Here’s an interesting thought experiment: how much memory will human civilisation need in 50 years time? By memory, I mean physical bits of information stored in any format imaginable: paper, books, silicon, magnetic disk, nanotubes, optical disks, holograms, or any unheard of as yet mechanism.
Surely there is an upper limit on the need for external memory? That upper limit may be because of lack of a technology capable of storing it, although I think this is doubtful (Moore’s law continues to be valid). To me, it seems more likely that we’ll run out of a requirement for memory sometime soon. But how soon? And what kind of applications can be imagined that continually demand more and more memory?
What will our requirements look like in 50 years time? Predicting how the world will look in 50 years time is an inherently risky task, but I’ll attempt it for the fun. I shall introduce a measuring system to make things simpler. We shall call the current memory requirement for any type of storage “c” and precede the c with an initial letter to distinguish it from other types of memory. Hence, “Mc” will refer to the current, worldwide amount of memory dedicated to storing music (Vc for video, Hc for health, and so on). In our thought experiment then, the requirement in 50 years will be expressed as a multiple of that current capacity: e.g. 10Mc or 100Mc, etc.
Let’s assume that people will still want to store music to carry with them. Let’s assume that music production reaches epidemic proportions over the next 50 years, increasing the annual output by 10 fold. This might be conservative actually. I think independent music labels will continue to prosper and grow as the means of distribution grow trivially cheap. And world music will become ever more popular as people’s taste changes. In addition, we could imagine a world where MP3 players (or their 2057 equivalents) hold upwards of 1000 (or more) times more songs than currently, so there would be almost no limit to what people could store, if they desired. And if MP3 players get merged with mobile phones, and if mobile phone penetration keeps getting bigger, than perhaps upwards of 4 billion of these devices would be available. Crunching the numbers gives us a figure of 10^5Mc. That is 100,000 times what we currently store as music.
Video is the next growth area for mobile devices. Already, mobile phones can record short videos digitally and YouTube is entering agreements to distribute content to phones. I can imagine that this trend is set to continue relentlessly, with everything from weddings to births to bar mitzvahs being distributed to friends and relations via mobile phones, and every single person on the planet watching and sharing videos at an ever increasing rate. In addition, law enforcement agencies are continuing to advocate the use of CCTV at every street corner. The time is not far off where thousands of these devices will appear in every city in the world, to be stored digitally and searched, as required, for evidence of criminality. All in all, I see the requirement for memory of these digital activities increasing exponentially, to 10^6Vc, although even this increase may be conservative.
The idea that we are becoming a society obsessed with recording our daily lives is gaining currency. YouTube sees every inane, trivial incident recorded and then shared with the world. Hypothetically, we might see a device that takes the pain out of actually having to carry a video recorder with you, or having to anticipate some funny incident that you see. I don’t necessarily see this kind of device becoming too popular, but even if only 1 million people use such a device, then we might expect 10^6Vc from this kind of thing too.
Wouldn’t your doctor love to have a complete record of your every physical move throughout your life, in order to better diagnose, prevent, and treat your every ailment? I can imagine a device capable of being embedded directly into your body that records constantly your heart rate, body temperature, sweating, exercise, blood pressure, glucose levels, cholesterol, brain activity, and so on. The device might produce 1MB of data every second of your existence (that’s a very rough guess) and then would transfer that data to a central database as you pass a wi-fi station. If almost every person living in the first world wore such a device for 365 days a year, then at that rate, the data generated would produce a prodigious need for memory capacity. At a figure of 2 billion people x 60 seconds x 60 minutes x 24 hours x 365 days x 1MB, I work that out to be a prodigious 6.3 x 10^16 MBs. Another way to phrase it is to call it 10 zettabytes, approximately. Too much to comprehend! Incidentally, I have to guess the current health data held about every average human at 10^3 MBs per year (for digitised x-rays, heart monitors, charts, pictures etc), so in my terms that makes approximately 10^13Hc.
Clearly, I could continue to speculate about the memory requirements of civilisation in 50 years (or 100, or 200…) but I hope you can agree that there might be applications out there that will eat up every conceivable bit of memory capacity that we produce.
Faced with this information, it looks like a case of “Build it and they will come”. Look at this classic website (unfortunately it seems to have ended in 2004) for a comparison of memory prices for hard drives across the decades. It never ceases to amaze me how cheaper a MB of disk storage could become. We cannot argue that price will prevent this vision occuring.
I have skipped over many concerns that might arise from these increases: legal, social, financial, logistical, technological, psychological. Perhaps that’s a topic for another day?