交易系统高可用性研究
|
摘要
随着网络技术的不断发展和计算机应用的不断深入，许多企业的日常业务都依靠计算机系统来进行，而这些系统大多是基于数据库的联机事务处理系统，往往维系着企业的生命。因此，人们对于计算机系统及依赖它的交易系统的高可用性的要求越来越高，人们不仅希望关键业务的数据信息能保持完整，而且希望联机交易能够不间断或者在最短的时间内自动恢复，这就有必要对交易系统的高可用性问题进行研究。
一体化银行卡消费系统是华夏银行的重要业务系统，是连接商家MIS系统和华夏银行卡业务系统的交易通道，而银行网关又是该系统的关键环节。目前，银行网关存在单点故障问题，降低了整个系统的可用性，也无法保证交易的一致性和完整性。因此，提高银行网关的可用性可以提高整个卡消费系统的可用性，对此进行研究具有现实的意义。
太原理工大学硕士研究生学位论文
本文从交易概念入手，分析了事务处理系统的发展状
况，阐述了事务的ACID特征以及事务处理的相关规范;然
后引出了高可用性的概念并分析了高可用系统的发展历程，
指出交易系统实现高可用性的必要性。
本文对造成系统故障的原因进行了分析，指出软件故
障是系统故障的重要因素。文中还对高可用系统理论作了
初步探讨，对系统可用度进行了分析。并对集群技术进行
研究，分析了双机集群的系统软硬件架构，对实现集群高
可用系统的关键要素进行了详细的阐述。
在以上理论的基础上，结合卡消费系统的实际情况，
本文改造了双机集群架构，改进了银行网关应用程序，设
计出了一套适用于一体化银行卡消费系统的高可用性方
案。此方案在提高整个系统可用性的同时，还保证了交易
的一致性和完整性。希望能为同类系统设计提供一些借鉴
和帮助。
With increasing improvement of network technology and the application of computer, many enterprises' daily businesses are mainly supported by computer systems. Most of these systems are the Online Transaction Processing Systems, and critical to enterprises. So there are more and more demand on the High Availability of computers and Transaction Systems. And people hope not only that the data of important business will be protected completely, but that online transaction won't be discontinuous or will recover automatically in the lest time. This is the problem of the High
Availability of Transaction System.
The Integrate System of Consume by Bank-card is an important business system of the Huaxia Bank. It is the transaction channel between the MIS System of merchant and the Bank-card Business System of the bank, and moreover, the bank's gateway is the key link of it. But at present the bank's gateway has single point failure problem, which drops lower the feasibility of the whole system and makes it impossible to maintain the consistency and perfection of transaction. So, to improve the feasibility of the bank's gateway is to improve the feasibility of the whole system. Research on this has realistic meaning.
The article, at first, puts forward the conception and the ACID characteristic of transaction, and expounds the development and the standard of the Transaction Processing System; then, introduces the conception and development progress of the High Availability, and finally points out that it is very necessary for Transaction System to achieve.
And in the article, the author gives the cause of system
failure, and indicates that Software failure is the major factor. He also makes some initial inquiry into the High Availability theory and analysis of the availability of the system in the article. With studying the cluster technology, the author analyses the framework of Hardware and Software of the two-computer cluster system, and describes the critical factors in the achievement of the High Availability of cluster.
Based on these theory and the actual condition of the Bank-card Consume System, the article transforms the framework of the two-computer cluster, improves the application software of the bank' gateway, and designs a set of High Available project suitable for the Integrate System of Consume by Bank-card. This project can improve the feasibility of the whole system and maintain the consistency and perfection of transaction as well.
Finally, the article is expected to be able to provide some useful experience and help to the similar system design.
