Nokia Xpress

~ the eulogy ~

Nokia Xpress was a browser that did not really browse. When you asked it for a page, the request went to a Nokia data center, which fetched the site, rendered it, threw away most of what was there, and sent your handset a compressed package that was up to ninety percent smaller than the original. The phone was not talking to the web. It was talking to Nokia, and Nokia was talking to the web on its behalf.

That sounds like a compromise until you remember what it was running on. Series 40 was not a smartphone platform. It was the software on the cheap Nokia bar phones that sold in enormous numbers across Asia, Africa, the Middle East and Eastern Europe, on hardware with a fraction of the memory a real browser needs and on data plans priced by the megabyte. A normal browser on that phone would either fail to render the page or cost the owner more than they wanted to spend finding out. Xpress made the mobile web technically possible and financially survivable on a phone that cost thirty dollars.

It arrived properly in 2012, in beta on Lumia in march and as the default across the Asha line that october, though the same compression engine had been shipping under the plain Nokia Browser name before the rebrand. Nokia put real engineering into it, including a web-apps layer so that developers could target the cheap handsets without writing native code. For a couple of years it was, by installed base, one of the most used browsers on earth, and almost nobody writing about browsers noticed.

By 2012 I was an engineer in Ankara with a phone that did its own rendering, so this was never my browser. It was the browser for the tier below mine, and that tier was much larger than mine. Turkey in those years still sold mobile data as though it were rationed, and the Asha phones were everywhere in the hands of people who were not going to pay smartphone prices for the privilege of loading a page twice.

~ the dossier ~

~ the unraveling ~

In january 2013 a security researcher noticed something that should have been obvious from the architecture. If Nokia's servers were compressing your pages, they had to be able to read your pages, including the ones served over HTTPS. Xpress was terminating the encrypted connection at Nokia's end, decrypting the traffic, compressing it, and re-encrypting it on the way down. Your bank session was in plaintext on a machine in Finland for as long as it took to squeeze.

Nokia's answer was that it did not store anything and that no employee could access the data, which was probably true and entirely beside the point. The company had built a legitimate interception layer between every Xpress user and every secure site they visited, and had shipped it as a default on hundreds of millions of phones without saying so plainly. Opera Mini worked the same way and took the same criticism. It was the price of the whole proxy-browser category, and the category simply hoped nobody would do the arithmetic.

~ how it actually ended ~

Microsoft bought Nokia's devices business in 2014 and immediately had a portfolio it did not want. The feature phone line was profitable but strategically pointless for a company betting on Windows Phone, and maintaining a global compression proxy for it was an ongoing server bill with no upside. On 26 january 2015 Microsoft announced that Opera Mini would become the default browser on Series 30+, Series 40 and Asha handsets. Opera would run the proxy, Opera would pay for the servers, and Xpress would be retired.

Support formally continued to the end of 2015 and then stopped. The compression servers went away, which is the detail that matters, because a proxy browser without its proxy is not a degraded browser but a dead one. Millions of phones kept the icon and lost the function. Microsoft sold the feature phone business to HMD and Foxconn in 2016 and left the whole category behind.

~ what it left ~

The compression proxy is one of the few genuinely good ideas the mobile web produced, and it kept getting rebuilt by people who had never heard of Xpress. Chrome shipped Data Saver on the same principle and killed it in 2022. Opera Mini is still going. The concept only makes sense when bandwidth costs more than compute, which was the condition in most of the world for most of the smartphone era and is quietly returning wherever data is sold in small expensive increments.

What Xpress really documents is a period when the phone in most hands was not a small computer but a terminal for one somewhere else. That arrangement was supposed to be a stopgap until hardware caught up. It turned out to be the shape of everything that came after, with the rendering moved back onto the device and almost everything else moved off it.

@}-->--

              )
             (_)
            .-'-.
            |   |
            |   |
            |   |
            |   |
          __|   |__   .-.
       .-'  |   |  `-:   :
      :     `---'     :-'
       `-._       _.-'
           '""""""

« previous  ·  random  ·  next »