Error masking probability of 1's complement checksums

In the transport layer of the TCP/IP protocol suite, both TCP and UDP use Internet checksum to protect headers and data. Internet checksum uses 1's complement arithmetic to detect errors in the content delivered by the data-link layer. Both research and experience have shown that there are a wide variety of error sources which cannot be detected by this lower layer. The error detecting performance of 1's complement checksum determines how many of these errors will be passed to higher layers, including the application layer. The performance analysis also influences protocol design and improvement, for example, header compression. Unfortunately, previous work on this topic only determined the number of error passing patterns and the probability for 2 and 3 bit errors, and the method used for determining the probability is hard to extend to more bit errors. We present a method to generate the formula of error passing probability. When too much calculation is needed to compute an exact result, we achieve a better estimation of the probability, which is around 3 percent of the upper bound achievable with previous techniques when 1's complement checksum is used in TCP/UDP.