A new Object Request Broker architecture in implementing a medical picture retrieval system

Area of Science:

• Computer Science
• Medical Informatics
• Software Engineering

Background:

• Implementing effective web-based interfaces for medical image retrieval systems is crucial for accessibility.
• Existing Object Request Broker (ORB) technologies, such as CORBA, present challenges in terms of type coherence, static Interface Definition Language (IDL) requirements, and maintenance difficulties.

Purpose of the Study:

• To describe the software architecture and implementation approach for the world wide web user interface of MediPict, a medical picture retrieval system.
• To present a novel ORB solution that overcomes the limitations of CORBA for the MediPict system.

Main Methods:

• Utilized a three-tier client/server architecture.
• Developed a custom ORB to address weaknesses identified in CORBA.
• Employed a unified programming language across the top and middleware tiers for improved coherence and maintainability.

Main Results:

• The implemented three-tier architecture with the custom ORB proved to be a beneficial approach for the MediPict web interface.
• The new ORB resolved issues related to non-coherent types, the need for a static IDL, and maintenance difficulties inherent in CORBA.
• Using the same language across tiers simplified development and enhanced system coherence.

Conclusions:

• The custom ORB developed for MediPict offers a viable and improved alternative to standard ORBs like CORBA for medical image retrieval systems.
• This approach enhances the usability and maintainability of web-based medical information systems.
• The findings suggest that tailored ORB solutions can significantly optimize the performance and development of specialized software architectures.