A department store

Client introduction

A department store is main board listed company in Hong Kong stock exchange, and it holds multiple A-share listed companies. After over 10 years of significant development, with annual sales of nearly 10 billion yuan, this company has become the leading company in China retail sales industry. At present, it has 40 stores in 20 cities throughout south China, southwest, north China, and east China areas.

Background and client demand

Each store’s sales management system of this company is composed of front-end application management service and Sybase database service running on RedHat 5.4 platform. As Sybase database service is the core back-end database of front-end sales management app service, it’s undoubtedly the most important part in the system. Before implementing this project, thecustomer's back-end Sybase database runs on the stand-alone system.

Project goal:

It provides a basic business continuity environment, to guarantee that all its stores’ sales management system run continuously without interruption, which improves their sale efficiency and reduces the human and financial cost of system management.

The solution

As the key point of this solution, the core back-end Sybase database of store sales management system should run continuously and uninterruptedly to guarantee high availability of the system.With careful screening and with consideration to client's existing resources, Rose Datasystems recommends the business continuity product that is based on data mirroring technology. That is to build a mirrored active/standby solution by the sales management systems’back-end Sybase databases of eight urban stores that have vast passenger flow, to realize business continuity.

Overall architecture description

Hardware platform of each store may not be the same. Some store may need to add one server as the standby server of Sybase database, and some store can use the front-end application server as the standby server, to make full use of hardware resources. When software and hardware environment are ready, user can start to build active/standby cluster based on data mirroring technology with the solution of Rose.


Here we take a store as example to elaborate it.

Normally, user can select one of the two servers as Sybase's active server. It will provide external service through active IP, and the data generated by active server is written into its own local disk. At thesame time, Rose solution will transmit the changed data that is captured in real time to standby server through network, to realize that data on two servers is consistent. On this basis, if the activeserver encounters a failure (sever going down; application system fault; or network fault, etc), the protected application cannot continue providing service. Then, under the premise of data consistency, via Rose solution, active server would switch Sybase database to standby server to let it continue providing service, and guarantee that production and management system would run continuously. 

Business system continuity

Monitor the running state of Sybase database in real time. When hardware/software failure occurs, it will automatically switch database service to the standby server, to ensure database continuity

Real-time data replication

Using online real-time data replication, it ensures data consistency of active and standby server. And it supports scheduled snapshots task, so it can create snapshot records for data regularly, to further ensure the data security.

Multiple monitoring methods

It monitors application running state in real time, and it could monitor CPU/memory resources usage rate in depth, and execute intelligent warning and strategic failover.

Flexible architecture

The solution architecture is quite flexible. User can create active/standby cluster without disk array. By making full use of existing software and hardware resources, user can build active/standby cluster solution easily which avoids repeating investment. 

Simplified operation and maintenance

Rose offers friendly graphic interface, and user can manage active/standby cluster remotely, and monitors cluster working status. It has a variety of alarm ways, such as: online status, online logs, SMS notification, mail notification and etc., which makes it easy for user routine management, so as to simplify and ease the operation and maintenance difficulty.