BingoBo promises to present a well-organized web with vertical domains and geographic locations by classifying it into categories and sub-categories. It takes the initiative in designing and implementing the next generation of search engine using Agent Technology with a solid infrastructure in place to share information among various local sub-sites and the "US web".
From an architecture design point of view, scalability is considered first to reduce data redundancy and keep saved data as compact as possible for fast retrieval. It is achieved in the following ways:
1. Let users roam instead of making data replication across all sites or sending mobile agents traveling from site to site to collect data per user requests Most of the time, users will stay in their local area and switch between the local sub-sites and the "US web" sub-site which has the majority of the online information across the country. However, when users need to travel from place to place or find information on behalf of somebody from a remote sub- site, they can easily click an icon of a different sub-site or choose it from a list of locations in the same way as Craigslist. The sites they have visited will be cached and listed on the right side of the user interface for them to switch in-between easily. If some information is from foreign countries, they will be available under the WWW category of the "US Web". Other countries will have separate BingoBo systems to support their language and geographic locations. 2. Data is pre-partitioned into vertical domains instead of partitioned on the fly when data comes in Since it removes the heavy burden of document indexing, a distributed relational database architecture is designed from the very beginning for each vertical market sector to ensure fast read/write and returning of query results. When users want to query for particular things, it quickly narrows down to the vertical domain by human intelligence. Data is stored in the separate databases underneath each vertical domain in clustering and with failover mechanism built in for high availability. A caching mechanism is built on top of each database clusters accordingly. Each database cluster may sit on different nodes of the local area network for each local site. 3. Data is dynamic with an expiration date attached instead of staying in the repository forever Most of the data is pumped from the commercial database of the businesses when they pick and choose some of their products and services to be promoted on BingoBo. Businesses will pay different prices for the promotion data within a short period of time and the regular data to let users choose from. This way, BingoBo covers the data from the hidden web and exposes it to the users for them to find up-to-date information, which is updated automatically to their Private Web.
Therefore, BingoBo provides a robust system to handle large volumes of structured data and share it across multiple sub-sites dynamically and seamlessly for user access. It retrieves high quality search results as the data is organized in such a way that it can be retrieved quickly and easily from the categories and sub-categories, because most of the data is classified in advance before they are entered into the system.