Thursday 28 March 2013

Package to packet to package

I've been in the internet business longer than most.  I sent my first trans-Atlantic e-mail in 1983, I marketed a book on the internet in 1985, I wrote one of the first web-pages in Ireland in late 1994.  

 In 1991 work was analyzing DNA sequences, so we used regularly to receive a copy of GenBank, the database of all the DNA sequences known to science, on a 2400 ft magnetic tape.  This was the size of an inch-thick dinner plate and arrived in a padded envelope from France.  After downloading the data, the tape itself was sent on to Indiana, then Houston and finally back to Lyon.  By mid 1992, we had left 19th century communication behind and were downloading GenBank over the internet.  The whole database then was only about 200 Megabytes in size (say, three MP4 TED talks in today’s e-currency) – it’s 1000x bigger now – but the connexion  was slow (overnight at least) and unreliable.  Traffic was lighter and the connexion accordingly faster and more certain at weekends. All too often, however, I’d come in on Monday morning to find that only a fragment of the database had arrived together with a laconic “BROKEN PIPE” error message – which evoked an image of packets (technical term for an aliquot of data) of As Ts Cs and Gs spewing out across the Atlantic seabed.  So I took to going into town on Saturday or Sunday to check progress and start it all again if the process had failed.  This was conscientious but tedious, and I was delighted when, foraging about in the basement at work, I found an acoustic coupler and brought it home.  You could plug the coupler into a computer and telephone at home and send commands to the server at work albeit at less than the speed of light – maximum 300 bytes/sec.  The phone handset sat into a foam-rubber surround to minimize noise and maximize signal but you could still hear it fizzing and clicking.  With that appropriate technology, from home I was able to . . . C H E C K  P R O G R E S S  at work and R E S T A R T   J O B  if necessary.  This was – marginally – quicker than making a 25km round trip on my bike.   

Over the next few months and years it got easier - the wires got fatter, connexions more reliable.  Perhaps more importantly, Genbank got too big to download locally at about the same pace at which it became possible to do all the necessary analysis off site on what we were still years from calling The Cloud.  It became possible because capable graduate students and tech-savvy post-docs across the globe were writing code to abstract particular sequences from the database and analyse them in particular, and even peculiar, ways.  Of course HTML, WWW and HTTP made it all so much easier to produce functional procedures than the primitive comms we'd used before.

Now here's the weird, I was in the pub before Christmas last year talking to a couple of the young effectives in our world of bioinformatics.  One of them said that he was routinely communicating with his US collaborators by FedExing a terabyte external hard-drive back and forth across the Atlantic. Parkinson's_law states that "work expands to fill the time available for its completion".  If there isn't already a similar law like "data proliferates to fill to capacity the conveniently accessible means for storing it" then I hereby claim/name it as Bob's Law of the Data Packet.

Was it the Red Queen or Madame de Pompadour who quipped about surfing the deluge to keep abreast of it?  I forget.  "Apres nous la brioche" was it?  "Surfing’s the source code man . . . swear to God"  Now that was Bodhi in Point Break.

No comments:

Post a Comment