Research Of Data Disaster Recovery Technology Based On Middleware

As mankind's reliance on computer information systems is growing, the vulnerability of information systems is getting more and more obvious. The uncertainty of a disaster will result in devastating damage for information system. In order to effectively protect data and information systems operations to maintain sustainable operation of the system, it has become the focus of attention to look for an effective way to protect information systems. At present, the disaster is the key subject for the research institutes, and companies also develop a lot of disaster recovery products, many of which have been adopted and put into practice by enterprises. Compared with the mature disaster recovery products aboard, the domestic products need a sound disaster recovery standards and norms due to the newly developing field; there is s trend that the domestic companies invested and developed blindly.Disaster Recovery which includes data storage, data replication, data backup (restore), failure detection, data transfer and many other technologies is a system engineering. It could build a comprehensive disaster recovery system, if it combines with the actual situation which is considered many factors. Data is the core of the information systems. An effective data protection is key to disaster recovery and the prerequisite for data disaster recovery systems are the most important aspect, the guarantee after a disaster, information systems, business data security available. Divided by distance, data disaster recovery, including local data disaster recovery technology and remote disaster recovery data disaster recovery technology, according to disaster recovery technology based on the implementation level is divided into hardware and software-based approach. The data disaster recovery system is usually implemented as the local data backup / restore functionality, and backup of data can be saved to the remote client, and then using data replication technology, through the network (IP network, SDH, ATM, etc.) will produce the data center synchronous or asynchronous replication to remote disaster recovery center, the local production systems can also be configured, such as high availability clustering technology, which can guarantee that the face of various disasters, the recovery strategy are to ensure data integrity, consistency and availability. At present, because of different implementation of distributed information system, the application of specific technical and differences of software, differences of business functions, scattered data storage location, the implementation of distributed information system disaster recovery often has more problems, such as high cost of foreign disaster recovery products, the burden of developed human, material resources, necessary time and expanding system causing revising disaster recovery functions. Middleware technology has a function of shielding the distribution system. It leads a certain extent on the system to increase reusability and reduce the complexity of the development. So middleware technology which will be applied to the disaster recovery areas should present good view.Based on the research of disaster recovery technology, middleware technology, the characteristics of distributed information systems, under the guidance of the formal model of disaster recovery systems, proposes the data disaster recovery model based on middleware. The model includes backup / restore, data replication, service switching functions and so on. Disaster recovery system achieve three layers including the application layer, the middleware layer of disaster recovery and service level. Service layer contains a variety of disaster recovery agent unit that deal with different data for different storage, different storage location using the corresponding technology. Disaster Recovery agent unit achieves disaster recovery data collection, processing, classification, transport and other specific operations of backup related to the specific data, while the layer also includes a storage media agent unit that transparent bottom of the backup data in different storage media. Disaster Recovery middleware layer is mainly responsible for dealing with disaster recovery operations, guiding the implementation of disaster recovery agent unit to enable service layer transparent to the application layer, which includes disaster recovery agent units management components, message passing components, data transmission component and other components. Application layer is the specific function settings of disaster recovery, such as the customization of disaster recovery strategy, the trigger of disaster recovery, the user log management. All the layers of disaster recovery communicate based on the disaster communication protocol that facilitates the different implementation technologies with every layer. Distributed information system for the extension can flexibly deploy the appropriate disaster recovery agent unit, when having the new data storage required to support disaster recovery, just need increase the corresponding disaster recovery agent unit. Comparison of fixed Disaster Recovery middleware layer, modifications are often at specific component. Data Disaster Recovery Model based on middleware makes the disaster recovery of distributed information system simplified, reusable, extensible. MOSS is an information exchange and collaboration web service platform that is issued by Microsoft in recent years. As Microsoft's enterprise web application platform, MOSS integrates Content Management, Business Intelligence, Workflow, Enterprise Search Engine, Enterprise Application Integration and other enterprise services. At present, there is a lot of business based on the MOSS platform to build its enterprise applications, for example, the opening ceremony of the Beijing Olympic Games using MOSS as its content search engine. Therefore, building the MOSS platform disaster recovery is more important. The technologies MOSS platform referred to include database, IIS6.0, ASP.Net2.0, workflow, search engine, etc; The deployment of MOSS platform is very flexible, such as Single-Server Configuration, Small Farm Configuration, Medium Farm Configuration and Large Farm Configuration; there is also a wide range of data storage, such as database, file system, system status, etc.Finally this article in accordance with the Data Disaster Recovery based on Middleware Model implements Data Disaster Recovery System of MOSS platform, through the MOSS large farm deployment shows that the MOSS data disaster recovery system architecture, moreover, by analyzing its business process, so that MOSS data disaster recovery system is able to adapt to different topologies. In the functional implementation, firstly, analyzing the data in MOSS farms, abstracting the disaster recovery data required that ensure MOSS farm available and data integrity through recovery function when encountering disaster; At the same time, describing the organizational form to disaster recovery backup data. Then this article analyses MOSS disaster recovery agent unit required, and shows the implementation of the necessary technologies and methods. Designing and implementing the MOSS Disaster Recovery middleware layer that includes Disaster Recovery Protocol component, customized message passing component, etc. Finally, the specific implementation to MOSS farm backup function, through selecting the disaster recovery data at interface and formulating the disaster recovery plan, could run backup job immediately or run at the specific time point. Backup job can make use of the flexible disaster recovery strategy before it starts running. MOSS farm recovery functions include the in-place recovery and the out-of–place recovery. In-place recovery is the function that restores the backup data to its original position and out-of-place recovery is the function that restores the backup data to a different MOSS farm, that is getting the data from a MOSS farm and restore it to another MOSS farm. The biggest difference between out-of-place recovery and data replication is that out-of-place recovery requires the backup data storage media offer the data, but it should also belong to the remote disaster recovery and improve the flexibility for MOSS disaster recovery. MOSS platform on disaster recovery of the experimental data prove that data disaster recovery model based on middleware is available and reliable, provide a data disaster recovery basis and guiding role for distributed information system.