Font Size: a A A

Based On High Availability Windows Nt System

Posted on:1998-06-04Degree:MasterType:Thesis
Country:ChinaCandidate:S J ZhuFull Text:PDF
GTID:2208360185995484Subject:Computer software
Abstract/Summary:PDF Full Text Request
Recently, on-line transaction processing (OLTP) applications are widely used in the commercial fields. The market of OLTP applications is growing continuously. Highly available systems, which provided service continuity, data integrity, shortly down time and long mission times, are needed for OLTP applications. After successfully completing development of Unix-based high availability system, the National Center of Intelligent Computing systems (NCIC) set up a project to research and develop Windows NT-based high availability system.This thesis designs and implements Windows NT-based high availability system — NCICHA. NCICHA is composed of two servers, which act as hot partners for each other, dual-ported disks, remote administration console and management software. In the event of a server failure, applications can be automatically switched over from the failed server to the surviving, server — without administrator intervention. The operation system will not need to be rebooted. As a result, clients of the failed server can quickly regain access to data and application via the surviving server. NCICHA has following characters: (1) Auto-spy devices status; (2) Under failure state, applications auto-takeover; (3) Configure , monitor , control system at remote console; (4) Dual-ported disk array guarantees data security.Among existing Windows NT-based high-availability systems, as we know, there is no one with concurrent access. This thesis gives out a high-availability solution with concurrent access, through designing and implementing a model of the share-disk cluster file system — HADFS. HADFS guarantees the data consistency in concurrent access to disk array. Whenever a process reads data from a file, it receives the most recently written data. HADFS fully supports the POSIX and increases the availability by failure removal and fast system reconstruction. The distributed global lock manager ensures the data integrity and consistency in concurrent multiprocessing applications.Firstly, this thesis gives the basic concepts of fault tolerant technology and the principle of Windows NT. Chapter 2 describes and...
Keywords/Search Tags:high availability, Windows NT, takeover, cluster file system, concurrent access, data consistency, system reconstruction
PDF Full Text Request
Related items