Brad Burnham 02 February 2006 Comments

Mathematics - How much is enough

After eBay’s purchase of Skype, it became clear to many of us that peer to peer file sharing networks were a potentially serious information technology services platform, so I got to thinking about how big this infrastructure really was.

iSupply reported last fall that total PC shipments for 2005 would be 211 million. IDC recently pegged the number at 209 million.

So let’s assume around 210 million PCs were shipped last year. I tried to get a handle on the size of the average hard drive by playing around with different search terms in Google, but this is not something Google does well. The “expert” on the sales floor at CompUSA told me the average drive they sell these days is 80 Gigabytes.

My friend Sam tells me that, if those numbers are correct, the total storage capacity of the PCs shipped last year is 16.8 Exabytes (10 to the 18th). Sam also pointed me to this post thinking it might help me wrap my head around how may bytes that really was.

Encouraged, because this seemed like a lot of bytes, I began to wonder if we would ever be able to store more bytes than we could produce.

It turns out that Gordon Bell, now at the Microsoft’s Bay Area Research Center has been thinking a lot about how many bytes a person can produce in a lifetime. In this related note he says that as long as you don’t want to store video it is eminently feasible to store a lifetime’s work. Despite Gordon’s admonition to stay away from video, I thought the best way to think about the most we would ever want to store was to imagine that every person on the planet would store full motion high definition video of every moment of their lives.

So I went back to my friend Sam, who told me the following. It takes about 1.6 Gigabytes to store an hour of HDTV or roughly 40 Gigabytes for a 24 hour day. That means it would take about 15 Terabytes to store an HD video of one persons life for one year. If you multiplied that by the roughly 6 Billion people on the planet, we would need about 87.6 ZettaBytes (10 to the 21st) per year to store all the bytes we could produce in that year.

As foggy as my memory of math is, I am aware that that is a big number and that the gap between 10 to the 18th and 10 to the 21st is substantial. What is remarkable to me, though, is that once you have stared those numbers in the face, you realize that if the storage capacity of PCs and the number of PCs shipped annually both continue to grow at the same pace, and you make the assumption that there is a limit to the number of bytes a human can produce, we are not that far away from the day when we will be able store all the bytes we can produce right on our PCs.

One could toss that away as another useless piece of information, except that it suggests that there will be a day – sooner than we may have thought – when our most talented engineering resources will turn their attention away from the problem of storing more bytes and to the problem of making better use of the bytes we have already stored. That transition, which may already be underway, will have a huge impact on the development of the information technology business.

I did not study a lot of science in high school so this reference to mathematics is likely to be my last effort to come up with a model from high school science that sheds light on the evolution of the information technology business.\n\nAfter eBay’s purchase of Skype, it became clear to many of us that peer to peer file sharing networks were a potentially serious information technology services platform, so I got to thinking about how big this infrastructure really was.\n\niSupply reported last fall that total PC shipments for 2005 would be 211 million. IDC recently pegged the number at 209 million. \n\n \nSo let’s assume around 210 million PCs were shipped last year. I tried to get a handle on the size of the average hard drive by playing around with different search terms in Google, but this is not something Google does well. The “expert” on the sales floor at CompUSA told me the average drive they sell these days is 80 Gigabytes.\n\nMy friend Sam tells me that, if those numbers are correct, the total storage capacity of the PCs shipped last year is 16.8 Exabytes (10 to the 18th). Sam also pointed me to this post thinking it might help me wrap my head around how may bytes that really was.\n\nEncouraged, because this seemed like a lot of bytes, I began to wonder if we would ever be able to store more bytes than we could produce. \n\nIt turns out that Gordon Bell, now at the Microsoft’s Bay Area Research Center has been thinking a lot about how many bytes a person can produce in a lifetime. In this related note he says that as long as you don’t want to store video it is eminently feasible to store a lifetime’s work. Despite Gordon’s admonition to stay away from video, I thought the best way to think about the most we would ever want to store was to imagine that every person on the planet would store full motion high definition video of every moment of their lives. \n\nSo I went back to my friend Sam, who told me the following. It takes about 1.6 Gigabytes to store an hour of HDTV or roughly 40 Gigabytes for a 24 hour day. That means it would take about 15 Terabytes to store an HD video of one persons life for one year. If you multiplied that by the roughly 6 Billion people on the planet, we would need about 87.6 ZettaBytes (10 to the 21st) per year to store all the bytes we could produce in that year.\n\nAs foggy as my memory of math is, I am aware that that is a big number and that the gap between 10 to the 18th and 10 to the 21st is substantial. What is remarkable to me, though, is that once you have stared those numbers in the face, you realize that if the storage capacity of PCs and the number of PCs shipped annually both continue to grow at the same pace, and you make the assumption that there is a limit to the number of bytes a human can produce, we are not that far away from the day when we will be able store all the bytes we can produce right on our PCs.\n\nOne could toss that away as another useless piece of information, except that it suggests that there will be a day – sooner than we may have thought – when our most talented engineering resources will turn their attention away from the problem of storing more bytes and to the problem of making better use of the bytes we have already stored. That transition, which may already be underway, will have a huge impact on the development of the information technology business.