1996 Fiscal Year Final Research Report Summary
A Real-Time Object Architecture for Large Scale Distributed Systems
| Project/Area Number |
06452244
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (B)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
情報通信工学
|
|---|
| Research Institution | Faculty of Science and Technology, Keio University |
|---|
Principal Investigator |
TOKORO Mario Keio University, Faculty of Science and Technology, Professor, 理工学部, 教授 (10051898)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
TAKASHIO Kazunori The University of Electro-Communications, Department of Computer Science and Inf, 電気通信学部, 助手 (40272752)
|
|---|
| Project Period (FY) |
1994 – 1996
|
| Keywords | Distributed System / Real-Time Computing / Object-Oriented Computing / Concurrent process Culculus / Scheduling Algorithm / DROL / DtCCS |
|---|
| Research Abstract |
As the utility of distributed systems is widely recognized, real-time computing in such distributed environments will play an important role for various applications such as distributed multi-media systems consisting of multiple computers connected by high-speed networks, large scale air traffic control systems where high-speed networks connect neighboring airports, and satellite monitoring systems that have to control multiple remote satellites.
Most of previously proposed real-time programming languages has been designed to describe centralized real-time systems. Thus, a great deal of effort has been made to construct statically predictable real-time systems. However, in distributed environments at which we aim, networks are not always reliable. This lack of reliability brings unexpected delay to message transmissions, or causes the loss of messages.Consequently, it makes difficult to analyze timing characteristics of software which communicates with other nodes with a static and unif
… More
ied method. Moreover, it also becomes difficult to analyze whole of system's behavior from a global view point.
Thus, the predictability on which in the category of the centralized real-time systems we have laid stress is meaningless. A new programming paradigm for distributed real-time systems is needed. This dissertation proposes the Distributed Real-time Object model, as a new programming model that can flexibly deal with such unexpected situations. Then, a programming language DROL that is an extension of C++ is designed and Implemented. A distributed real-time object defined in the programming language DROL is an active object which encapsulates timing information and supports two strategies, best service and least suffering.
In order to demonstrate the effectiveness of the distributed real-time object model and the design and implementation of the programming language DROL,we describe programming examples of DROL and evaluation results of runtime performance concretely.
We also present a process calculus for timed concurrent systems and study its theoretical properties. The calculus is an extension of Milner's CCS by introducing two timed related notion : a tick action and a timeout operator. The operational semantics of the calculus is given by a labelled transition system. It allows to describe not only the functional behavior of concurrent systems but also variant time properties of them, such as quantitative time, timed behavior on the passage of time, and communication delay. It enjoys most of the fundamental framework of CCS almost unchanged. Less
|
