Processing

The basic requirement of any integration project is enabling disparate application systems to communicate with one another.  In order to make the output of one system useable as input to another, there are three levels of compatibility which must be addressed.  The most obvious is the presentation level.  For example, a employee’s last name that is stored as 20-characters in one system may have to be shortened to fit a 15-character field in the target application.  Going beyond simple reformatting, the next level requires manipulation of the data being exchanged.  If the personnel system contains each employee’s annual salary, but the payroll records contain only monthly salaries, salary figures transferred from personnel to payroll will have to be divided by twelve to be useable in the payroll system.

The transformations described so far are fairly simple, and are often handled adequately with automated mapping tools.  The real challenge comes at the third, or information, level.  To “understand” the transferred data, the receiving application must be aware of the associated context.  In the two previous examples, both systems contained similar context, employee’s last name and salary, so the data could be mapped simply between systems.  Often this is not the case.  Suppose financial records for a loan applicant contain information on the candidate’s credit history, current debt level, present salary, and monthly expenses.  The new record to be created for the applicant in the loan review system has a risk rating field with a value of “high,” “medium,” or “low.”  While the risk rating is not contained in any of the other records, the available information is sufficient to determine the risk factor.  To convey this information with conventional integration tools, a program, or “script,” would have to be created as part of the application-to-application interface.

ZONAR's Intelligent Information Interchange (I3) Processing Technology reaches inside computer-related presentation and data packaging layers to operate on the actual information content at the core.  Rather than depend on record formats and data type definitions, I3 is designed to process information based on its logical meaning regardless of packaging.  In the loan approval example above, transformation of the financial records to Information Objects would automatically generate the “Risk Rating” object because it is a property of the applicant which is implied by the objects already collected.  This approach provides enormous benefits by extending the utility of enterprise information currently trapped in computer data systems and eliminating the labor-intensive generation of custom transformation scripts.

The processing required to create and manipulate Information Objects is based on the following fundamental concepts of I3 Technology:

I-Space
In order to process information correctly and reliably in any legitimate context, its exact meaning must be known.  I3 Technology conveys this information content by defining every aspect of the information that affects its interpretation and explicitly stating the value of the information object for each aspect.  This is best illustrated by viewing each Information Object as an object in Information Space (I-Space).  Each aspect of the information is a separate dimension in I-Space.  Typical dimensions of information relating to physical objects might be color, weight, size, etc.  Each dimension may or may not contain a value for a particular Information Object, depending on how completely that object is defined.  Information Objects with common information content will overlap in I-Space, and Information Objects with identical content will occupy identical positions in I-Space, regardless of their record structures or data formats.  The I-Space dimensions of an object are reflected in the tags which are selected from the I3 Semantic Repository to build the XML representation of the Information Object.

Fuzzy Values
All real-world information is inherently "fuzzy" -- that is, we don't really know most things with perfect precision.  A value itself may be imprecise, like an "expensive" purchase; or we may not be sure of the value, it may be negotiable, or it may have changed.  In order to deal with the reality of imprecise information, the I3 Processing Technology handles all values as fuzzy.  This means they may contain ranges, sets, or even qualitative descriptions of quantitative items (a "large" room, for example).  An Information Object with fuzzy values may also be used as a query specification, since it describes the I-Space in which information results are sought.

Information Algebra
Information transfer takes place when new information is added to an existing context.  This is accomplished by relating the new information to each of the dimensions of the existing context (in I-Space).  ZONAR's I3 Processing Technology includes "algebraic" functions that operate on Information Objects to produce useful information transfer.  These include functions that merge, generalize, extract, subtract, normalize, and restructure Information Objects to produce results that fit into any desired enterprise information context.  This Information Algebra operates on the Information Objects directly, regardless of their original data sources or ultimate data destinations.  This makes them independent of any other applications.

Intelligent Agents
Information may often convey meaning beyond what is obvious on the surface, especially to someone expert in the relevant subject matter.  Those familiar with the Sherlock Holmes mysteries will recall Holmes’ uncanny ability to deduce seemingly unknowable information from the facts he has directly observed.  This technique is utilized to some extent by nearly everyone when they encounter new information.  The statement “Jack’s mother is a 35-year-old woman.” does not specifically state that Jack is a person, but that information can be reliably inferred from the fact that his mother is a woman.  It is almost as certain that Jack is less than 15 years old.  A less reliable, but still useful, assumption would be that Jack is a boy.  All of this information is relevant to Jack and should be included in the Information Object that represents him.  To implement this aspect of information processing, I3 includes an evolving set of Intelligent Agents which complement the information content of specific Information Objects.  Each of these agents is expert in the "meaning" of its Object and can produce all of the information that could be derived from the explicit content by an expert human.  These agents manage PersonName and Address Information Objects, for example, to provide the full sets of logical component parts and attributes (such as determining City from Zip Code, or Zip Code from Street Address, City, and State), regardless of the original data content and format.

[Back to Top]

[Next]

[Previous]