Packet switching: the hot potato technology 31 Mar 2011Posted by Tony Law in Impact of IT, ITasITis, Technorati.
Packet switching is technology at the core of today’s networked world. But it wasn’t always so.
Today’s press carries tributes following the death of Paul Baran, one of two researchers who independently defined the fundamentals of packet switching (the other was Donald Davies in the UK). Who? you ask – outside academic computer science I doubt if most of us have heard of either of them. Independently they developed the fundamental concepts of packet switched networks: breaking messages up into small self-addressing units (so that the network doesn’t get clogged by large messages), which travel independently through the network and are re-assembled in the correct order at the destination. Baran, apparently, described this as handing-off the units “like a hot potato”.
The abstract to one of Baran’s early papers (Baran, 1964) describes this
“… distributed communication network concept in which each station is connected to all adjacent stations rather than to a few switching points, as in a centralized system. The payoff for a distributed configuration in terms of survivability in the cases of enemy attack directed against nodes, links or combinations of nodes and links is demonstrated. … [The] feasibility of using low-cost unreliable communication links, even links so unreliable as to be unusable in present type networks, to form highly reliable networks is discussed. The requirements for a future all-digital data distributed network which provides common user service for a wide range of users having different requirements is considered. The use of a standard format message block permits building relatively simple switching mechanisms using an adaptive store-and-forward routing policy to handle all forms of digital data including digital voice. This network rapidly responds to changes in the network status.”
It seems to be the classic story of a technology development which probably wouldn’t get funded in a utilitarian commercially-oriented research assessment. There were technical objections and, in the early 1960s, it would have been expensive to build. But, as is well documented, the US Defense Department was looking for resilient technologies to reduce the vulnerability of their networks. Packet switching (the name was coined by Davies, not Biran) was part of the answer. And, the political environment was one where super-powers were trying to reduce the likelihood of accidental military catastrophe: recognising that communications would play a key role, the work was never classified as Secret. The rest, as they say, is history.
Baran went on to found the influential Institute for the Future (I didn’t realise how long-established that institution is: it dates to 1968) and several technology-based companies. Packet switching went on to become the foundation of modern networking. It’s even displacing analogue networking from its natural base in the world of telephony, where implementation has to make a fundamental switch. For data, it’s important that all the packets arrive and are reassembled; loss of a packet is handled by a request for re-transmission. For a pseudo-analogue output such as video or telephony, loss of occasional packets is less important than avoiding gaps in rendering the output. But, as Baran’s early paper shows, he foresaw this use too.
• Paul Baran obituary, Guardian, 30 Mar 2011 (paper publication, 31st Mar)
• Internet Pioneers: Paul Baran, ibiblio.org (this includes one or two links to academic references)
• On Distributed Communications Networks, Baran, P., IEEE Transactions on Communications Systems, 12.1, 1-9 (March 1964), accessible from IEEE Xplore
• Paul Baran and the Origins of the Internet, Rand Corporation (rand.org)
• UK National Physical Laboratory (NPL) & Donald Davies, Living Internet
For a technical description of packet switching in the context of other network technologies, see e.g. Distributed Systems & Computer Networks, Sloman, M., & Kramer, J., Prentice-Hall, 1987, p 133