W3C home > Mailing lists > Public > www-tag@w3.org > September 2011

Re: Amazon Silk

From: Karl Dubost <karld@opera.com>
Date: Thu, 29 Sep 2011 14:53:25 -0400
Message-Id: <7919D04F-8386-4E82-B87A-57449AD9EB42@opera.com>
Cc: Mark Nottingham <mnot@mnot.net>, <www-tag@w3.org>
To: Robin Berjon <robin@berjon.com>

Le 29 sept. 2011 à 06:30, Robin Berjon a écrit :
> Is it really devaluing the Web or is it addressing a technical need that we don't cover?

I do not speak for Opera.
What is happening have interesting consequences, that we need to fully understand. This morning I was tweeting away about the notion of BigCo/cloud networks. The Web has a nice feature which is to be distributed without a single owner. It creates plenty of issues, it is imperfect in terms of speed, it breaks more or less gracefully but the distributed nature of it (aka the URI) is one of the main features, if not the main feature.

Now let's push it a bit. Each of the participating companies in what I call "Cloud Networks" try to solve a set of issues for their users base. 

* Opera Mini/Opera Turbo [1] gives a better experience in 
  terms of bandwidth and gives access to user on feature 
  phones with low capabilities.  

* Google with PageSpeed [2] intends to remanufactured the 
  pages on the fly so the developers do not have to do 
  it themselves with the possibility to change your 
  CNAME [3] to have all traffic going through Google servers.

* Amazon Silk is another case of balancing the optimization 
  of Web pages on different server/client. [4]

Now what is happening when a big part of the Web traffic goes through a unique vendor? How these services will scale and be monetized. Imagine for a second that we enter into a competition for these networks (exactly like airlines companies trying to maximize their consumers base). We could imagine that at a point 

* these networks decide to pay the owners of these websites knowing exactly which one receive which traffic.
* the owners optimize their sites for the capabilities of this specific network (more gain, more users). We see that type of behavior for browsers already.
* the network to be subjected to the law of a specific country where the headquarter is.
* the network to abide to some powerful organizations with very specific policies (religious, ethics, etc.)

This is not an encouraging set of thoughts. Are there things that we need to fix in the Web architecture (protocols and languages)? Speed is the marketing argument to gain users, but we have to be careful that it doesn't answer the monetization part of it. There is also the strong asymmetry in download/upload internet access which forbids tons of user engagement in businesses. It pushes the creation of big services able to provide the bandwidth that individual users can't have and reduce the distributed nature of the Web. 

Many questions.

[1]: http://www.opera.com/mobile/specs/
[2]: http://code.google.com/intl/fr/speed/page-speed/
[3]: http://code.google.com/intl/fr/speed/pss/docs/setup.html
[4]: http://amazonsilk.wordpress.com/

Karl Dubost - http://dev.opera.com/
Developer Relations & Tools, Opera Software
Received on Thursday, 29 September 2011 18:53:59 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:56:40 UTC