Original URL: https://www.theregister.co.uk/2005/05/16/googlenet/
GoogleNet - the ultimate embrace and extend?
End of the web nigh - again?
We've been reading stories about the "end of the internet" ever since the internet was exposed to the public more than a decade ago. Telcos, media companies, infrastructure hardware providers, and operating systems monopolists have all taken their turn as the villain in this particular drama. So far not one has succeeded.
But the really scary thing about "the end of the internet" was never how easily it could be achieved, but that it might be welcomed by most of the people who actually use it. Now Google has helped drive the point home.
At one stage it was clear what the benefits and drawbacks of open and closed networks were. Using the open internet, you could create an application such as a Napster, and have millions of users overnight. You weren't beholden to a telco, or a specific vendor's proprietary protocol to do so. Therefore, a creative software author would succeed on merit, and not because it paid the necessary backhander to the gatekeeper.
The inherent lack of security on the open internet was acknowledged, as were the issues of copyright and rights holders' compensation. But the consensus was that the advantages outweighed the disadvantages, and the faithful believed that technology would eventually fix both of these. Nirvana was just around the corner.
Now, with the open internet riddled with spam, fraudsters, and the toxic payload from the latest worm - and the ghosts of worms past - the advantages are no longer as clear cut. For mobile phone users, spam is the exception, not the rule.
Rights holders' and users will only be satisfied by well-trusted social contracts such as a compulsory license, and not by some as-yet-uninvented technological marvel, we now know. (Digital music can be freely shared, and the rights holders compensated, for about half of what an American mobile phone user pays in chiselling fees and surcharges every month.) And in a Windows world, security will only be fixed by ending the open internet, as we've discussed before, by restricting the capabilities of the client device.
In the dot com era ago, it was hard to find a bright young thing to defend France's antique Minitel computer system, but we can now see it achieved what it was intended to do - facilitate communication, casual sex and shopping - rather cheaply and well, and with a minimum of downtime. ("All over France people don't understand computers, but they're wired anyway, don't really grasp that they're wired, and use the systems like crazy," we observed seven years ago).
Today, unfortunately, it's the open internet evangelists who look like cranks.
So how will it all end?
In his most recent column, Robert X Cringely sketches one scenario, as he discusses the potentially far-reaching implications of Google's beta Web Accelerator proxy.
World Wide Google
Google's proxy dares commit what open internet evangelists say is one of the most heinous crimes imaginable. It puts some intelligence into the network, speeding up web page access for end users. Whether this ploy will succeed or not, we'll set aside for a moment. (Privacy advocates can spare a moment to feel vindicated, although for privacy-concerned end users, the light at the end of the tunnel turned out to be the proverbial train.) You don't have to accept either Cringely's logic, or his conclusion, to see that the idea could be very popular indeed.
The accelerator, he writes, "co-opts every ISP and web page owner. If surfing can be doubled in speed for nothing, of course nearly everyone will go for it. But that means every AOL customer becomes a de facto Google customer and this page becomes a de facto Google service that costs them nothing to produce."
With Google massaging the traffic, Cringely takes this to its natural conclusion.
"If Google adds power to its part of the Accelerator, you don't have to add power to your end, meaning your old PC can last longer. Part of that has to come from Google assuming a larger role over time, taking responsibility for rendering Flash, for example. And they'll do it. And we'll let them. At some point, Google might even offer its own hardware device, optimized for the Accelerator. At that point, you'll buy your PC from Google, use Google as your ISP, surf an Internet that is really the Google cache, be fed ads and sold content from Google servers. Its a GoogleWorld that requires no AOL, no Microsoft, no Intel, no HP or Dell -- only Google, cable companies, telephone companies, users, and of course advertisers and web page producers."
Because this can happen, of course, doesn't mean this will happen. There are many powerful entrenched interests who have reasons to object to Google taking over the internet as we know it. When we last looked there was still competition, of sorts, between advertising brokers, client computer manufacturers, and ISPs. They're sure to object to the "value add" part of their business disappear. In fact, you may disagree with almost all of Cringely's analysis, but you'll have a hard time arguing that end users of the World Wide Wait will turn down a free lunch. Why should they?
Academic discussions of the merits of different kinds of computer networks often dissolve into caricature, with arcane abstractions that bear little resemblance to the end result. But the speculative "GoogleNet" that Cringely outlines introduces entirely a new flavor of network that somewhere between what we think of open and closed networks today . Google can vow not to be a gatekeeper in usual sense, and can probably uphold that claim on a technicality, but it effectively becomes the de facto conduit through which the world's privatized web traffic is piped.
A fascinating prospect in itself. Let's complicate it even further.
East of Eden
We'll modestly suggest one factor which European or American pundits rarely, if ever, acknowledge.
In the US and the EU, the internet is almost synonymous with "the web". If you examine internet coverage in the west since the dot com bubble burst, almost all of it refers to non-web activities which are being recast as web protocols: blogs or social networking as a usenet replacement, for example, or push technology (RSS) as an email replacement, or podcasts as as an ftp replacement.
(This is what happens when you have a lot of unemployed web designers, as we had after the bubble burst, but so it goes).
But in Asia, the web is a minor part of the internet user's activities, and once you deduct Hotmail and Yahoo! Mail usage, it's almost negligible. A typical Asian internet user is online to play LAN games and do instant messaging. In cafes, you'll see "internet" promoted as a separate item below "Games" and "Chat" - almost as if it's an afterthought. A PC in the East is a rather more colorful Minitel that hosts multiplayer games.
Fetishizing the web has blinded many in the West, and not just techno-utopians, to what people really want from computer networks, as successful "closed" computer networks from Minitel, and possibly GSM with its SMS service prove.
So while Western experts lose their minds to utopian fantasies such as "the web is now nature", the Asian mind looks for rather more practical uses for technology. And it's Asia that will be the largest influence on the computer networks in the next decade, with a billion Chinese being urbanized over the next decade, pulling Indo-China along behind it, and half a billion newly affluent on the Indian sub-continent. Will these billions of users want to rely on a GoogleNet?
For those of us who can remember the original Dr Who TV series, a terrible old joke springs to mind. The original human-exterminating Daleks, which resembled large garbage cans, had very effective death rays but moved around on small casters, a not insignificant design flaw. (Fixed in the latest series.) They successfully conquered East Anglia, but were thwarted by the first range of hills.
Some consolation to cling to, then. But seriously, the "Google takes over the Web" scenario outlined by Cringely isn't quite the same as "Google takes over the internet". It's just a whole lot more plausible.
GoogleNet vs AsiaNet, anyone? Or was there ever an "internet" to take over? Give it your best shot. ®
Google puts the brake on Web Accelerator
India, China poised to feast on US IT complacency
Stalling net must dump TCP/IP and overlords Tech heavyweights explain how to destroy the Internet
Internet is dying Prof. Lessig
The Stuckist Net: what is your post-Palladium future?
Daleks to sprout legs