Saturday, August 31, 2013

Tubes: Behind the Scenes at the Internet (Andrew Blum)

When first published, Andrew Blum’s book Tubes: Behind the Scenes at the Internet was much reviewed and dissected in techie corners of traditional media as well as the internet. I finally got around to reading it while on holiday.

Various internet failures – at a country-wide level as well as down the side of his own living room couch – led the writer and technology correspondent to track down how the internet worked. But instead of examining the protocols and software stacks involved, he traced the people who help connect the physical network ‘tubes’ together to bring data from one side of the world to the other.

In 2006 Alaskan Senator and chair of the US Senate Commerce, Science and Transportation Committee Ted Stevens was ridiculed in the media when he described the internet as “a series of tubes”. Yet Andrew Blum’s travels took him to across the world to see huge fibre bundles being pulled across oceans to connect continents together; internet exchange (IX) facilities hooking networks together; a regional Network Operators Group where expertise was shared and new peering agreements were formed; and the seventy five year old professor who installed one of the first nodes of ARPANET at UCLA back in 1969. Lands End, Amsterdam, and rural America.

Reading Tubes reminded me of 1990s articles in trade magazines like DataComm that excitedly explained how the MAE-East internet exchange was based in a parking garage.

Andrew Blum discovered that location was important yet unpredictable. Many of the buildings be visited were both anonymous and remote. The reasons why an internet facility evolved and grew at a particular site would be varied and the consequences often long-lived with enormous datacentres now being built seemingly in the middle of nowhere due to the nearby high-bandwidth connectivity and cheap sources of energy.

At an internet exchange in Ashbury he poetically wrote:
There are different kinds of connect. There are the connections between people, the million kinds of love. There are the connections between computers, expressed in algorithms and protocols. But this was the Internet’s connection to the earth, the seam between the global brain and the geologic crust … among the landscapes of the Internet, it was the confluence of mighty rivers, the entrance to a grand harbour. But there was no lighthouse or marker. It was all underground, still and dark—although made of light.

Often, Andrew Blum fell on his feet with enthusiastic and charmingly eccentric guides – internet infrastructure speleologists – who helped him understand how the tubes fitted together and how the data moved around. In marked contrast, he visited a Google data centre but didn’t get any further than the lunchroom. Facebook were more open and welcoming. The book closes with a perceptive comment from one of their data centre managers, Ken Pratchett:
If you lose rural America, you lose your infrastructure and your food. It’s incumbent for us to wire everybody, not just urban America. The 20 percent of the people living on 80 percent of the land will be left behind. Without what rural America provides to urban America, urban America couldn’t exist. And vice versa. We have this partnership.

Andrew Blum added:
If in Oregon that was once about timber and beef, it now extended to data, of all things. The Internet was unevenly distributed. It wasn’t everywhere at all—and the places where it wasn’t suffered for it.

While technology-based, the book is equally accessible to nerds and non-geeks. A good primer on how the internet works and how it breaks.

There must be a follow-up book for Andrew Blum – though perhaps not so universally appealing – to talk to the men and women who established some of the processes and software that are still fundamental to the internet’s operation: the people behind internet address allocation, DNS, SMTP, Mosaic, Netscape, etc as well as information entrepreneurs like Rick Gates who conceived and ran The Internet Hunt from 1992-5.

(In the days before web browsers and search engines, you had to answer Rick’s questions by explaining where on the internet – using FTP and near-extinct tools like Gopher – you could find the answers online. Often there was a bonus question which he didn’t have the answer to but would inevitably be found somewhere online. Good memories from my days at QUB!)

No comments: