based on. Host-based data replication technology, you can not consider the isomorphism problem, as long as the host is the same operating system can be, and there is also support between heterogeneous hosts, such as NetVault Replicator can support heterogeneous servers Data replication, can support remote real-time replication across the WAN. The disadvantage is the need to take up a little host resources.
Based on storage systems. Using the data replication software provided by the storage system, the copied data streams are passed between the storage systems and are independent of the host. The advantage of this approach is that data replication does not take up host resources, the shortcomings of the need to disaster recovery center storage systems and production center storage systems have strict requirements, generally need from the same manufacturer's storage system, so that the user's disaster The selection of the storage system of the center has brought limitations.
based on. This technology is developing, using the new features of the fiber switch, or the use of management software to control the fiber switch, the storage system virtualization, and then management software management of the pool volume management, volume replication, volume mirroring and other technologies to achieve Remote replication of data. More typical of Storag-age, Falcon and so on.
Application-based data replication. This technology has some limitations, are for specific applications. Mainly using the database itself to provide the replication module to complete, such as OracleDataGuard, Sybase Replication and so on.
Analysis and selection of disaster backup solutions
In the process of implementing computer applications, banks have been deeply aware of the importance of system availability to banking. In addition to ensuring high availability to the local room, they also recognize that there is still a need for protective measures such as natural disasters, equipment damage and other catastrophic events to ensure continuous operation of the business system The corporate image, to ensure the normal financial order of society.
In fact, the prevention of disaster construction is part of crisis management. According to an American study, 75% of the company's business will be completely paused and 43% can no longer be reopened after the disaster, and 20% of the companies are in two years Was forced to declare bankruptcy. On the other hand, the impact analysis for the disaster shows that the industry's longest tolerable information system downtime is: financial industry -2 days; sales -3.3 days; manufacturing -4.9 days; insurance industry -5.6 days The On average, the general industry can tolerate the information system downtime for 4.8 days (Source: University of Minnesota). If the loss of operating income, the financial sector suffered the most serious losses, up to 50% of daily operating income.
This shows that the financial industry's dependence on information technology, especially in other industries, more difficult to tolerate because the computer center to stop the service caused by heavy losses. At the same time, the realization of disaster recovery is not a mere technical problem, it will be involved in the entire enterprise may encounter problems, so most of the business decision-makers are aware of the disaster backup plan for the importance of the enterprise.
The disaster recovery plan requires a thorough preparation, in particular the need to analyze the impact of the disaster on the business and the corresponding disaster recovery strategy, at the same time with the latest information technology, put forward the best recovery program. After the establishment of the system backup plan, must also be repeated in advance testing, and at any time to adjust to be improved, a complete disaster backup solution was established.
According to the definition of the international standard SHARE 78, the disaster backup solution can be divided into seven levels according to the level of the main consideration listed below, and there are seven different levels of corresponding disaster backup solutions from low to high. The bank's financial regulatory system can design and implement plans based on the importance of the data and the speed and degree of recovery that needs to be achieved.
• The range of backup / recovery
• The status of the disaster recovery plan
• The distance between the production center and the backup center
How is the connection between the production center and the backup center
How does the data transfer between the two centers?
Allow how much data is lost
How to ensure that the updated data is updated at the backup center
• The backup center can begin the ability to back up work
In 1992, Anaheim's SHARE 78, M028 report, the automatic remote recovery task is defined in seven levels:
1. No off-site data
Is defined as the need for no information storage, there is no need to establish a backup hardware platform, there is no need to develop contingency plans, the data is only locally restored, no data sent to the off-site. This approach is the most low-cost disaster backup solution, but in fact this disaster backup does not really have the ability to back up disaster because its data is not sent away from the local place, and data recovery is only exploited Local record.
2. Pickup Truck Access Method
Disaster backup programs need to design a contingency plan that can back up the required information and store it in remote locations, and then selectively build a backup platform based on the specific needs of the disaster backup, but does not provide a hardware platform.
Is a standard way for many central backups. After the write is complete, the data will be sent to a place away from the local area, along with some procedures. After the disaster, a complete set of systems and application installation actions need to be completed on an unbooted computer. The system and data will be restored and re-connected to the network. This disaster recovery scheme is relatively costly (requiring only the consumption of the transport tool and the consumption of the storage device). But at the same time there are difficult to manage the problem, that is difficult to know what kind of data in what kind of place. Once the system can work, the standard approach is to first restore the critical application, the rest of the application needs to be restored. In this case, recovery is possible, but it takes some time, and depends on when the hardware platform can be provided ready.
3. PTAM truck transfer mode + hot backup center (PTAM + Hot Site)
Equivalent to Tier 1 plus a disaster center with a capability center. The hot backup center has enough hardware and to support the installation requirements for critical applications. For very critical applications, in the event of a disaster, must be in the off-site with the hardware platform is running to provide support. The way this disaster is backed up relies on the use of PTAM to store daily data in offsite storage, and when the disaster occurs, the data is moved to the center of a hot spare. While moving data to a hot backup center increases the cost, it significantly reduces the time it takes for a disaster to back up.
4. Electronic transmission (Electronic)
Is based on the Tier 2 with electronic links to replace the vehicle for disaster backup. The recipient's hardware platform must be physically separated from the production center, and after the disaster, the stored data is used for disaster backup. As the hot backup center to maintain continuous operation, thus increasing the cost. But it is to eliminate the need for delivery tools to improve the speed of disaster backup.
5. Active Secondary Site (Active Secondary Site)
This disaster backup requires both centers to be active and manage each other's backup data, allowing xp auto backup software actions to take place in either direction. The receiver hardware platform must be physically separated from the other platform, in which case the workload can be shared between the two centers, and the two centers are backed up between each other. Between the two centers, copies of the online key data are kept interoperable with each other. In the event of a disaster, the key data needed can be quickly restored through the network, through the network switch, the critical application of the recovery time can be reduced to the hour.
6. Two-Site Two-Phase Commit
Is managing the selected data on a mirrored state (based on a single commit range, updating the data both locally and in the remote database) on Tier 4, that is, before the update request is considered satisfactory, Tier 5 Data required for production center and backup center are updated. We can imagine a scenario where the data is mapped between the two centers, synchronized by the remote two-phase commit, because the critical application uses double online storage, so when the disaster occurs, only the data in the transmission is lost, The recovery time is reduced to the hour level.
7. Zero data loss (Zero Data Loss)
Can achieve zero data loss rate, while ensuring that the data is automatically and automatically transferred to the backup center. Tier 6 is considered the highest level of disaster backup, while both local and remote data is updated, taking advantage of dual online storage and full network switching capabilities. Tier 6 is the most expensive way in disaster backup, and the fastest way to recover, and the recovery time is reduced to minutes. For Tier 6's disaster backup solution, two remote copy technologies can be implemented, namely, PPRC synchronous remote copy and asynchronous remote copy.
At present, the financial industry to take the bank, some banks need their financial regulatory system has two different locations, 365 days a year to run continuously. One of them is the main data center, for daily operations, and the other for the disaster backup center. The second data center should be kept fully synchronized with the master data center to support disaster recovery. This applies to offsite backup.