The Browser Will Become Everything - Let's Do The Maths


If you don't think everything can - no will - move to the cloud, this again.

In a previous posts Amdahl's Law Meets Moore's Law - Death Of The PC and Commodity Markets In Cloud Computing - It Has Started! I have shown how the end of the PC as a dominant platform is coming to and end and how cloud computing will allow all our compute power needs to be provided as a commodity. However, what about really intensive visual stuff like drawing and gaming?
The PC will die here too and all this will move into the cloud.
By the end of the upcoming decade the browser will be everything - the platform will be the browser.
How do I come to this radical assertion? Let us consider some simple mathematics (I really don't understand why this is not being said more often and more loudly):
Modern HDMI (the high definition media interface standard) runs at 10 GB/s. In the UK fast broad band is now 50MB/s and 15 years ago when I was doing my PhD it was 15KB/s (I remember because I worked from home over a 14400 modem - I have rounded up). That gives a exponential rate of change of the order of 1.71 per year:
Year Internet Speed
1994 15000
1995 25650
1996 43861.5
1997 75003.165
1998 128255.4122
1999 219316.7548
2000 375031.6507
2001 641304.1226
2002 1096630.05
2003 1875237.385
2004 3206655.928
2005 5483381.638
2006 9376582.6
2007 16033956.25
2008 27418065.18
2009 46884891.46
2010 80173164.4
2011 137096111.1
2012 234434350
2013 400882738.5
2014 685509482.9
2015 1172221216
2016 2004498279
2017 3427692057
2018 5861353417
2019 10022914343
2020 17139183527
As we can see from the numbers and the graph, my model estimates being able to pipe raw HDMI over the UK's sluggish broadband sometime in 2018/2019. The reality is that the new multi-core processors that are now comping on line will be able to handle compression and decompression so easily that the date at which local video sources are obsolete will be much sooner than that; I suspect the tipping point will be in the latter half of the decade, but exactly where is impossible to predict realistically.

Latency

There an issue of latency. However, this is becoming ever less so. We are learning tricks so that people do not notice latency; these are largely coming out of the gaming industry where latency issues in multi-player games have been overcome even over contemporary slow networks. Once the path from the cloud to the browser is over glass and clouds are on the same continent as the browser, latency goes away as an issue. It is worth noting that humans are not able to perceive latency less than 10 milli-seconds; the real figure is more like 100 milli-seconds for most situations.

What does this mean for the industry?

  • People will rent software, not buy. This will be for games as well as video and photo processors/editors.
  • Personal compute devices will have much longer life spans.
  • Issues like aesthetics will become more important.
  • As I suggested before, the local operating system will stop being a relevant differentiator.

What does this mean for the user?

  • Very little - life will just get easier.
  • Computing will get more stable (installations/upgrades etc).
  • The playing field will be levelled. If we want to just use Photoshop for a small job we will be able to rent a few hours rather than stump up missive costs for a full installation.
  • Games will merge with other media all emanating from a common cloud source. People will move straight from watching the film into playing the game into online chat onto editing a fan-fiction 3D animated romance scene based on the plot.