Faroo suffers from the mother of all chicken and egg problems. Their model is based on indexing the sites that all the Faroo searchers visit. Until that number is measured in millions it will be a useless service. Faroo is clearly aware of this issue but I am not sure they have a good answer for it.
So the real breakthrough may come from one of the existing P2P networks. The problem here looks like a mind-boggling array of alternatives, each of which has their ardent admirers. This is one area where de factor standardization would help. As long as the P2P winner is open source who cares who wins? There is no toll gate being erected.
It would then be possible to layer a Faroo like search application on top. The basic functionality of a search engine is not complex. The network effect is what matters.
One company that may have some of the pieces is Limewire. It is open source, runs on the Gnutella network, it is spyware free (they monetize on a freemium model, with an upgrade to a paid Pro edition) and seems to be backed by a stable company.
Another company that could make a play here is Sun. After all, P2P is totally in synch with their ever-prescient “the network is the computer”. That has a better ring than “the server farm is the computer”. If Sun enabled this next disruptive wave they would re-gain their relevance at the center of the IT industry.
The P2P model puts data back on our hard drives. That means fixing the data management, security and other issues that continue to plague PCs. But that would be a small price for re-gaining control over our data. Social networks believe that my data is an asset for them to sell to advertisers (in return for some simple features). I disagree and so do many other people. This is a visceral issue. Its my data, dammit. P2P puts data back in user’s control.
Surprisingly, there is very little buzz about P2P today. Searching on P2P gets posts/articles that are mostly 2001/2002 vintage. So I am sure nobody will Digg this post or comment on it. But I do harbor a suspicion that the one big disruptive play in search will come from P2P.