SlipStream Data Inc.

CA

Comparison of SlipStream Data's Compression and Other Approaches

SlipStream Data Inc., has developed products which contain proprietary compression, and network optimization technologies. The compression technology used in the SlipStream products is unique and patented.

Compression Technology

Background - Data compression methods can be divided into two types: lossy data compression and lossless data compression. Lossy methods are popular for compressing rich-media, such as video, audio or pictures. Lossy compression is achieved by discarding information without severely affecting the user's perception of the subject. A trade-off is created between quality and compression. In the case of text data, lossy compression is not appropriate. Every byte must be recovered to maintain the integrity of the data.

SlipStream Data uses lossless compression for text data including plain-text, HTML, Javascript and Content Style Sheets (CCS), and lossy compression for images. The core technologies behind its acceleration solutions are its innovative, on-line, and tuneable lossless compression algorithms which can achieve as much as 20: 1 on-line compression for web text data objects, resulting in much faster loading of web pages.

Conventional lossless data compression methods typically employ Huffman coding, run-length coding, arithmetic coding algorithms, Lempel-Ziv algorithms, and their variants.

Huffman Coding and Run-Length Coding - Huffman coding and run- length coding were invented in 1957 and 1966, respectively. Because of their simplicity, they are still widely used in the lossless component of a lossy compression system such as MPEG and DV.

Arithmetic Coding - Arithmetic coding algorithms and their variants are statistical model-based algorithms. To use an arithmetic coding algorithm to encode a data sequence, a statistical model is either built dynamically during the encoding process, or assumed to exist in advance. Several approaches have been proposed to dynamically build a statistical model. These include the prediction by partial match algorithm, dynamic Markov modelling, context gathering algorithm, and context-tree weighting method

Lempel-Ziv Algorithms - In contrast, no statistical model is used in Lempel-Ziv algorithms and their variants. During the encoding process, the original data sequence is parsed into non- overlapping, variable-length phrases according to a string matching mechanism, and then encoded phrase by phrase. Each parsed phrase is either distinct or replicated with the number of repetitions less than or equal to the size of the source alphabet. Phrases are encoded in terms of their positions in a dictionary or database.

Other conventional universal compression methods include the dynamic Huffman algorithm, the move-to-front coding scheme, and some multi-pass compression algorithms including algorithms based on the Burrows Wheeler Transform. These conventional methods are either inferior to arithmetic coding and Lempel-Ziv algorithms, or not applicable to on-line compression for transmission.

SlipStream Data's Compression Technology

SlipStream Data's compression solutions use a relatively new source coding theory to construct on-line, tuneable, context- dependent compression algorithms. Instead of compressing the original objects, SlipStream Data's solutions compress the numerical set of rules. To achieve on-line compression, the solutions have an engine to construct these rules dynamically and then encode these rules on the fly as web data flow in. The end results are efficient compression, fast sequential encoding and decoding, and small memory usage - all these are important and desirable factors in on-line compression for transmission.

Since the set of rules is dynamic, when the structure of web text data objects is changed, the corresponding rules will be updated; similarly, when the content of web objects is updated, the corresponding rules will be updated accordingly. This unique feature fits very well with the world of web pages, where the content of web objects is updated constantly and on the other hand, the structure of web objects remains approximately constant and changes infrequently. This unique feature of its solutions enables SlipStream Data to achieve astonishing on-line, tuneable compression.

Network Optimizations

Most of the Internet applications rely on TCP/IP protocol for communication of data over the network. The TCP/IP is designed to provide robust and reliable communication over different types of networks and conditions. However, as a trade-off, TCP/IP provides poor performance with noisy, low bandwidth, and high latency networks such as wireless and dial-up networks. Moreover, popular Internet applications like email and web- browsers make and break connections several times during a single session further exacerbating the performance problem. SlipStream Data uses a proprietary network architecture and protocol that eliminates the problem of poor performance of TCP/IP caused by slow-start up and frequent handshakes.

United Online IKANO Communications TELUS Bell Canada AOL Canada America Online France Telecom Terra Spain Northwestel Onspeed (UK) Service Providers in more than 40 countries worldwide

Millions of users worldwide are benefiting from our technology for increased acceleration on dialup, wireless and broadband connections.

541510