Developers along desire to evaluate their new technologies in realistic situations. The developers for the Internet are not excepted. The general experimental issues for Internet technologies are efficiency and scalability. StarBED enables to evaluate such factors in realistic situations.
Actual computers and network equipments are required if we want to evaluate software for the real Internet. In StarBED there are many actual computers, and switches which connect these computers. We reproduce close to reality situations with actual equipments that are used on Internet. If developers want to evaluate their real implementation, they have to use actual equipments.
There are 680 actual PCs in StarBED, in order to realize a large-scale topology. Furthermore, each node on StarBED can run 10 virtual machines with VMware, which enables constructing a 6800-node topology. The experiment topology is constructed with VLAN, and it can adopt various topologies.
It is difficult to control experimenter's programs on many computers by human operations. The experimenter has to logon the experiment nodes and type commands the scheduled timings. Moreover, human operation is not so accurate, and it causes a decrease in the reliability of the experiment. We develop an experiment support software, SpringOS, which supports experimenter's operations. SpringOS consist of a dozen modules, each having a specific purpose.
With SpringOS, an experimenter's program (target) is executed by an experiment driving program (driver.) The driver uses an experiment description language. The experimenter can write complex scenarios (e.g., state machine, rendezvous and timeout) using the language. The driver name is Kuroyuri, which is a part of SpringOS.
StarBED is only a nick-name. The official name of the facility is Hokuriku Research Center. This facility is founded and managed by NICT.
The design of StarBED and SpringOS is our role. the StarBED team. Please contact us, if you have any questions.