Computer programmers are taking the honey bee as a model

The increasing rise in the level of shopping over the Internet is bringing a number of major problems in its wake. Customer shopping behavior may be entirely different to that which is generally expected, and may differ among customers. The variable traffic resulting from this leads to sudden loading for on-line sales Internet servers. (Server: a computer in a network that stores application programs and data files accessed by the other computers in the network.) Experts from Oxford University and the Georgia Institute of Technology are working together in order to develop technologies to manage such loading. The researchers are taking a community whose traffic has already been effectively regulated as a role model. It is the behavior of honey bee colonies that is being imitated in technologies aimed at easing the burden on servers at times of exceptionally heavy traffic.

The sudden surges of shoppers or stockmarket trading, the ups and downs of an internet auction represent a major difficulty for firms running servers. In order to raise their profits to the maximum level, they need to monitor their computers at every moment to keep them capable of adapting to changing levels of demand by means of rapid interventions. However, the fact that only a single web application can be loaded onto a computer at one time constitutes a handicap. Switching between applications causes an interruption during the 5-7 minutes necessary for the computer to be reconfigured, in other words, it results in losses.

A similar problem arises in terms of the tasks carried out by honey bees. Flower patches vary in terms of quality. Therefore, one might think that decisions such as how many bees should be sent to each patch and how long they should spend there would represent a problem in a colony that wishes to maximize the nectar collection rate. Yet thanks to their highly effective work system, bees are able to resolve this with no difficulties being experienced.

Around one-fifth of the bees in a hive serve as nectar collectors. Their duty is to rove between flowers and collect as much nectar as possible. When they return to the hive, they hand over their loads to food-storer bees that maintain the hive and store foodstuffs. These bees then store the nectar in the honey combs. A nectar-collecting bee is also assisted by its fellows in judging how good its flower patch is. The bee waits to see how long it takes to find an unoccupied food-storer bee. If this waiting period lasts a long time then the nectar-collecting bee takes this as an indication that its flower patch is not of the very top quality, and that other bees have for the most part carried out successful searches. On the other hand, if it is met by a large number of food-storer bees to take over its load then the probability of that load being of high quality increases.

The bee that makes use of this information decides whether its flower patch is worth the expenditure of any more effort. If it is, then it performs its famous waggle dance so that the other bees will follow it. The length of this dance shows how profitable this flower patch is likely to be.

Sunil Nakrani of Oxford University and Craig Tovey of the Georgia Institute of Technology adapted the honey bees' strategy to the problems of Internet hosts. Each server assumed the role of the nectar-collecting bee, and each customer request that of the flower patch. In this way, doctors Nakrani and Tovey developed a "honey bee" algorithm for Internet server "hives." (Algorithm: a logical sequence of steps for solving a problem that can be translated into a computer program.)

A host performs the task the bees carry out by waggle-dancing by producing an advert and sending it to other servers in the hive. The duration of this advert reflects the importance and profitability of that server's customers. Other servers read the advert and behave like worker bees following the instructions in the waggle dance. Considering the advert and their own experiences for analysis, they decide whether or not to make a switch from the customers they are serving at the moment to customers whom the server sending the advert is serving.

Doctors Nakrani and Tovey tested the honey bee algorithm they developed against that known as the "greedy" algorithm currently used by most Internet host providers. A greedy algorithm is backward looking. It divides time into fixed periods and allocates servers to customers for a period by working out what would have been the most profitable arrangement in the preceding period. Researchers revealed that at moments when the traffic was highly variable, the honeybee algorithm exhibited a 20% better performance than that of the greedy algorithm. It may be that servers working with the honeybee algorithm will spread in the near future, at which time it will be more appropriate to refer to the Internet as an "Intercolony."

This study, carried out by scientists, shows, with a most appropriate analogy, what rational solutions there are in nature. The problem facing the Internet servers is very similar to that solved by the honey bee colony. Indeed, the success of the study, performed by applying the honey bee colony model, is an indication of this. But where does the solution offered by honey bees to computer programmers stem from? Even though computer programmers can take honey bee behavior as their models, the bees themselves have no such opportunity. That is because although the computer programmers' imitation of bees is the product of a judgment-based process of conscious thought, bees have no such ability to think. Solving the problem requires intervention, such as first of all to be aware of the problem, to analyze the factors giving rise to the problem, revealing their effect on the problem in general and on one another, and finally to decide from among various options.

Of course such a problem solving process cannot have taken place in a bee colony consisting of 20 to 50 thousand members. There can only be one rational explanation for the way in which so many living things save energy by implementing the most efficient nectar collecting strategies, although one would normally expect to see chaos and confusion. The knowledge of the problem in the bee colony and its solution are the work of an omniscient Creator. There can be no doubt that it is God, Creator of the heavens and Earth and all that lies between, Who created the honey bee colony. The strategy applied in the colony is the product of God's inspiration. God reveals this in the following verse:

Your Lord revealed to the bees: "Build dwellings in the mountains and the trees, and also in the structures which men erect. Then eat from every kind of fruit and travel the paths of your Lord, which have been made easy for you to follow." From inside them comes a drink of varying colors, containing healing for mankind. There is certainly a sign in that for people who reflect. (Qur'an, 16:68-69)

1- "Honey bees and internet optimisation," The Economist, April 15 2004.