If you’ve ever tried to save music videos from your browser onto your hard drive, you’ve run into then name Akamai. Usually the final video, once you’ve removed all the frames and html and everything surrounding it, is hosted from an Akamai address with a bunch of numbers as the URL.
It turns out Akamai is a very old compamy, originally founded in 1995.
Like other Net infrastructure plays, Akamai got swept up in dot-com fever. Following its 1999 IPO, the stock price soared from $26 in late October to $345 on New Year’s Eve. But when the Internet bubble burst, many of Akamai’s customers went bust or just disappeared. Then Lewin was killed on September 11 on American Airlines Flight 11, which crashed into the North Tower of the World Trade Center in New York. In 2001, Akamai lost $2.4 billion; a year later, the share price bottomed at just 56¢.
Akamai stock has tripled over the past year to $43.
Lessons to be taken here:
- the value of persistence
- the foolishness of financial markets (i.e. the lemming effect)
- the value over time of good ideas
- the importance of load spreading for speed in web applications (i.e. almost all the big companies including Google, Apple and Microsoft are using Akamai – if there were an easier or cheaper way, they’d be using it)
To underline that last point, the article cites Friendster as a web application which had crumbled under success (bandwidth demands) in spring of 2006. In March after adding Akamai, repsonse time was brought down to 3.4 seconds from 9.4 seconds.
That figure – more than three seconds for response time – is not nearly good enough, but much better. When building a web application, one must be quickly prepared for the unfortunate consequences of success.
As an (unhappy) Typepad customer, I can testify from personal experience to how poorly scaling can be managed. Fortunately I own ecto which means my entries are created offline with comparative security from crashes and timeouts and just posted as a background process.