一体化银行卡消费系统是华夏银行的重要业务系统，是连接商家MIS系统和华夏银行卡业务系统的交易通道，而银行网关又是该系统的关键环节。目前，银行网关存在单点故障问题，降低了整个系统的可用性，也无法保证交易的一致性和完整性。因此，提高银行网关的可用性可以提高整个卡消费系统的可用性，对此进行研究具有现实的意义。
太原理工大学硕士研究生学位论文
本文从交易概念入手，分析了事务处理系统的发展状
况，阐述了事务的ACID特征以及事务处理的相关规范;然
后引出了高可用性的概念并分析了高可用系统的发展历程，
指出交易系统实现高可用性的必要性。
本文对造成系统故障的原因进行了分析，指出软件故
障是系统故障的重要因素。文中还对高可用系统理论作了
初步探讨，对系统可用度进行了分析。并对集群技术进行
研究，分析了双机集群的系统软硬件架构，对实现集群高
可用系统的关键要素进行了详细的阐述。
在以上理论的基础上，结合卡消费系统的实际情况，
本文改造了双机集群架构，改进了银行网关应用程序，设
计出了一套适用于一体化银行卡消费系统的高可用性方
案。此方案在提高整个系统可用性的同时，还保证了交易
的一致性和完整性。希望能为同类系统设计提供一些借鉴
和帮助。
With increasing improvement of network technology and the application of computer, many enterprises' daily businesses are mainly supported by computer systems. Most of these systems are the Online Transaction Processing Systems, and critical to enterprises. So there are more and more demand on the High Availability of computers and Transaction Systems. And people hope not only that the data of important business will be protected completely, but that online transaction won't be discontinuous or will recover automatically in the lest time. This is the problem of the High
Availability of Transaction System.
The Integrate System of Consume by Bank-card is an important business system of the Huaxia Bank. It is the transaction channel between the MIS System of merchant and the Bank-card Business System of the bank, and moreover, the bank's gateway is the key link of it. But at present the bank's gateway has single point failure problem, which drops lower the feasibility of the whole system and makes it impossible to maintain the consistency and perfection of transaction. So, to improve the feasibility of the bank's gateway is to improve the feasibility of the whole system. Research on this has realistic meaning.
The article, at first, puts forward the conception and the ACID characteristic of transaction, and expounds the development and the standard of the Transaction Processing System; then, introduces the conception and development progress of the High Availability, and finally points out that it is very necessary for Transaction System to achieve.
And in the article, the author gives the cause of system
failure, and indicates that Software failure is the major factor. He also makes some initial inquiry into the High Availability theory and analysis of the availability of the system in the article. With studying the cluster technology, the author analyses the framework of Hardware and Software of the two-computer cluster system, and describes the critical factors in the achievement of the High Availability of cluster.
Based on these theory and the actual condition of the Bank-card Consume System, the article transforms the framework of the two-computer cluster, improves the application software of the bank' gateway, and designs a set of High Available project suitable for the Integrate System of Consume by Bank-card. This project can improve the feasibility of the whole system and maintain the consistency and perfection of transaction as well.
Finally, the article is expected to be able to provide some useful experience and help to the similar system design.
