Font Size: a A A

Researches On Key Issues Of Dependable Middleware Technology In Object-Oriented Disbributed Environment

Posted on:2007-01-20Degree:DoctorType:Dissertation
Country:ChinaCandidate:Q L LiFull Text:PDF
GTID:1118360185956753Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
The rapid technical progress of network communications and object-oriented software increases requirements of people to distributed object applications since 1990s. Dependable computing plays an important role in many distributed object applications. However, traditional fault-tolerant technology mostly aims at centralized computing environment, lacks of object-oriented mechanism for dependability guarantee, and increasingly exposes its limitations in face of a series of problems such as node failure, object crash, value fault and network partition. Therefore, traditional fault-tolerant technology is not applicable in distributed object environment. It's significant and valuable for both of researchers and engineers to investigate the dependable computing technology in distributed object environment.On the background of middleware, distributed object computing, dependable computing and fault-tolerant computing, the author explores dependable middleware technology in distributed object environment. Some targets of the research works include the replication consistency technology based on group communication service, the active replication technology with majority voting, fault management in distributed object environment. Some innovative contributions of the dissertation are enumerated as:· Proposing a replication consistency protocol combined with group communication in partitioned network environment. Owing to the dynamics, uncertainty of network environment and characteristic of applications, distributed object environment will produce network partition frequently. Traditional replication consistency protocols, which are based on virtual synchronous communication model, strongly depend on reliable connection of network hardware and have obvious limitations in partitioned network. Thus a new replication consistency protocol, which is aimed at partitioned network environment, is proposed to solve these problems. The protocol introduces eventual path to achieve update information in the case of network partition, adopts dynamic linear voting algorithm to ensure global...
Keywords/Search Tags:dependable computing, fault-tolerant technology, distributed object technology, middleware, replication, group communication, majority voting, fault detection, recovery
PDF Full Text Request
Related items