Methods for compressing and decompressing data packets. A data packet is classified as related to an Internet service and a function is determined that the classified data packet is intended to accomplish for the Internet service. A state that corresponds to the classified data packet is determined based on a state-machine and one or more states of the state-machine. The state-machine and the one or more states are associated with the determined function in a conversion file. Compression is achieved by detecting in the classified data packet a data pattern associated with the determined state in the conversion file and replacing the data pattern with a token associated with the detected data pattern in the conversion file. Decompression is achieved by detecting a token associated with the determined state in the conversion file and replacing the token with a data pattern associated with the token in the conversion file.