Communication patterns of bees gave an idea how Internet servers can be optimized to manage massive loads, without being overwhelmed with requests.
Georgia Tech researchers developed a communication system inspired by honeybee dance to help single-task Internet servers move between tasks as needed, reducing the chances of a Web site being overwhelmed with and locking out potential visitors. The team led by Prof. Craig Tovey designed the system by recognizing that bees and Internet servers have one thing in common - limited resources to be deployed for the best results. Researchers studied the bees' strategies for distributing resources in a constantly changing environment to see how the strategies could be applied to Internet servers.
The research focuses on the technique called Swarm Intelligence, a branch of Artificial Intelligence based on collective behavior. So far, the honeybee method improved service between 4 and 25 percent in tests based on real Internet traffic.
Georgia Tech researchers developed a communication system inspired by honeybee dance to help single-task Internet servers move between tasks as needed, reducing the chances of a Web site being overwhelmed with and locking out potential visitors. The team led by Prof. Craig Tovey designed the system by recognizing that bees and Internet servers have one thing in common - limited resources to be deployed for the best results. Researchers studied the bees' strategies for distributing resources in a constantly changing environment to see how the strategies could be applied to Internet servers.
The research focuses on the technique called Swarm Intelligence, a branch of Artificial Intelligence based on collective behavior. So far, the honeybee method improved service between 4 and 25 percent in tests based on real Internet traffic.
No comments:
Post a Comment