Thanks to swarm intelligence, bees’ colonies can organize foraging activities very efficiently. Though there is no central control, forager bees can spread out in the area surrounding the hive so to produce as much honey as possible in the given conditions.
Whenever a forager bee finds a new source of food, it can easily tell how valuable it will be for its sisters. The forager bee will gather some nectar from the new source and take it back to the hive. Then it waits for bees inside the hive to collect it and store it away. If this happens slowly, that means there is a lot of food in the hive. Consequently, the source of food found might not be regarded as valuable by its sisters. If the process happens fast, then it means that sources of food are scarce. In this case the forager bee who found the new source of food will start its waggle dance to point the others in the right direction.
In the 90s, a team of researchers from Georgia Institute of Technology came up with a theory that predicts forager bees’ distribution in relation to where the sources of food are located. While developing this model, researchers noted how bees do not spread in an optimal way – at least, not in the specific meaning of the word in system engineering – so they thought bees’ evolution did not reach its full potential.
Fortunately, their research did not stop here. They cast aside theoretical formulations for a while and focused on field work. Researchers marked 4000 forager bees individually through a combination of colors and numbers. They placed the hive in an area with variable food sources and observed the bees’ foraging activities for a week. It turned out that bees spread in the most efficient way. They knew how to adapt to an environment that goes through continuous changes such as the blooming of flowers at different times in the year.
It took quite some time to find the right application for this model. Eventually, professor Sunil Nakrani was the one to find it. He was looking for a solution to the very variable demand of servers providing web hosting services. He talked about this problem with professor Craig Tovey. Little did he know that professor Tovey was in the team mentioned above, so he had a solution on hand.
According to the demand for each application, server providers of web hosting services give a certain space to their clients. Each server can run only one application at a time. It takes some preparation time before a server can switch to a new application, and whenever that happens, there will be a money loss. The market of server providers is very variable, so it’s only natural that they will look for the most efficient way to allocate their money.
The study of foraging activities in bees led researchers to develop an algorithm capable of assessing a client’s economic potential, so that this information can be shared with the other servers – just like forager bees pointing their sisters in the direction of a food source with their waggle dance.