Oltp Vs Dss Systems

Writen by Kostis Panayotakis

Information systems are classified into two major categories, according to international developments: A. On-line transactional processing systems (also called operational systems)

B. Decision support systems (DSS)

Α. On-line transactional processing systems OLTPs are systems which serve transactions with suppliers, partners and customers, as well as internal business transactions. They support operations throughout the value chain of the Organization:

• Supply Chain Management (SCM)
• Production support (e.g. MRP, Advanced Planning & Scheduling)
• Customer interface management (e.g. sales, order management and billing) (CRM)
• Finance and Accounting (ERP)
• Sales force automation
• Web channel operations (eCRM)
• Internal workflow support systems

Β. Decision support systems DSS provide management at all levels of the Organisation, with information which supports understanding of the current Business position and taking informed decisions (fact based management). OLTP vs DSS systems Even though OLTP (on-line transactional processing) and DSS (decision support systems) functionalities may overlap (e.g. an OLTP system may provide some operational reporting functionality used for decision support), it is clear that the purpose of the 2 categories differs, given that they serve different functions and different User groups in the Business. Therefore the development philosophy of the two categories differs radically. Specifically, differences are identified on the following criteria (1 for OLTP, 2 for DSS): System functional requirements:

1. Clearly specified given that the system serves specific functional needs – the predetermined transactions
2. the determination of a complete requirement set is a challenge, given that there are dynamically changing informational requirements.

Capture of current and historical information:

1. Current state information is captured (some historical data may exist only to serve potential future transactions)
2. Recent and historical information is captured (current may not be captured, given that data from the OLTP are retrieved at regular intervals)

Data models used:

1. Complex, focused on business entities (in terms of relational databases it is called normalized data structure (e.g. 3NF))
2. Different approaches exist. The simplified denormalised dimensional structure gains momentum, since it allows easier understanding by business users and optimized execution of complex queries.

Information level of detail:

1. Detailed data per transaction are kept
2. Detailed data are kept in a different structure and are enriched by 'dimensional' information which allows analytical processing. Moreover, aggregated data like KPIs (key performance indicators), are calculated and stored in persistent storage.

Volume of data:

1. The volume of data is relevant to the size of the Business and the penetration of IT in it.
2. The data volume handled by a DSS, is multiple of that of the OLTP systems on which it is based, given that it maintains multiple historical snapshots

Copyright 2006 – Kostis Panayotakis

Material relevant to business intelligence can be found at www.pleroforea.com