Table of Contents 

dynamically builds a page to conform to the look and feel and navigation features of each Web site. Our wireless data services are device 

independent and provide a platform that enables our wireless carriers to support HDML, SMTP, SMPP, and a variety of emerging protocols 

such as VoiceXML and PQAs for Palm VII. Our services are compatible with a variety of gateway technologies including WAP gateways from 

Nokia, OpenWave and Ericsson.  

Our Web Server Technology includes other features designed to optimize the performance of our information infrastructure services, 

including:  

a HTML compressor that leverages semantics preserving modifications of file content to reduce size, thereby reducing download time 

for users; 

an adaptive Keep Alive feature that maximizes the duration of HTTP connections, based on current server load, thereby improving the 

user experience by reducing latency; and 

a proxy server that provides the capability for real time integration and branding of applications that reside remotely with third party 

providers. 

Database Technology 

We have developed database technology to address the specific requirements of our business strategy and information infrastructure 

services. We designed our Co operative Database Architecture to function efficiently within the unique operating parameters of the Internet, as 

opposed to commonly used database systems that were developed prior to the widespread acceptance of the Internet. The architecture is 

integrated with our Web Server Technology and incorporates the following features:  

Our Heterogeneous Database Clustering allows disparate data sources to be combined and accessed through a single uniform 

programmatic interface, regardless of data structure. These clusters facilitate database bridging, allowing a single database query to 

produce a single result set containing data extracted from multiple databases, which is a vital component of our ability to aggregate 

applications from multiple sources. Database clustering in this manner reduces dependence on single data sources, facilitates easy data 

updates and reduces integration efforts. In addition, our pre search and post search processing capabilities enable users to modify 

search parameters in real time before and after querying a database. 

Our Dynamic Parallel Index Traversal mechanism utilizes the search parameters supplied by the user to determine the appropriate 

database index (from among multiple indices) to efficiently locate the data requested. Further, an index compression mechanism 

allows us to achieve an efficient balance between disk space and compression/decompression when storing or accessing data. 

In a response to a database query, conventional databases access previously displayed results in order to display successive results to a 

given query, thus increasing response time by performing redundant operations. Our Automatic Query State Recovery mechanism 

decreases response time by maintaining the state of a query to allow the prompt access of successive results. This feature is 

particularly important, for example, when an end user query retrieves a large number of results. 

We incorporate a natural word search interpreter, which utilizes familiar category and topic headings that are traditional to print 

directory media to generate relevant and related results to queries. By supporting a familiar navigation feature in our services, we 

provide end users with a more intuitive mechanism to search for and locate information. 

For our merchant services we have developed a comprehensive, enterprise wide data warehouse. This data warehouse contains 

information relating to merchants, products, services, users, customers, profiles, storefronts, purchases, site traffic and metrics. The 

aggregation of this information in one place allows us to leverage our development efforts and reduce redundant information. 

9