TCPIP协议族，第3版（影印版）下载 附录C.pdfVIP

APPENDIX C Checksum This appendix briefly describes three common checksum techniques: traditional, Fletcher, and Adler. C.1 TRADITIONAL Traditionally, the checksum calculation for IP, UDP, and TCP uses one’s complement arithmetic. We use three number systems, binary, hexadecimal, and decimal. The first two are done manually to show the concept; we use a flowchart for the third. Calculation in Binary To show a binary checksum calculation, we use the data in Figure 11.9 in Chapter 11. First we calculate the partial sum as shown in Figure C.1. We add each column and carry to the next columns, if necessary. We then calculate the sum. Finally, we get the checksum, which is the one’s complement of the sum (each 0 is changed to 1 and each 1 to 0). We step through the calculations: ❏ When we add the first (rightmost) column, we get 7. The number 7 in binary is 111. We keep the rightmost 1 and carry the rest to columns 2 and 3. ❏ When we add the second column, we include the carry from the first column. The result is 8, which is 1000 in binary. We keep the first bit (rightmost) and carry the rest (100) to columns 3, 4, and 5. ❏ We repeat the above procedure for each column. ❏ When we finish adding the last column, we have two 1s for which there are no col- umns left for addition. We add these two 1s to the partial sum in the next step. ❏ If there is no carry from the last or previous columns, the partial sum is the sum. In this example, there are carries; these are added to the parti