Development of a model driven architecture and software quality based framework for a distributed medical software system.

摘要

The use of "distributed" medical devices within an operating room to assist with surgical operations is very important. Medical devices are used to assist physicians with surgical procedures that would normally not be accomplished without some sort of computer-based interaction. These medical devices must be able to react to a surgeon's request with appropriate speed and precision. Medical devices provide greater automation of surgical and monitoring procedures and improved data collection capabilities that can be leveraged in medical diagnosis and treatment. These distributed medical devices have stringent performance and data quality requirements. Because of those constraints it is often a difficult task to decide which middleware will be used to provide the backbone of communication amongst these devices. Middleware, software that bridges multiple devices in a distributed system, is playing an increasingly important role across many application domains. Significant advances over the past years have made middleware both modular and customizable through the development of state-of-the art object-oriented frameworks and quality of service (QoS)-enabled component middleware. This brings out the problem of "how devices developed using different middleware technologies can and will interoperate". There are many commercial off the shelf (COTS) options that currently exist. COTS middleware have been considered as a viable solution for use with distributed medical device systems. When developing middleware for high-confidence medical devices (HCMD), many QoS constraints such as performance and throughput, as well as requirements must be taken into consideration. Many COTS offerings are not robust enough to fulfill QoS requirements as an out of the box implementation. These issues in particular, make selecting middleware for use with distributed medical device systems a non-trivial process. There are many middleware options available to the user including, but not limited to, developing an in-house software system. The problem that arises with the proliferation of in-house middleware software systems is that interoperability amongst these systems becomes virtually impossible. HCMD systems should be developed based upon an underlying framework that is system agnostic and promotes system interoperability. This framework should be abstracted from the implementation specific code and developed at the model stage of the software development process. When the framework is completed it is also important to get an understanding of the robustness of the models. Because the model-driven process requires the developer to build systems at a level of abstraction above implementation code, the developer must define the design quality of the system at that level of abstraction as well. It is important to gather metrics data regarding the object-oriented (OO) system design that can identify the quality of a system before one line of code is written. This can also go a long way in convincing the stakeholders of the software project at an early stage that the software being built is easily maintainable, scalable, and is a quality design. This thesis describes the development of a middleware framework and metrics gathering process for a quality design for use with HCMD software systems using the Object Management Group's (OMG) Model Driven Architecture (MDA) as the base for the development. We are building an MDA based framework that can be used by medical device software developers that are interested in building interoperability into their systems. We are also developing standards and processes for gathering metrics that define the design-time quality of the system.