A note on the project focus

It should be noted that the main focus of this project is functionality, not efficiency. While NIST Switch should be efficient enough to allow meaningful use on stock PCs over 10Mb and 100Mb Ethernet, and eventually over ATM, there is no need for million packet/second processing rates. As a rough guideline, a better than 10,000 packet/second performance rate (100 $\mu$sec/packet) should be readily achievable on current (as of 4/98) machines in the $1000-$2000 range, assuming the total per-packet overhead for NIST Switch components stays under 40 $\mu$sec. While this level of performance would not be adequate for comparably-priced commercial products, it should be more than sufficient for use in a research setting, even as a basis for many comparative performance studies.