The Active Information Platform (AIP)TM
PointofData's vision called for a system capable of storing real-time multimedia data collected on servers, personal computers and portable devices and to permit the exchange of that data in a free-agency marketplace with the ability to hide the details of content while advertising its availability. The same data sources - under the control of the creators of the data and with their permission - would be used for information discovery as well as search and analysis applications, including real-time data layering, geographical display systems and real-time applications, such as instant advertising. Data would never be deleted; it would be accumulated continuously for the life of the individual or entity, and each event collected would be associated with a time and place.
These requirements necessitated a massively distributed, scalable, event-oriented data management system which could tap into a variety of data sources, create scalable local databases and make use of the best available information analysis technologies for custom applications. PointofData designed, developed and delivered the Active Information Platform (AIP) to fulfill these requirements. The Active Information Platform has five key features: Dynamic IndexingTM, Scalable/Distributed Database, HyperSearchTM, Intelligent AgentsTM and Distributed SecurityTM.
Dynamic Indexing eliminates shortcomings of existing indexes, such as the need to reindex data or maintain multiple indexes. Dynamic Indexing permits efficient range-value and approximate searching, performs efficient searches on non-contiguous key segments or ranges and produces very small indexes for a given set of source data. A Dynamic Index for full-text searching is a fraction of the original data's size, and the size of the index as a percentage of the original data size may actually shrink as entries are added, depending on the content of the original text. By contrast, in many full-text search systems, it is not uncommon for the index to approach or surpass the size of the original data.
Scalable, Distributed Database
PointofData seamlessly layers and integrates information to provide comprehensive records of time-stamped events. The Active Information Platform treats each data point, whether an SQL row or an edited document, as a discrete event and allow these events to be related, permitting the creation of associative relationships between information. Events are defined as objects with a series of named attributes. Each attribute associated with an event may contain structured or unstructured data, and there may be any number of versions of the event data. Events within the Active Information Platform may be related explicitly by defining new events containing connectors to existing events, or events may be related implicitly by patterns found in or synthesized from attribute data. All of this data can be stored in 80% less space than conventional technology requires.
HyperSearch finds information, whether its structured or unstructured, stored online, nearline or offline, with a single query. Additionally, Hypersearch identifies relevance, or patterns and relationships within data, as rapidly as single words. A user can establish parameters and search the data using them. For example, the user can define altitude, time and location as parameters, combine these parameters and specify ranges within them. The query might look like this: What planes traveled between 10000 and 20000 feet, between 1:00 and 2:00 pm, within thirty miles of Seattle? The user can then change details relating to a parameter in order to narrow in on desired information without having to submit a new query. In this instance, that would mean adjusting the ranges or number of miles from Seattle.
Intelligent Agents easily interface with data stores and data gathering devices. Each agent functions as a discrete event manager, interacting with other agents hosted on systems as small as a cell phone or as large as a mainframe. Intelligent Agents function in active (or push) mode and reside on live systems, as opposed to passive (or pull) mode where information is indexed from backup logs or redo logs to another location. In active mode programmable agents allow a customer to automate business processes, collect or receive data, or enable communication between devices via a messaging system.
The Active Information Platform includes a patent-pending Distributed Security module, which protects sensitive data sources. This module offers data protection based on query and usage patterns, as well as imperative security rules. Intelligent Agents and their data sources can further be rated by trustworthiness and controlled according to predefined standards, using PointofData's Trusted SearchTM module.