Ants invented the internet?

Ask who invented the Internet and you’ll spark off an argument with everyone championed from DARPA to Nikola Tesla. However, two Stanford scientists claim that the inventor may have had six legs, antennae and a taste for disrupting picnics. Professor of biology Deborah Gordon and professor of computer science Balaji Prabhakar say that red harvester ants (Pogonomyrmex barbatus) use the same Transmission Control Protocol (TCP) in foraging that the internet uses to manage data transmissions – making a sort of “Anternet.”

A common image of an ant colony is that it’s the ultimate dictatorship, with each ant condemned to a life of total regimentation. It’s nice metaphor, but it’s actually false. In fact, in an ant colony, no one is in charge. True, there’s a queen, but she doesn’t do anything except lay eggs. As to the rest of the ants, no one makes decisions, no one tells anyone else what to do and no one supervises the work. This raises the question, how does anyone know what to do?

The answer is that they don’t. Instead, like computers, ants use algorithms. They follow a simple set of rules that tell each ant, “If this happens, do this.” For example, every morning harvester ants send out patrols to look for food. If no or few patroller ants return, then the forager ants who collect the food stay in the colony. However, if a large number of patrollers return, then the foragers leave the colony and follow the chemical trail left by the patrollers to the food.

It’s tempting to say that the foragers “know” that there’s food, but that’s not true. The foragers don’t know anything. They are simply and blindly obeying instructions that tell them how to react to what they encounter.

That ants use algorithms has been known for decades and is even used in computer science as a model for optimizing operations. What wasn’t known was just how much like computers ants really are. Gordon, who has been studying ants for 20 years, was observing the times at which harvester ants send out foragers. When she noticed a particular pattern of behavior, she called in Prabhakar, who soon had a revelation.

"The next day it occurred to me, 'Oh wait, this is almost the same as how [internet] protocols discover how much bandwidth is available for transferring a file!'" Prabhakar said. "The algorithm the ants were using to discover how much food there is available is essentially the same as that used in the Transmission Control Protocol [TCP]."

The TCP algorithm controls how the internet handles data. A computer breaks down a file into a set of packets. These are sent to another computer, which sends back a confirmation for each packet received. If these confirmations come back too slowly, then the transmitting computer slows down. If they come back quickly, it speeds up.

Harvester ants work the same way. A forager ant won’t come back to the colony until it finds food. If a lot of foragers find food, they return at a fast rate and more foragers are sent out. However, if few return, then few leave the colony. In both the internet and the ant colony, it all boils down to a question of bandwidth – except in the ants’ case, they’ve been doing it for millions of years.

They even anticipated the slow and fast phases of TCP. The TCP algorithm sends out a large number of packets at the start of a transmission to determine available bandwidth and then adjusts the speed accordingly. Harvester ants do the same thing by sending out a large number of foragers and then adjusting the rate. Also, like TCP, the ants will stop sending out foragers if none return in about 20 minutes.

Prabhakar believes that a study of ant algorithms could lead to simple, scalable distributed networks and other applications. However, it seems unlikely that these will involve replacing server farms with the ant variety.

Source: Stanford University