| In light of the development of China's security trading industry and owing to the need of the security exchange companies for satisfying their business development as well as the requirement by the superintendent authorities of securities exchange for a centralized control and management, the previous modes of both the securities exchange and the management of the transactional data based on the individual stock exchange local divisions are no longer suitable for such a situation and it is more and more urgent for us to adopt a centralized exchange and trading system based on the financial companies of the headquarter-class in order to uniformly manage and control all transactions and trading data within a whole corporation. The centralized exchange system of such a kind shall be beneficial for lowering exchanging and agent costs for the security dealers and realizing a co-sharing of information resources, consolidating the centralized control and management, minimizing operational risk as well as enhancing the comprehensive competitiveness of the security exchange dealers.During the process of realizing a centralized securities exchange, it shall become a vital issue for the system as for how to clear the single-point failure by fully considering the redundant backup in all links beginning from the network equipment, communicating lines, routing strategy and the key-server up to the data storage and access methods during the software and middleware designs etc. so as to improve the operating efficiency and ensure an uninterrupted service provided by the core systems.Based on the research of the centralized security exchange practice currently adopted by Chinese security exchange dealers as well as with the help of this author's experience in clearing off failures in the security exchange system for more than 10 years, this article has integrated the theories of computer network, database middleware and clustering service etc. including the network redundant technologies of HSRP (Hot Standby Routing Protocol), DDR(Dail-on-Demand Routing) , FEC/GEC (Fast Ethernet Channel/Giga Ethernet Channel) and Spanning-Tree (Spanning Tree Protocol) etc. as well as the clustering service of Windows 2000 server and the middleware load balance to create a complete set of redundant backupsolution for clearing off the single-point failure in the centralized security exchange, in order to ensure that the requirements of high efficiency, high stability and high security necessary for the security exchange system are met. This solution is quite practicable and advanced, which is suitable for all security transaction dealers in China to be deployed in the centralized security exchange system.This article has discussed the realization of clearing off failure in the centralized security exchange system of the corporation headquarter class in the real trading environment, which enables this system become one of a few primary successful solution cases in Chinese security trading industry. A series of fault toleration technologies were adopted in the process to solve the single-point failure in network equipment, communication lines, servers, middleware as well as in the communications of offer submission and transaction receiving purposes etc. In the process of building the network platform, two sets of redundant equipment were adopted in the primary and backup routers and trunk exchangers at every corporation headquarter and trading business center, i.e. the primary communicating line of 2M and the redundant line of ISDN, as well as the dynamic route protocol OSPF (Open Shortest Path First) in the WAN primary line and the floating static router in the backup line. Together with the deployment of technologies of HSRP, DDR, Vlan, Trunk, Spanntree, it is possible to establish the following clearing off single-point failure- between LAN exchangers, between LAN exchangers and routers ,between routers' interfaces, and between routers and the WAN links. Two HP DL760 G2 servers with one disk array cabinet were adopted as the redundant equipment for the transaction server. In addition, the clustering technology of Windows 2000 was also used to clear off the single-point failure in the transaction servers. During the process * of designing the centralized transaction software, we have adopted the data storage and access method that is capable of separating the transacting data from the historical data and supporting a multiple historical query and a middleware that supports a dynamic load balance, the multiple chair offering that supports a load balance. Those functions have fundamentally resolved the redundant problems in the data storage and access processes, in the middleware of the connection and communication and in the multiple chair offerings etc. as well as enables an uninterrupted service by the backupsystem during happenings of failure in the data server, middleware and m the offering equipment and communicating lines. In the connecting issue with the stock exchanges, we use the satellite and land special lines provided by the Shanghai and Shenzhen Security Exchange to secure the backups that are transmitted both in air and on the land.As we have comprehensively adopted all of the key elements - the redundant backup technology of a centralized security trading system, the single-point troubles are overwhelmingly wiped out and the system running efficiency is greatly enhanced while the failure rate is also considerably decreased. In addition, the resource co-sharing is realized, which has made it possible for the security exchange companies to lower their daily operational cost. Before the adoption of the centralized exchange system with the single-point failure recovery function, the overall trading systems of the stock exchanging merchants were very often broken off as a result of failures in the communication line, the hardware equipment and the trading system software etc. At that time, the average broken off time was 15 minutes per month for the subordinate trading stores of the corporation. Since the centralized trading system was adopted with the function of the so called single-point failure-recovery, the technical failure causing an exchange stoppage has never happened, while the average failure time has consequently decreased to zero. Even if some uncontrollable failures do happen in the communication line etc., the system can be automatically shifted to the backup line within a period of a couple of seconds without people's intervention, which has assured that the security exchanging core system can provide a uninterrupted service. Since the failure rate is greatly decreased, the stock dealing departments are able to save their cost in compensating the trading errors, as well as' avoid the effort paid to dealing with the black out incidents. Moreover, as the centralized security exchange system is established based on a corporation headquarter class, a maximum sharing of all resources is realized for all of the trading shops subordinated to the enterprise, i.e. the equipment and device of the trading server, the price offering chair and the price offering channel etc., which is able to save an one-time investment of 5 million yuans as well as economize the operating cost of 1 million yuan per year for a stock trading corporation who has 20 subordinatebranches (among them, 10 are newly established). As a result, the economic efficiency is enhanced for the enterprise. During its promotional process, the project was well praised and warmly welcome by the relevant users, hence it can be so said that the project has a quite high practical value with a big popularizing potential. |
PDF Full Text Request