TOPIC

DECISION-SUPPORT SYSTEMS (DSS)

Database-centered client/server applications fall into two categories: They are:1.Decision-Support Systems(DSS) 2.Online Transaction Processing(OLTP)

DECISION-SUPPORT SYSTEM(DSS)
Decision-Support Systems (DSS) are used to analyze data and create reports. They provide the business professional and information hounds with the means to obtain exactly the information they need. A successful decision-support system must provide the user with flexible access to data and the tools to manipulate and present that data in all kinds of report formats. Users should be able to construct elaborate queries, answer what if questions, search for correlations in the data, plot the data, and move it into other application such as spreadsheets and word processor documents.

Decision-support systems are not generally timecritical and can tolerate slower response times. Client/server decision-support systems are typically not suitable for mission-critical production enviornments. They have poor integrity controls and limited multitable access capabilities. Finding information may involve large quantities of data, which means that the level of concurrency control is not very granular.

Decision-support systems are built using a new generation of screen-layout tools that allow nonprogrammers to build GUI front-ends and reports by painting, pointing, and clicking Point-and-click query builders take the work out of formulating the question.

ONLINE TRANSACTION PROCESSING(OLTP)

OLTP Systems are used to create applications in all walks of business. These include reservation systems , point-of-sale, tracking systems, inventory control, stockbroker work stations, and manufacturing shop floor control systems. In OLTP systems, the client typically interacts with a Transaction Server instead of a Database Server.

Comparing Decision-Support and OLTP Systems

Decision-support applications can be created directly by end-users. Network administrators are still needed to help set up the client/server system, and Database Administrators may help assemble collections of tables, views, and columns that are relevant to the user. The design of client/server systems for OLTP is a lot more involved. If the OLTP system stops, your business stops.

Comparing the programming Effort for Decision Support and OLTP

Client/Server Application Decision Support Client Off-the-shelf decision-support tool with enduser scripting. Canned event handlers and communications with the server. Custom application. GUI tool lays out screen, but the event handlers and remote procedure calls require programming at the C or Java level. Server Messages

Off-the-shelf SQL queries and database server. joins. Tables usually defined by DBAs as part of a data warehouse.

OLTP

Custom application. Transaction code must be programmed at the C or Java level. The database is offthe-shelf.

Custom function calls are optimized for performance and secure access.

Database Needs : DSS versus OLTP

FEATURE Who uses it? Timeliness of data Decision-Support Database Needs Information hounds Needs stable snapshots of timestamped data. Point in time refresh intervals are controlled by user. Reports can be reconstructed using stable data. Sporadic OLTP Database Needs Production workers Needs current value of data. Reports cannot be reconstructed.

Frequency of data access

Continuous throughout workday. Workrelated peaks may occur.

FEATURE Data format

Data collection

Data source known?

Decision-Support Database Needs Multiple levels of conversations, filtering, summarization, condensation, and extraction. From multiple sources-internal and external. No, it comes from different applications, databases, the web, and ERP systems.

OLTP Database Needs Raw captured data. No derived data. Detailed and unsummarized transaction data. From single application. Yes, most of it is generated by single application.

FEATURE

Decision-Support Database Needs Timed snapshots or Yes, you can key off multiple versions? a snapshots date/time. Each snapshot is a version unless you overwrite it during refresh. Data access pattern Mostly single-user access. Intense usage on an occasional basis. For example, when a report is due. Can data be Read-only, unless updated? you own the replica.

OLTP Database Needs No, continuous data. Single version.

Multiple users updating production database.

Current value is continuously updated.

FEATURE

Decision-Support Database Needs Flexibility of access Very flexible via a query generator, multi-table joins, and OLAP.

OLTP Database Needs Inflexible, access to data via precompiled programs and stored procedures. Performance Relatively slow. Fast response time is a requirement. Highly automated, repetitive tasks. Data requirements Fuzzy and unstable. Well understood. A lot of detective Known prior to work and discovery. construction. Subject-oriented data. Information scope Data can come from Finite. Whatever is anywhere. in the production database. Average number of 100s to 1000s of Less than 10 records accessed records in sets. individual records.