This Week's Reviews! NETWEBLY HOME! SUBMIT YOUR WEBSITE FOR THE FIRST ANNUAL NETWEBLY AWARDS! the netwebly guide to the Internet - The Intelligent Web Guide><a name=
... Search the Guide
this week's
the netwebly report
about us
rating system

Trouble in Electric Ladyland

The best just got better

Meet The Weasels
The web site designers your mom warned you about

Free High Speed Access Arrives
What to Expect

Wanna e-trade?
No thanks

Software we'd Like to See:
From the people who brought you Autonuke: software to warm your heart. Introducing Bidnessbot

Taking your Business online? Look before you link
The web site that launched a thousand (annoying) e-mails.

Do unto Microsoft
Our top ten alternative punishments for Microsoft...

The Motley Fool
Cool fool

You're thinking something. Tell us what it is.




Solving the bandwidth problem...
Akamai Technologies
It wasn't that long ago that experts were wondering if the Net would able to survive a dramatic growth rate that was unlike anything we'd ever seen...

No one was questioning the desirability of the Information-based society, or the commercial implications of electronic commerce.

The problem was with the Net's infrastructure itself.

With millions of Internet users flooding data pipelines to major web sites with simultaneous requests, traffic slowed to a crawl.

It became almost impossible to log on to many popular sites at peak times. Simple requests for hypertext files took forever. Downloading a large file meant abandoning your computer and spending a couple of hours doing something else to kill time during the agonizingly slow process of the file's transfer across the clogged networks of the data highway.

Newcomers could hardly be blamed for complaining about the World Wide Wait.

It was becoming increasingly obvious that something had to be done, and that the person or company that came up with a solution would become a savior - and a fantastically wealthy savior at that.

A mathematician working at the Massachussets Institute of Technology (MIT) guessed that he might have a solution.

Acting on a casual suggestion from World Wide Web inventor Tim Berners-Lee, Tom Leighton posed the problem to a group of students.

Could the network load-balancing problems that were reducing Net traffic to a crawl be solved with the use of advanced mathematical alogorithms?

Leighton believed the right alogorithms would be able to predict network troublespots ahead of time, giving network administrators the ability to divert traffic around potential hotspots.

Leighton and his students went to work. After three years of intense research, Akamai technologies was born...

The Akamai approach to handling Internet traffic relies on a global network of servers, operating outside of the traditional Internet. Rather than allowing traffic to be funneled into a concentrated area, causing the infamous bottleneck problem, the system relies on regionally located servers to help solve the problem.

A user on the east coast, might be directed to duplicated content on a server in Massachussetts, rather than to a company's central server in Colorado.

The algorithims which Leighton and his students worked so hard to perfect function a little bit like a digital version of the human nervous system, constantly directing and redirecting traffic to the various servers in the Akamai network - exactly the solution Berners-Lee had hypothesized might work.

To operate on a global scale, the Akamai system required the consent of neighborhood ISP's to establish the necessary network of local servers.

The company has been successful in recruiting these service providers. By early 2000, Akamai had a network of thousands of servers with more being added almost every day.

So far the Akamai approach has proven itself to be very effective - large high traffic web sites have been quick to sign up with the program and the system has shown itself capable of handling significant traffic.

The reward?

Leighton and the students who started out working on a math project are now running one of the most successful companies in the Internet Industry, and will play an important role in the future of the medium.

Not a bad payoff for solving a math problem.

back to top

Add URL | This Week| netwebly guide | netwebly report

© the netwebly guide 2000