With the advent of IP over Avian Carrier (see RFC 1149) and the recent addition of Quality of Service in RFC 2549, it becomes interesting to try to extend these protocols to other animals.

To date, the Habitrail appears to be the best medium with gerbils in the ether. While this has not yet been tested or proven to be an effective means of computer communication, it is none the less interesting.

The habitrail is a clear yellow tube that allows gerbils to wander about within the confines of the tube. This in may ways resembles network topology in some environments. Furthermore, gerbils and habitrails are both available in many Home Area Networks.


The basic essence of gerbil-net has several different protocols.

  1. Tie a message into the leg of the gerbil. While this allows for more information per packet, it does have the disadvantage that it may be necessary for a human or complex Rube Goldberg machine to read the information.
  2. Use a black gerbil to represent a '1' and a brown or tan gerbil to represent a '0'. While data rates are slow, this may be the best method and most research has been around this protocol.
  3. Use a one-time gerbil and tattoo the packet in a machine readable form on the gerbil's back.

To correctly implement gerbil-net, it is necessary to take into consideration some of the facts about gerbils.

  • Male gerbils are very territorial. Attempting to implement a network that contains more than two gerbils will result in catastrophic packet collisions. This is often thought to be a Bad Thing and should be avoided.
  • Male gerbils when left with female gerbils do exactly what nature intended them to do - they reproduce (yes, that is right, these packets run in promiscuous mode). While this may not have noticeable immediate consequences, it will lead to packet duplication at some point in the future which will doubtless confuse gerbil-net protocol II.
  • Female gerbils will eat their young. If using gerbil-net protocol II, and trying to use the above feature to get redundancy in the network this may cause packet loss.
  • Gerbils are living creatures and die. A dead gerbil in the tubes is not something that people want to see and will clog communication between the two points requiring alternate paths to be taken. To clear out the gerbil it is recommended that the network be shut down for a time while a parity ferret is allowed to clean up the lost packets.
  • Gerbils poop. Packet droppings must be cleaned out on a regular basis. Packet sniffing is inadvisable.


After much thought, the 'best' gerbil-net topology is that of a token ring where only one host is releasing gerbils to the network at any given time. The receiving host will then store the gerbils until such a time that host wishes to send packets. If using gerbil-net protocol II, the 'token' is represented as an albino gerbil (as opposed to the brown and black gerbils). If the store of gerbils is used up, this will be the last packet sent out and allows the host to possibly receive some data before it is time to send again.


With wireless networks being all the rave today, gerbil-net has been extended to this medium with limited success. The easiest and most obvious is to use the habitball. The difficulty with this is that the gerbil tends to get lost and rude teenagers try to go bowling with it.

Still, the habitball is more successful than gerbil-net catapult where gerbils are launched at the desired destination in a catapult. The prime difficulties with this method of wireless gerbil-net communication is that packet order is very difficult to distinguish and unless there is a football stadium sized bowl of jello the gerbils tend to arrive in a sloppy red color which is very difficult to decode using any protocol.


Security can be implemented on the classic habitrail base G network by changing the tubes from the clear yellow to an opaque style (such as PVC tubes). This makes it much less interesting for the packets and takes some coercing to get them into the network.


Repeaters for gerbil-net are being postponed until cloning experiments are successful, even then it may take several months for the packet to be re-released.

The pneumatic gerbil-net accelerator necessary for connections with normal high speed networks has been put on hold until it has solved the puree problem.