引文
[1] http://www.tongtech.com/zcfw/index2.asp?page=2
[2] 交易中间件与XA规范，http://www.huihoo.com/middleware/trade_middleware.html
[3] http://www.landersoft.com/fwzx/wdzl.asp
[4] 孟小峰，于戈，事务处理：概念与技术，北京：机械工业出版社，2004
[5] 司功闪，陆斌，胡樱，高可用性网络基础，北京：中国电力出版社，2002
[6] 姚世勇，高可用性Web集群服务器系统研究与实现，武汉大学，2002．5
[7] Chuck Hill，高可用系统的设计，电子工程专辑，2001．7
[8] David Kalinsky，高可用系统的硬件和软件设计模式，电子工程专辑，2003．5
[9] Chuck Hill，高可用系统软件启动模式设定，电子工程专辑，2001．7
[10] 潘学芬，一个简单有效的高可用性解决方案，中国呼叫中心发展产业研究报告，2001．8
[11] “一秒千金”培训手册，http://www.microsoft.com/china/ntserver/ha/hal.asp
[12] 刘庆超，东丰，赵晓鹏，计算机集群技术概述，蓝色航标，2002(1)
[13] 王春森，系统设计师教程，北京：清华大学出版社，2001
[14] 高文，服务器聚集系统中高可用性分析与设计方法，中国科学院计算技术研究所，2001．4
[15] 高文，祝明发，基于生灭过程的机群系统高可用性分析与设计，微电子学与计算机，2001(8)
[16] 黄翔，集群技术和进程迁移简介，http://www-900.ibm.com/developerWorks/cn/linux/cluster/mosix/part1/index.shtml，2002．1
[17] 唐俊奇，负载均衡技术的三种实现方法，微电脑世界，2003(4)
[18] 章文嵩，LVS集群的体系结构，http://www-900.ibm.com/developerWorks/cn/linux/cluster/lvs/part2/index.shtml，2002．4
[19] Philip M．Lewis，Arthur Bernstern，Michael Kifer，Databases and Transaction Processing，北京：高等教育出版社，2002．10
[20] ARP problem in LVS/TUN and LVS/DR, http://www.linuxvirtualserver.org/docs/arp.html
[21] 杨威，交换机集群技术在企业网的应用，微电脑世界，2002(17)
[22] 周金萍，徐丙立，姜小光等，Windows系统编程，北京：人民邮电出版社，2002．7
[23] Chris Oggerino，High Availability Network Fundamentals，北京：中国电力出版社，2002
[24] Monty Wright, An Overview of High Availability and Disaster Recovery for DB2 UDB, http://www-900.ibm.com/developerWorks/cn/dmdd/library/techarticles/0304wright/0304wrighteng.shtml, 2003.4
[25] Deploying Microsoft Windows NT Server for High Availability, outlook.microsoft.com/ntserver/docs/availabl.doc
[26] Delivering High Availability Solutions With Red Hat Cluster Suite, http://www.redhat.com/whitepapers/rhel/RHEL3ClustrSuiteWPPdf.pdf
[27] Clustered Failover High-Availability Solution, http://www.netapp.com/products/filer/clustered.html
[28] IBM WebSphere V4.0 Advanced Edition: Scalability and Availability, http://www.redbooks.ibm.com/pubs/pdfs/redbooks/sg246192.pdf
[29] 杨建宁，企业网的高可用性，微电脑世界，2002(15)
[30] http://www.dell.com/us/en/esg/topics/products_clstr_pedge_000_cluster_ha.htm
[31] Data Recovery and High Availability Guide and Reference: Introducing High Availability and Failover Support, http://www-3.ibm.com/cgi-bin/_db2www/data/db2/udb/winos2unix/support/document.d2w/report?fn=db2hae71db2hal1.htm#HDRHACMP
[32] Blair K. Adamache, Clustering for Scalability, http://www7b.software.ibm.com/dmdd/library/techarticle/adamache/0110adamache.html
[33] Data Recovery and High Availability Guide and Reference: High Availability on AIX, http://www-3.ibm.com/cgi-bin/db2www/data/db2/udb/winos2unix/support/document.d2w/report?fn=db2hae71db2ha12.htm#HDRHACMPES
[34] 范建华，胥光辉，张涛等，TCP／IP详解卷1：协议，北京：机械工业出版社，2000
[35] 曾声奎，赵廷弟，张建国等，系统可靠性设计分析教程，北京：北京航空航天大学出版社，2001
