The end of the web as we know it?

Is the internet in imminent danger of falling around our ears?

You could be forgiven for thinking so, after reports of a speech by Patrick Gelsinger, senior vice president and chief technology officer at the leading microchip manufacturer Intel, warning that ‘the worldwide web could collapse in the next few years’ (1).

But what Gelsinger actually said was that the technology underlying the internet is inadequate to meet our current and future needs – which is hardly surprising, when you consider how old this technology is, and how our needs are continually evolving. The really interesting question is how we take the internet forward.

The internet is showing real shortcomings, as users increase and applications diversify. But these shortcomings tend to be bundled together with broader concerns over spam, viruses, hacking, and all the other sundry ills of the world. The result is a general gloom.

As the online publication the Register puts it, ‘the noise quotient of the internet has increased somewhat and, what with all the spam, the pop-up windows, the blog noise and Google being broken, things just aren’t what they used to be’. The Register has also highlighted the problem of unwanted ‘background radiation’ on the internet, ‘as internet static is multiplied exponentially’, and the problem of ‘a stream of nasty pollution we call malicious traffic, that will never stop’ (2).

Elsewhere, Steve Linford, head of the anti-spam project Spamhaus, has predicted that the internet may be only six months away from ‘meltdown’ due to spam. The only thing is, he predicted this 13 months ago (3).

Pat Gelsinger’s message earlier this month was actually more optimistic. Speaking at the Intel Developer Forum in San Francisco, his speech was preceded by an onstage conversation between himself and technology pioneer Vinton Cerf, widely known as the father of the internet. By using this order of line-up, Intel was sending out the audacious message that it will lead the next wave of fundamental innovation. Well, somebody’s got to do it.

Cerf explained that when it comes to the underlying technology, ‘the basic protocols are still pretty much unchanged from what we specified back in the 1970s’. While this testifies to the ingenuity of his groundbreaking work, it also suggests that we have failed to break new ground since. As Gelsinger put it, the internet is ‘a 1973 Buick’. And in spite of ‘all this work to refit it, reconfigure it and add things to it…under the hood, it’s still a 1973 Buick’ (4).

Some important work is being done to improve the system. For instance, the current standard for internet protocol addresses, IPv4, is being replaced by the superior standard IPv6 (see Unleashing the net, by Sandy Starr). And the Internet Corporation for Assigned Names and Numbers – one of the key organisations overseeing the technical administration of the internet, chaired by Cerf – recently launched IPv6 on its root servers (5).

But problems with the system remain. In the words of the Internet Mark 2 Project, which addresses ‘the range of problems which have appeared as the internet gets older and bigger’, ‘the internet is required to perform a number of important functions not included in the original design. New protocols have been developed, and various patches have been applied to base protocols, not always evenly. It seems appropriate to examine whether the current system, people, and processes are still appropriate.’ (6)

The key protocols which Cerf helped to pioneer – the transmission control protocol and the internet protocol (known collectively as TCP/IP) – allowed different kinds of technological systems to exchange information easily, so that users wouldn’t have to concern themselves with the systems’ underlying complexity. As Gelsinger put it in his Intel Developer Forum speech, the invention of TCP/IP ‘overlaid, it hid, it pushed down that complexity, and ignorance was bliss’. But given the new demands we are placing upon these 30-year-old protocols, ‘all of that underlying complexity and heterogeneity is re-emerging’. So the challenge, according to Gelsinger, is ‘creating an overlay that allows us to again abstract and look past much of that complexity’.

The trouble is, when Cerf was first working on the internet it was the arcane province of academia and government, whereas today it plays a key role in business and society worldwide. There are now so many competing interests bound up in this technology, that seeking agreement on fundamental changes to it is no easy task. As Cerf put it, ‘the net’s really a big collective entity. It’s a federation of many different companies and organisations’. It is therefore ‘impossible to just rip everything out and replace the existing infrastructure in a short period of time’.

