In the information systems in general become the lifeblood of business operations today, by constructing a remote disaster recovery mechanisms to ensure business continuity, corporate managers have been generally recognized.However, many types of remote disaster recovery technology, and the most complex structure, not only into hardware and software equipment and the huge time cost, ongoing maintenance costs were prohibitively high, so, despite the large number of interested builders, but most hesitate.
To tape-based backup for remote disaster recovery, is widely used means business, managers need to regularly back up your data off the tape, and then transported to the remote through the preservation of human can, if there is a disaster, should firstIT infrastructure repair, and then restore from tape systems and data.This approach, while the remote disaster recovery of the lowest cost, but the lengthy recovery time (usually several days or even weeks) caused by operating losses, and backup / recovery time caused by the non-real-time data loss, it is clear beyondthe scope of today's enterprise can bear.Ideal for remote disaster recovery framework should be based on unlimited wide-area network based on remote data replication (Remote Replication) technology, the online (on-line) off-site disaster recovery, so as to meet the business requirements of RTO and RPO.
Through the network connection, the local copy to a remote data preservation, sounds easy, but in a complex IT architecture, in order to ensure successful implementation is not easy.After several years of technical development, online remote disaster recovery to produce the following two types: host-type (Host based) and the storage system type (Storage based).
Host-based remote disaster recovery: the limitations of large
Host-based remote disaster recovery, simply, is installed on the server through the data replication software (such as VerITas Volume Replicator), or applications of data replication / disaster recovery tools (such as Oracle Flash Recovery Area), using TCP / IPthe capacity of the network connect to a remote standby server to asynchronous data replication.
Host-based remote disaster recovery has the advantage of less server environment, low cost, users need to replace the existing system architecture too much, do not worry about the compatibility of the back-end storage systems, only paydisaster recovery software licensing fees and hardware costs can end.However, if a large number of the server environment, management will increase the complexity of the overall input costs will increase.Another drawback is that it is in the application software is installed on the host, the host running the operation will consume resources, if the hardware level is not high, the application can Nenggei impact.
Long-range disaster recovery storage systems: the high cost
Long-range disaster recovery storage systems, by definition is based on the storage system (optical disk array, NAS) model.Built through the storage system firmware (firmware) or operating system, through the IP network, or DWDM, Fibre Channel Interface link to synchronous or asynchronous data replication to a remote way.Well-known long-range disaster recovery plan storage systems have SRDF, TrueCopy, PPRC and so on.
Compared with the host-based remote disaster recovery, storage systems, the advantages of long-range disaster recovery operation is to separate data on the operation of the system resources of the host is relatively small.In addition, most of the operating mechanism is the use of the mirror (mirror) to copy the data and help speed up the cache I / O access, both ends of the different time points of data, coupled with the storage system itself has a certain degree of fault tolerance, with a certainoperational performance and reliability.
The long-range disaster recovery storage systems, the biggest constraint on the construction of its expensive cost.The user must end in the local side and disaster recovery are two sets of the same storage system configuration, not only the procurement costs are high, but also subject to a single device manufacturers, the future expansion of the bound and inflexible.In addition, the Fibre Channel storage system if you want to build a remote disaster recovery, and disaster recovery must be in the local end of the installation of a FC-to-IP adapter, the hardware cost will be more than 50,000 U.S. dollars, plus the cost of network bandwidth, thenThe overall cost investment will be staggering.If the enterprise has not been carefully evaluated prior to installation, then build a remote disaster recovery storage systems based IT spending is likely to cause a black hole, increasing the financial burden.In addition, the storage system-based disaster recovery disaster recovery mode for the consistency of the database there is a big flaw in the disaster recovery a little more structure to the existence of non-applicability.
In addition to these two architectures, the companies do no better choice? No! Recent years, the United States, FalconStor Software, Inc. (FalconStor), represented by the network storage software solutions, successfully developed the storage networkOff (Storage Gateway) as the core of the storage network-based (Storage Network based) remote disaster recovery scheme, which combines not only host-based long-range disaster recovery and storage system architecture advantages, but also breaking the construction cost remains a remote disaster recoveryno less than the ice, becoming the most popular at this stage remote disaster recovery technology.
Long-range disaster recovery storage network architecture, is the front-end application servers and back-end storage systems, storage area network between (SAN), adding a layer of storage gateway, the gateway and network gateway as we know it is different from the United StatesFalconStor Software's solutions, for example, combines the IPStor storage management software and dedicated manager, front-end server host connections, back-end-attached storage devices, and its role in the storage network is like the traffic police, all the I /O is controlled by its management.
Storage network-based remote disaster recovery: inclusive
Or in FalconStor's IPStor as an example, IPStor remote disaster recovery operation mode, a combination of host-side agent (Agent) and snapshot function (TimeMark), creating a snapshot copy of host data, copied over the network to a remote disasterPreparation of the site.IPStor is the only change of use of the data model for data replication, therefore, the lower the bandwidth required, and the local link between the disaster recovery side, you can use relatively inexpensive IP networks.
Storage network-based remote disaster recovery advantages can be said is a culmination of the above two architectures.Since the data replication is performed by storage gateway, application server, database implementation agent only, as opposed to host-based remote disaster recovery for its performance impact is very low.In addition, through the storage virtualization gateway technology, integration of front-end servers and back-end heterogeneous platforms with different brands of storage devices, local and disaster do not need to pair the device side configuration, the user can RTO and RPO, in the remoteestablishment of a complete hot backup center, when the local disaster took over the business operations immediately; or to end only in disaster recovery installation of storage temperature and station configuration, the first protection of data integrity and security of the repair is completed locallyfurther recovery.
In addition to the above do not take the host operating resources, and do not have the limitations of storage platform characteristics, the cost of storage networks is the biggest advantage of long-range disaster recovery.
First of all, when you need to replace the original building IT infrastructure, only in the original by adding storage storage area network gateways, local hosts and storage devices can be any brand, the disaster recovery side of the host and storage devices are not required andthe same local user can even use lower grade side in the disaster recovery storage systems (such as SATA disk array), according to statistics, the investment costs can save up to 30-50%! for those who are keen on building a remote disaster recovery needsThe IT budgets of the customers are very limited, long-range disaster recovery storage network is the best choice.
Second, special agent for the database, storage, network-based remote disaster recovery to ensure that the database and start with a complete disaster recovery capabilities, without worrying phenomenon not start.
More importantly, the data on the storage of more value-added features such as snapshots, can make all kinds of data can be instantly vandalism ability to recover, it is one of the most complete disaster recovery system, it coversthe scope of the disaster far more than the aforementioned fight against all kinds of ways.In such a disaster recovery system, the construction of disaster recovery is no longer difficult, disaster is no longer a very terrible thing.