Q & A on 100Mbps Network Performance Q: My 100Mbps network has a lower efficiency than 10Mbps network. I wonder why it is the case: I think the overhead is mainly due to the header, and the Proportion of the header should be smaller when a larger package is sent. Therefore, 100Mbps network should have a higher efficiency than 10Mbps network. However, the result is different from my expectation. Where did I make mistakes? Thanks. A: The header and maximum data sizes for a 100Mbps Ethernet are exactly the same as for a 10Mbps Ethernet. Thus, if the measurement of throughput is made by sending the same data on each network, none of the difference will be caused by the headers. Here are a couple of possible explanations. First, if the two networks are not both isolated, traffic from other stations can interfere with measure performance, causing the observed throughput to appear lower than it should. Second, the system being used for measurement may not be capable of transferring data at the higher speed. Any component may be a bottleneck: the network interface card, interrupt processing, processing by protocol software, lack of adequate buffers, or operating system overhead such as context switching or I/O.