details

UTBS In Detail

Performance

The time-critical handling of mass data is a special challenge. This, amongst others, takes place in a real-time scenario, where a continuous data stream is processed real time. Any delay by a data processing being too slow unavoidably results in a backlog growing more and more. This process can not regain lost time and thus waiting periods increase.
In other scenarios of time-critical processing of mass data certain response times may not be exceeded, exceptional circumstances have to be recognized and have to be handled.
UTBS bases on a network of single systems (nodes), which are responsible for the execution of different tasks in independence of each other. By splitting general tasks into parts, which implicates a logical distributed system, there is no difficulty to process any number of records in shortest time by adding nodes. This performance only partially or with considerable investment of resources can be realized by ”physical distributed systems” as distributed database systems or distributed CPU-resources.

Total History

In conventional mass data systems records are handled in such a way, that conclusions to former contents or the development of the contents at later time are only possible in a restricted way.
However just the existing of history, a ”memory of knowledge”, is a necessary precondition of ”logical inferences” using existing knowledge and empirical facts to extract conclusions and to convert this conclusions into operations. ”logical inferences” are necessary preconditions for operation appropriate to the situation conditions especially in the sphere of ”artificial intelligence”.
UTBS does not change records and record information, UTBS continues information by generating new records. Thus a complete and qualified history is guaranteed, which allows a complete reproducible development of an event.

Complex Matching

Only the complete reproducible development of an event facilitates the establishing of conclusion systems within the scope of handling complex matching situations, which can be regarded as a starting point of each form of ”intelligence”.
Already today, within the frame of banking transactions, UTBS recognizes difference situations and their development using more than 20 characteristics to reveal these situations and (in the future) to execute user defined tasks for every determined case of difference.

Concept of Security

A change of control parameters of production systems, even administrated by proficient personal, is a potential reason of inconsistence, which leads to serious problems and lost time in data processing.
In ideal case parameter values, which lead to inconsistencies within a system or a connection of systems, will be intercepted by corresponding testing routines beforehand. UTBS guarantees this by falling back on complete history also containing the changes of parameters. This safety concept makes it possible, that trained laypersons without being an expert are able to manage and control the system at any time.

Development Status

Today the system covers approximately 350,000 lines of code, it is written in COBOL as well as in Visual Basic .NET and is finished regarding special central routines and central processes. At present an extensive parameterisation and dynamic sampling is implemented aiming at maximum flexibility for various application scenarios in mass data context.
In the sector of banking UTBS has already product character in a direct way. In other fields UTBS appears as a kernel system, which is a basis for trade-specific solutions in the sphere of mass data.

References

A predecessor system of UTBS developed for special requirements of a German major bank had been in use since 1987 in different stages of expansion for card supported transactions. A monthly volume up to 150 millions of transactions thereby had been handled.

back to the top