How to do well in disaster recovery of information

2022-08-04
  • Detail

How to do a good job of information system disaster recovery

the information system disaster recovery specification puts forward the basic requirements that China's information system disaster recovery should follow. On the premise of meeting this requirement, enterprises can have a variety of methods and means to implement disaster recovery. Now let the experts teach you some tips

three data centers' disaster recovery ensures the highest business continuity

problems: users adopt a large centralized business model, which not only has a large amount of business, but also cross regions, and puts forward higher requirements for data security and business continuity

with the increasing requirements for the reliability of the disaster recovery system, the traditional disaster recovery schemes of the two data centers can no longer meet the requirements. By integrating synchronous replication technology and asynchronous replication technology based on two data centers, storage manufacturers have launched the three data center disaster recovery architecture with the highest safety factor at present. According to the different technologies used, it is divided into three implementation methods

three implementation modes

cascade mode is the most basic and earliest mode. A synchronously copies the data to B, and B asynchronously copies the data to C. If there is a problem with site B, switch between site a and site C. This disaster recovery method can not fully guarantee the real-time continuity of business

a synchronously replicates data to B, and asynchronously replicates data to C. There is no connection between B and C. once a disaster occurs, B and C need to re-establish the connection, resulting in a large amount of time for initial data replication between B and C, and the requirement for data transmission link bandwidth is also very large

a synchronously replicates data to B and asynchronously replicates data to c. a standby link is added between B and C through asynchronous replication

in case of a disaster, the backup link between B and C should be started. B only needs to copy the differential data to C, without data initialization replication. This method can synchronize the data between B and C in a short time, and realize the disaster recovery system between B and C in a short time. Hitachi Data System first introduced this technology a year ago. The key point is that it realizes the differential backup between B and C

in this way, any two sites can back up each other for disaster recovery without data loss. Therefore, it realizes the full disaster recovery of the three sites in the true sense and ensures 100% business continuity for users. It is the highest level of disaster recovery scheme at present

China International Electronic Commerce Center (CIECC) has adopted the third data disaster recovery scheme to realize the disaster recovery of the three data centers in the real sense and ensure 100% business continuity

"pull in" asynchronous replication technology

hds delta resync three data centers adopt Hitachi universal replicator (HuR) for disaster recovery. The perfect combination of local and cross local replication is very economical and practical for more enterprises and applications

hur can asynchronously replicate data on USP or other manufacturers' storage systems externally connected to USP. Its uniqueness and progressiveness lies in that it does not push the data from the production center to the remote backup center, but pulls it to the disaster recovery center

this disaster recovery method based on HuR remote asynchronous replication technology is called "pull disaster recovery". In this way, all performance losses are concentrated at the end of the disaster recovery center, which will not affect the performance of the production center. If disaster recovery is done in the traditional way, since the initiative of data replication is controlled by the equipment in the production center, a large number of resources of the storage system in the production center will be consumed, resulting in the reduction of the performance of the production center

hur is a unique innovation that goes beyond the traditional replication function. It redefines the execution mode of asynchronous replication

hur not only greatly reduces resource consumption, but also establishes a new standard for data protection. By adopting the "pull disaster recovery" technology based on USP, HDS's three data centers multi target with delta resync disaster recovery scheme can guarantee the performance of the user's production center to the greatest extent for users, truly realize the full disaster recovery between the three data centers, achieve 100% business continuity for users, and greatly improve the operation efficiency of their IT systems

using VTL to realize intelligent disaster recovery

problems: transmitting a large amount of data between the active and standby centers not only consumes a large amount of network bandwidth, affects the transmission speed, but also increases the cost of disaster recovery

there is also a very prominent problem in the disaster recovery system, that is, the amount of data transmitted is large. Large scale data transmission over the network will seriously affect the timeliness of the system. To solve this problem, sepaton's virtual tape library incorporates replica deletion technology. The main idea of replica deletion technology is to delete redundant data among existing data, so as to achieve the effect of data compression and reduce the amount of data transmitted. Sepaton's replica deletion technology can achieve a minimum 25:1 data deduplication ratio for mixed data. This can effectively reduce the data backup window and meet the data transmission requirements of RPO and RTO

