Sunday, October 16, 2011

The web in 60 seconds (and its storage requirements)

Have you ever wondered how is the web evolving?

A recent research has shown some facts which have been summarized in this picture:


Let's now concentrate on some of these facts. By this research we see that every 60 seconds:

  • 6600 pictures are uploaded on Flickr
  • 25+ hours of videos are uploaded on YouTube
  • 98000 twits are published
  • 21500+ blog posts are sent (summing up the ones on tumblr and the others)
  • 695000 status updates on Facebook
Fire up your calculator now and check these facts:

  • Consider, for the sake of simplicity, that one picture is 2 Mb on average. Every 60 seconds we use 13,2 Gb of storage for new pictures on Flickr.
  • By the fact that the picture shows that 600+ videos are uploaded every 60 seconds, this means that one video is 2,5 minutes long. Assuming (as a worst case) that every video is uploaded in HD, by simplicistic calculations we can say that one video uses 500 Mb of storage. Therefore, we burn 300 Gb on YouTube's hard drives every minute.
  • One tweet is 140 characters long. This means that (having no meta-data per tweet, which is of course impossible) we have 140 bytes per tweet, which leads to 13,4 Mb of tweets per minute. Of course, I'm not taking into account the possibility of attaching pictures to them.
  • A good post's length should be (by some computations on this) 2500 characters in length. Assume this is the mean value, and that no one is attaching pictures and other content in blogs. Assume blog posts are just plain text, so no formatting, which means no more characters for css/html style. Again, no meta-data is stored along with any blog post. This leads to 2,5 Kb per post and 53,8 Mb per minute.
  • A Facebook status update is bounded by 420 keystrokes. Taking this as an upper bound, under the same assumptions as before, we have 420 bytes per update and 278,4 Mb per minute.
This means that considering only these social activities, every minute we use 313545,6 Mb of storage around the world.

Nowadays we have 1 TB hard drives widely available. If social sites were only using this kind of storage, in 3,34 minutes one disk would be filled up by our social bullshit. (And yes, with this post I'm contributing!)

This means that as we have 1440 minutes in one day, we trash 431,14 1TB hard drives per day. Which is 157365,27 per year.

My computations are very simplistic. But maybe this is the sole reason why Cloud Computing is taking its step. And we all should give up at trying to develop the new best algorithm for optimizing computation. As we maybe only need storage...

No comments:

Post a Comment