| 摘要 |
The present invention provides a system and method for compressing a data sequence comprising a plurality of records, each record having a plurality of fields and each field being arranged to contain a data item. The system comprises: comparison means which, for a current field within a current record other than the first record in the data sequence, is arranged to compare the data item in the current field with the data item in the corresponding field of a preceding record; and match indication means, responsive to a signal from the comparison means indicating that the data item matches the data item in the corresponding field, for replacing the current field data item by a token indicating the match. The comparison means is arranged to repetitively perform the comparison process on a predetermined number of fields in a plurality of records of the data sequence. Preferably the system is incorporated within a server computer, the server being arranged to output the data records of the data sequence as compressed by the system for transfer over a network to a client computer. The technique of the present invention is a very quick algorithm, taking very little resource. It does not prevent the use of more traditional compression techniques, and is simple. Further, it recognises the field structure of the data and uses this as a method to achieve good compression. Its algorithm is not affected by the host computer architecture nor that of the transport layers. Additionally, it can reduce the CPU resources required at the client side, thereby improving performance above and beyond the data reductions. <IMAGE> |