[36] 实现高可用性的集群技术选择，微电脑世界，2002(9)
[2] 交易中间件与XA规范，http://www.huihoo.com/middleware/trade_middleware.html
[3] http://www.landersoft.com/fwzx/wdzl.asp
[4] 孟小峰，于戈，事务处理：概念与技术，北京：机械工业出版社，2004
[5] 司功闪，陆斌，胡樱，高可用性网络基础，北京：中国电力出版社，2002
[6] 姚世勇，高可用性Web集群服务器系统研究与实现，武汉大学，2002．5
[7] Chuck Hill，高可用系统的设计，电子工程专辑，2001．7
[8] David Kalinsky，高可用系统的硬件和软件设计模式，电子工程专辑，2003．5
[9] Chuck Hill，高可用系统软件启动模式设定，电子工程专辑，2001．7
[10] 潘学芬，一个简单有效的高可用性解决方案，中国呼叫中心发展产业研究报告，2001．8
[11] “一秒千金”培训手册，http://www.microsoft.com/china/ntserver/ha/hal.asp
[12] 刘庆超，东丰，赵晓鹏，计算机集群技术概述，蓝色航标，2002(1)
[13] 王春森，系统设计师教程，北京：清华大学出版社，2001
[14] 高文，服务器聚集系统中高可用性分析与设计方法，中国科学院计算技术研究所，2001．4
[15] 高文，祝明发，基于生灭过程的机群系统高可用性分析与设计，微电子学与计算机，2001(8)
[16] 黄翔，集群技术和进程迁移简介，http://www-900.ibm.com/developerWorks/cn/linux/cluster/mosix/part1/index.shtml，2002．1
[17] 唐俊奇，负载均衡技术的三种实现方法，微电脑世界，2003(4)
[18] 章文嵩，LVS集群的体系结构，http://www-900.ibm.com/developerWorks/cn/linux/cluster/lvs/part2/index.shtml，2002．4
[19] Philip M．Lewis，Arthur Bernstern，Michael Kifer，Databases and Transaction Processing，北京：高等教育出版社，2002．10
[20] ARP problem in LVS/TUN and LVS/DR, http://www.linuxvirtualserver.org/docs/arp.html
[21] 杨威，交换机集群技术在企业网的应用，微电脑世界，2002(17)
[22] 周金萍，徐丙立，姜小光等，Windows系统编程，北京：人民邮电出版社，2002．7
[23] Chris Oggerino，High Availability Network Fundamentals，北京：中国电力出版社，2002
[24] Monty Wright, An Overview of High Availability and Disaster Recovery for DB2 UDB, http://www-900.ibm.com/developerWorks/cn/dmdd/library/techarticles/0304wright/0304wrighteng.shtml, 2003.4
[25] Deploying Microsoft Windows NT Server for High Availability, outlook.microsoft.com/ntserver/docs/availabl.doc
[26] Delivering High Availability Solutions With Red Hat Cluster Suite, http://www.redhat.com/whitepapers/rhel/RHEL3ClustrSuiteWPPdf.pdf
[27] Clustered Failover High-Availability Solution, http://www.netapp.com/products/filer/clustered.html
[28] IBM WebSphere V4.0 Advanced Edition: Scalability and Availability, http://www.redbooks.ibm.com/pubs/pdfs/redbooks/sg246192.pdf
[29] 杨建宁，企业网的高可用性，微电脑世界，2002(15)
[30] http://www.dell.com/us/en/esg/topics/products_clstr_pedge_000_cluster_ha.htm
[31] Data Recovery and High Availability Guide and Reference: Introducing High Availability and Failover Support, http://www-3.ibm.com/cgi-bin/_db2www/data/db2/udb/winos2unix/support/document.d2w/report?fn=db2hae71db2hal1.htm#HDRHACMP
[32] Blair K. Adamache, Clustering for Scalability, http://www7b.software.ibm.com/dmdd/library/techarticle/adamache/0110adamache.html
[33] Data Recovery and High Availability Guide and Reference: High Availability on AIX, http://www-3.ibm.com/cgi-bin/db2www/data/db2/udb/winos2unix/support/document.d2w/report?fn=db2hae71db2ha12.htm#HDRHACMPES
[34] 范建华，胥光辉，张涛等，TCP／IP详解卷1：协议，北京：机械工业出版社，2000
[35] 曾声奎，赵廷弟，张建国等，系统可靠性设计分析教程，北京：北京航空航天大学出版社，2001
[36] 实现高可用性的集群技术选择，微电脑世界，2002(9)