Internet Servers May Learn a Thing or Two from Honeybees, According to Researchers

According to researchers at the Georgia Institute of Technology, Internet servers may use the organization of a swarm of honeybees, like how they managed to collect nectar without a central command and limited resources, to improve the efficiency of their servers.

A communications system developed by Georgia Tech was inspired by the bee dance. The communication systems help servers to move between tasks, lessening the chance that a website will be overloaded with requests and potentially shutting down.

Internet servers are the ones running the web sites and there are usually a number of servers devoted to one web site. The servers provide the computing power to access the website and requests to access a website is very hard to predict so servers usually overload when a certain web site is accessed by different people at the same time. When a server overloads, the web site shuts down and couldn't be accessed resulting to a loss in traffic or potential customers for that website.

According to the researchers, honeybees usually have a limited number of workers and they have to adapt to changing environments since nectar may be abundant in one flower and scarce on the other but the honeybees always managed to bring in a steady flow of nectar whether it be rainy season or Spring. How these honeybees are able to keep a steady flow of nectar unto their bee hive is what researchers at the Georgia Tech adapted to their communications system.

Crave Tovey, professor in the H. Milton Stewart School of Industrial and Systems Engineering at Georgia Tech has been studying the efficiency of honeybees for some time and he realized that bees and servers have similar barriers to efficiency. Through conversations with Sunil Nakrani, a colleague from the University of Oxford, Tovey realized that the bees' strategy for allocating their resources can be applied to Internet Servers.

Bees are able to allocate their resources through "dances", a scout bee is usually responsible for searching nectar and once he has found a good source, he returns to the bee hive and performs a dance, telling the other bees the direction of his find. The other bees learn the steps and fly off towards the location. According to Tovey, as long as there is still nectar on the source, the returning bees continue the dance routine or until another forager bee returns with a new dance routine, telling the other bees of a better source of honey.

Tovey and Nakrani's communication system adapts this bee dance; one server acts like the forager and receives the request for a web site and the dance is replaced with an "internal advertisement". The duration of the advertisement will depend on the demand to access the website and the longer the advertisement stays on the "dance floor" the more servers are devoted to serving requests to access the website.

Tovey's research was published in the journal Bioinspiration and Biomimetics.

SOURCE:

Georgia Tech, "Bee Strategy Helps Servers Run More Sweetly." Gatech.edu