Tuesday
About a year ago I posted this wondering how much data there was in a truck. Dai Davies, director of Dante, has answered my question (via the BBC)
Before now the highest data transfer speed was achieved by putting the tapes in a van and driving them to where they need to be analysed. Delivery vans can carry lots of tapes at the same time which means that Europe's roads have a relatively high bandwidth.
You can send a few hundred megabytes per second through DHL
I am still waiting for an answer to how many trucks there are though.

I suppose the amount of data in a truck can be worked out from the (volume of the inside of the truck) divided by (the volume of each tape) multiplied by (the amount of data on each tape).
As far as bandwidth goes, factors to consider include how many trucks you have, how close together they drive, the time it takes to make the journey (which also affects the latency), and how fast you can load and unload the trucks. (Pretty much the same factors you have to consider when transitting data electronically).
Of course, there's no point transferring data by truck faster than the tape drive(s) can read or write, unless you have a large buffer (er, warehouse).
Posted by Rob Fisher at September 16, 2003 03:05 PM
This is actually an end-of-chapter exercise in the excellent book 'Programming Pearls'. A motorcycle courier can easily carry 400 DVDs in his panniers. This is over a Terabyte of data so the bandwidth is enormous compared to domestic/small office network connections (it would take over a year to put this much data down a 256 kbps DSL line). As Rob Fisher points out, a drawback is that latency is high.
Posted by David Gillies at September 18, 2003 12:12 AM
I recently completed a consolidation project. We moved 2T (Terabytes) of data in a transit van and then in an aircraft. Lyon, France to Gothenborg in Sweden. Took a total of seven hours door to door. Dumping the data to tape took a long time due the the speed and number of tape readers.
We also transfered 2T of data (that was being continuously updated) over two 11M pipes. IIRC the data was fully mirrored in less than four days.
Posted by Matt at September 20, 2003 12:18 AM