So rather than proposing the radical step of revising or replacing the key existing protocols – which is what the Internet Mark 2 Project would like to see happen ‘within a five- to 10-year time frame’ – Intel is proposing instead to add a new overlay to these protocols, just as these protocols were themselves laid over pre-existing systems. Gelsinger argued that ‘it’s time to rethink and replicate that exact architectural thinking for the situation of today’, and described Intel’s plans for a ‘computational services overlay, allowing us to…abstract and isolate many of the difficulties of that underlying network layer’.

Intel’s proposals may be ambitious, but it’s worth asking just how innovative you’re being when you set out to ‘replicate’ the ‘exact architectural thinking’ of 30 years ago. If innovation is understood in terms of the extent to which it disrupts and transforms what has gone before, then a truly innovative change to the internet would ‘rip everything out and replace the existing infrastructure’, as Cerf put it (8).

Such drastic changes aren’t even on the agenda today. This indicates that even the most ambitious proposed changes to the internet aren’t that innovative by past standards.

The main model flagged up by Gelsinger was the PlanetLab project. PlanetLab is ‘an open platform for developing, deploying and accessing planetary-scale services’, and these services include overlay networks of the type being developed by Intel. PlanetLab describes itself as, among other things, ‘a microcosm of the next internet’ (9).

This sounds inspiring, and it is, but these technologies have their problems. Intel boasts that PlanetLab ‘could…enable new capabilities, such as persistent storage, the idea of giving the internet a “memory”. For example, 100 years from now a piece of data could still be found, even though the original computer on which it was posted no longer exists. In addition, this research could influence the future design of servers and network processors.’ (10) Those of us concerned about the latitude that the authorities have to spy on us, would be suspicious about an internet with a 100-year memory.

Technological developments of this scope are double-edged. They bring us new opportunities to communicate and collaborate, but they also bring the authorities new opportunities to collate information about us and pursue their own dubious interests. All of us are inconvenienced by problems such as spam and viruses, but the benevolent heading of ‘security’ – under which measures to counter these problems are usually proposed – also means interference in our affairs and regulation of our speech and behaviour.

The inflexibility of certain basic aspects of the internet has been an impediment to authoritarian measures. Pressure has long been exerted by the authorities to amend basic internet technology in order to make governance and regulation easier. Any significant change to the technology underlying the internet, including that proposed by Intel, will almost certainly be seized upon to tighten up regulation.

At the Intel Developer Forum, Cerf explained the ‘regulatory problem, as governments wade in and try to come to terms with this evolving internet’, while Gelsinger described ‘a freight train of regulatory concerns coming at the internet’. More worryingly, Gelsinger argued that since ‘privacy and security laws’ around the world are ‘different’, ‘we can’t embed those into underlying protocols. We need the network to be able to respond and address those different regulatory regimes.’ (11)

Technology enthusiasts’ idea of the internet as an unregulable Wild West was always something of a fantasy. But the international dimension of the internet has helped to safeguard freedom, because a decentralised medium evades the rule of law in specific jurisdictions.

An internet whose underlying technology was designed to accommodate ‘different regulatory regimes’, on the other hand, would not be the universal medium we know today. It would be divvied up and dragged back down to national limits.

Rather than debating the substance of Intel’s proposals, critics have focused instead upon Intel’s status as a Big Bad Corporation. It’s true to point out, as does the Observer’s technology columnist John Naughton, that ‘Intel resembles a drunken Ibiza clubber, in that both are motivated by a single, overweening objective. The clubber’s is to get laid; Intel’s is to make sure that every electronic device sold has one of its chips inside it.’

But it’s not surprising that changes to the internet are going to be driven by organisations with powerful commercial interests. Indeed, greater investment in this technology by major players would be a welcome change, given the underinvestment that has come with the post-dotcom boom hangover. The enlightened attitude is not to regard all commercial interests with suspicion, but to ask to what extent they correspond with our collective interest.

