Whether an organizations need to implement off-site data storage is driven by regulations or a healthy sense of self-preservation, many IT managers are finding that traditional tape backup (with off-site tape storage) is not enough. Fortunately, there are several solutions—both low- and high-end—that can help IT managers move and synchronize data between primary and backup sites.
While synchronous mirroring solutions are still extraordinarily expensive (especially from a bandwidth cost perspective), there are a wide range of products available that can take the data you create at your primary site and intelligently ship it to a secondary location. Because they leverage the WAN, many of these solutions can work over the suggested 200- to 300-mile distance.
Host-based replication products, such as NSI Softwares Double-Take, are software-based solutions that run on servers and allow them to asynchronously replicate data to backup servers over an IP link.
One advantage of a solution such as Double-Take is that it is affordable: Because it uses host resources, customers dont have to invest in additional hardware, and the software has a starting price of only $2,495. However, because Double-Take uses host resources, it could create a fair amount of overhead in a highly transaction-intensive environment.
Moving up the price ladder are array-based replication products such as EMC Corp.s SRDF (Symmetrix Remote Data Facility) and Hitachi Data Systems TrueCopy. These products replicate data between arrays without interfering with the host.
However, one of the biggest reasons array-based systems are expensive compared with host-based systems is that vendors of array-based products do not allow mixing and matching of hardware in their replication solutions. For example, EMCs SRDF requires that the primary and secondary sites have Symmetrix arrays, and IT managers dont have the option of using lower-end arrays as their replication targets at backup sites.
Companies that do not want to get locked into a proprietary solution should consider a new class of network-based replication that has emerged during the past few years: hardware appliances that sit on the network between the primary and backup sites. (For more on such a system, see story on Kashya Inc.s KBX4000.)
The primary advantage of these systems is that they are vendor-agnostic, allowing any-to-any replication. As a result, IT managers can use a network-based replication device to replicate data from an expensive, high-performance Fibre Channel array at a primary site to a low-cost SCSI- or ATA-based array at a backup site.
With this flexibility, IT managers can save a fair amount of money. However, these systems do not have a proven track record, so organizations should tread carefully for now.