Long-distance data replication, once considered prohibitively expensive for most companies, has become more affordable as asynchronous replication technology matures. Disaster recovery sites can now be affordably separated by hundreds of miles, overtaking metro-area disaster recovery data centers -- which are cheaper but conspicuously risky in light of recent disasters and public works failures. Businesses are warming to asynchronous replication, which is good news for the channel since it brings services opportunities.
The benefits of the current generation of long-distance asynchronous data replication technology are considerable: After a period of unacceptable performance degradation levels, it is now bandwidth-friendly and doesn't impact production application performance. IT managers can now decrease network bandwidth costs at the expense of increasing the amount of recently changed data they are willing to lose in a disaster. The service-level metric that defines how much data loss is allowed by the business is called
Value-added resellers (VARs) have real opportunities to help clients navigate the vast array of product selection services and network design offerings. There are three basic classifications your customers might fall into that will guide the choice of data replication product. Just understanding the pros and cons of each of the following types of products will take a huge load off your customer:
- Array-based replication, suited for sites that have a homogeneous storage array and feature-rich external storage containing data to be replicated.
- Host-based replication, for sites with a colorful mix of storage arrays or data residing on local storage.
- Network-based replication, for sites at which neither the storage nor the hosts can be made to support a single replication mechanism.
Business impact analysis
Customers face tough challenges in determining how much protection each application needs. VARs should make clients aware that the cost of availability has to be weighed against the business impact of an outage. Clients with no disaster protection today and those who are integrating asynchronous replication into their existing DR plans should create or at least review their application business impact studies. Solution integrators can add value by lending their objective perspectives and expertise to the client's business impact analysis.
Sizing WAN speeds vs. RPO
Asynchronous replication gives customers a powerful set of choices. If they increase the bandwidth available for asynchronous replication, generally they will also reduce the amount of data lost, improving the application's RPO. The inverse is also true -- reducing the bandwidth available for replication reduces costs but also increases the amount of data lost, degrading the RPO. Since WANs often involve complicated contracts, changes to circuit bandwidth should be made with confidence. Integrators can lend their expertise to determine the amount of bandwidth required to meet a specific RPO.
Network design and circuit sales
Clients may have a lot of experience configuring LAN networks, but WAN connectivity is a whole different matter. It takes a lot of experience -- which customers may not have -- to navigate the network vendor selection process.
Most customers cannot afford the luxury of dedicated WANs for data replication. Instead, the network link that connects the data centers carries traffic of all types. Without quality of service (QOS) functionality in the network, all traffic is of equal priority. No matter whether the customer considers replication traffic more or less important than the CEO's email, integrators can pitch by implementing this critical service.
Asynchronous data replication is making regional disaster protection possible for organizations of all sizes. The expertise required to get through all the decisions and implementations is significant. These challenges are merely disguised opportunities for the true value added integrator.
This was first published in November 2007