So Naughton’s cynical comment that ‘since all the nodes of this new layer are based on Intel chips, it might be better to call it the Intelnet’, and remarks by other commentators that Intel’s proposals amount to ‘Chipzilla’, impugn the company for the most positive aspect of its intervention in this field – the fact that it is willing to step in, be ambitious, and take the lead (13). What we should be impugning instead is the way in which Intel seems willing to usher in a new regime of stifling regulation.

Innovation in information technology presents the libertarian with a dilemma. Innovation is crucial, if we are to realise the potential of this technology to change the way we work, play and live for the better. But such innovation tends to be put to uses that undermine our freedom, especially given the illiberal tenor of our times.

To stick with decrepit technology, on the grounds that replacing it opens up a can of regulatory worms, smacks of Luddism. Not only should we be championing innovation in order to meet our needs – we should also be raising our expectations of what constitutes innovation, and arguing that there is ultimately no limit as to what our needs might be.

We shouldn’t be afraid to make changes to the technology that underpins the internet, and we should give short shrift to the idea that any changes are necessarily tainted by big business. But we should be sensitive to who is driving these changes and who benefits from them. It is not Luddism or facile anti-capitalism that will protect us from the Big Brother of the future, but a ruthlessly critical attitude toward regulation, and a staunch defence of freedom.

According to Vint Cerf, the internet is ‘still pretty primitive…kind of in the stone age’ (12). Here’s to moving the internet out of the stone age, but keeping our wits about us while we do it.

Sandy Starr has consulted and written on internet regulation for the Organisation for Security and Cooperation in Europe, and for the European Commission research project RightsWatch. He is a contributor to Spreading the Word on the Internet: Sixteen Answers to Four Questions, Organisation for Security and Cooperation in Europe, 2003 (download this book (.pdf 576 KB)); From Quill to Cursor: Freedom of the Media in the Digital Era, Organisation for Security and Cooperation in Europe, 2003 (download this book (.pdf 399 KB)); and The Internet: Brave New World?, Hodder Murray, 2002 (buy this book from Amazon (UK) or Amazon (USA)).

Read on:

Can technology can spam?, by Sandy Starr

‘Responsible’ regulation, by Sandy Starr

Standard questions, by Mark Birbeck

Trusting technology, by Norman Lewis

(1) Internet faces collapse, warn computer giants, Owen Gibson, Guardian, 14 September 2004

(2) Watching the Net’s background radiation, Andrew Orlowski, Register, 27 November 2003; The polluted internet, Kelly Martin, Register, 27 August 2004

(3) See Sick of spam? Then blame Alan Ralsky. He emails a billion of them a day, Paul Harris, Observer, 1 June 2003

(4) Keynote speech, Patrick P Gelsinger and Vinton G Cerf, Intel Developer Forum, 9 September 2004

(5) See Next-generation IPv6 address added to the internet’s root DNS zone, Internet Corporation for Assigned Names and Numbers, 20 July 2004. See the definitions of IPv6 and root server system, on the Whatis.com website

(6) Homepage of the Internet Mark 2 Project website; Executive summary (.pdf 217 KB), Internet Analysis Report: 2004 – Protocols and Governance, Ian Peter, Internet Mark 2 Project, 2004, p4

(7) Keynote speech, Patrick P Gelsinger and Vinton G Cerf, Intel Developer Forum, 9 September 2004; Executive summary (.pdf 217 KB), Internet Analysis Report: 2004 – Protocols and Governance, Ian Peter, Internet Mark 2 Project, 2004, p12

(8) PlanetLab is…, on the PlanetLab website

(9) Intel, HP join top academic researchers to expand the usefulness of the internet, Intel, 24 June 2003. See the definitions of protocol, transmission control protocol and internet protocol, on the Whatis.com website

(10) Keynote speech, Patrick P Gelsinger and Vinton G Cerf, Intel Developer Forum, 9 September 2004

(11) Chips down for net as we know it, John Naughton, Observer, 19 September 2004; Intel vision of the future, p2pnet.net, 9 September 2004

(12) Keynote speech, Patrick P Gelsinger and Vinton G Cerf, Intel Developer Forum, 9 September 2004