The Connected Web

A South African IT company has transferred 4GB of data between its offices by carrier pigeon, the BBC reports, achieving a data transfer rate 25 times faster than a simultaneous transmission sent via ADSL:

"A Durban IT company pitted an 11-month-old bird armed with a 4GB memory stick against the ADSL service from the country's biggest web firm, Telkom. Winston the pigeon took two hours to carry the data 60 miles — in the same time the ADSL had sent 4% of the data. Telkom said it was not responsible for the firm's slow internet speeds."

The story reminds me of an interview published several years ago with since-lost data storage guru Jim Gray, in which he discussed the merits of transferring large volumes of data via UPS in preference to broadband:

"I'm sending complete computers. We're now into the 2-terabyte realm ... I tend to write a terabyte in about 8 to 10 hours locally. I can send it via UPS anywhere in the US. That turns out to be about seven megabytes per second ... UPS takes 24 hours, and 9 hours at each end to do the copy."

Recalling the notion of 'sneaker net', an early belt-and-braces form of networking in which "you would pull out floppy disks, run across the room in your sneakers, and plug the floppy into another machine," Gray described his storage-appliance-by-UPS method as TeraScale SneakerNet. Even taking the cost of the hardware into account, it was still cheaper than the cost of using up all that bandwidth, he added. Today the economics would probably be even better if Backblaze's sub-$8000 67TB storage pods were used.

Of course, data transfer by carrier pigeon (or even some couriers) is neither as secure nor reliable as using the cloud. But the moral of the story is to remember just how long it can take to transfer large volumes of data across the cloud when planning for backup and disaster recovery. A presentation I attended last year at CloudCamp London by Alan Williamson of RSS traffic analytics provider Mediafed underlined how the volumes can rapidly mount: it had so much data in the cloud, it had calculated that it would take three weeks to download a complete back-up of its S3 data.