The principle of

replica deletion technology is that the data entering the system for the first time will be scanned and the database will be established; When new data enters the system again, the system will compare the same points of old and new data by bytes, mark those same data with pointers, and delete those same data after confirmation. The replica deletion technology can restore in the case of full backup, incremental backup and differential backup. Its method is to directly locate the source data according to the pointer, so as to ensure the maximum recovery performance. The processed data can be used for the transmission of sepaton site2. At this time, it is not necessary to locate the data subject, just pass the pointer. This can save the data transmitted on the network, thus optimizing the bandwidth and shortening the time. The replica deletion of sepaton is performed after the backup and within the virtual tape library, so it does not rely on any third-party software resources

adopting the virtual tape library solution can not only greatly reduce the hardware procurement cost, but also save storage space and energy consumption by adopting the disk to disk (D2D) method, which can be described as green

integrating heterogeneous storage hardware with virtualization

facing problems: there is already a complete set of storage infrastructure, and there are multiple disk arrays of different brands, which increases the complexity and cost of management

at present, many users have built a centralized data storage platform based on network storage technology and integrated all business data into this centralized storage platform. However, due to various reasons, the back-end of the centralized storage platform is composed of disk arrays from multiple manufacturers. This heterogeneous disk array environment brings inconvenience to users

it is difficult to manage heterogeneous environments

disk arrays from different manufacturers need to be managed and used using their own management tools, so the management and use of the entire storage platform is very complex. Disk arrays from different manufacturers cannot replicate array based data with each other, so it is impossible to protect array based data between them. Disk arrays from different manufacturers have compatibility problems in host access, especially in multipath control. This restriction makes a set of hosts can only use the disk array of a certain manufacturer, and the storage space can not be maximized

if users need remote disaster recovery, from the perspective of management and cost, they can only adopt the construction of host based disaster recovery system, such as the TDPs technology of NetApp

storage virtualization is a good recipe

heterogeneous storage environment has brought great problems to storage use management, data protection, disaster recovery system construction and so on. So how to solve these problems

the simplest way is to use the disk array of the same manufacturer to replace the storage disk array of other manufacturers, so as to achieve a homogeneous storage environment. Although this method is simple, it is very wasteful. A large number of original investments of users cannot be protected. At the same time, users' choices are limited to a certain manufacturer, which will indirectly lead to an increase in subsequent procurement costs

a reasonable solution is storage virtualization. NetApp's V-series storage virtualization products can help users solve this problem

disk arrays from different manufacturers and those from other manufacturers that may be added in the future can be directly connected (or through San) to NetApp V series storage virtualization devices, which are connected to the user's original San network

for the user's original heterogeneous disk array, after the virtualization shielding is performed through the NetApp V series storage virtualization device, the front-end host can use and manage the back-end disk array according to the unified interface and interface provided by NetApp, regardless of which manufacturer the back-end disk array comes from. In this way, it seems that the entire storage subsystem comes from one manufacturer (NetApp), which greatly simplifies the use and management of the entire storage subsystem, reduces the management and use costs, and improves the reliability of the system

because NetApp's V-series storage virtualization device, like NetApp's primary storage series FAS, adopts NetApp's patented dataontap operating system, which has the same functions as the primary storage. Therefore, after the device virtualization operation, those users' original heterogeneous disk arrays can fully utilize NetApp's advanced storage functions, such as WAFL, flexvol, snapshot, snaprestore, etc. The use of these functions, on the one hand, can improve the performance of data access, simplify the management of storage space, and improve the utilization of storage space; On the other hand, it improves the ability of data protection and greatly improves the security and reliability of data

generally speaking, users' original heterogeneous disk arrays can only support SAN storage access architecture. After being virtualized by NetApp V storage virtualization device, since the virtualization device supports NetApp patented san/iscsi/nas integrated support technology, these disk arrays from different manufacturers can simultaneously support three architectures to meet the storage requirements of different applications

by using NetApp's virtualization devices, the user's entire storage environment logically looks like a set of NetApp's Fas series disk arrays at the back end. Therefore, if users use NetApp's V series storage virtualization devices to complete the virtualization of heterogeneous storage environments, they can fully use snapmirror technology to achieve remote disaster recovery based on storage system data

a set of NetApp Fas series disk array is configured in the disaster recovery center, and the remote replication of data is realized between the V series virtualization devices in the production center and the NetApp disk array in the disaster recovery center by using snapmirror, thus realizing the remote disaster recovery of data based on the storage system. After virtualization with V-series storage virtualization devices, you can also use snapmirror's many to one data remote replication capability to achieve a disaster recovery center to provide disaster recovery services for multiple customers at the same time

cdp bypass scheme achieves minute level recovery

problems: some large and medium-sized users have high requirements for the continuity of their key businesses and hope to maintain 7 × 24-hour continuous operation

generally, the prevention of gradual disasters mostly depends on the backup system, so this kind of disaster recovery system does not have the ability to prevent a full series of disasters (including natural disasters, etc.).

Copyright © 2011 JIN SHI