Issue link: https://cp.revolio.com/i/1012043
4 | THE DATA CENTER JOURNAL www.datacenterjournal.com D isasters tend to be rare, however, so companies can easily look back at the months or years of smooth sailing and put off preparation, to their detriment. But the motivation is seldom just laziness or blind trust in an untroubled future. Disaster preparedness and recovery can involve technical challenges that carry high price tags, especially if a company decides to tackle the prob- lem alone. Implementing a data center DR site can involve extra capital costs for real estate, IT equipment and other infrastructure, as well as maintenance and other operational costs. And the technical knowledge to implement an off-site DR scheme comes at a premium that many companies can't afford. But riding the as-a-service wave is disaster recovery (DRaaS). By delivering off-site resources for data center operators through the cloud, DRaaS providers allow customers to outsource much of this onerous task. e question, then, is whether cloud- based DRaaS is truly a match for a purely in-house (or even colocation- based) scheme. is same question has arisen with many other aspects of the cloud, and although reports of the in- house data center's demise are greatly exaggerated, use of cloud services is growing: median operational budgets are rising, whereas median capital budgets are flat, according to Com- puter Economics. CLOUD OR COLO? Colocation offers a ready solution for data center operators that want a DR site without having to build or buy an entire remote data center. It converts many (but not necessarily all) expenses from capital to operational costs. A major challenge remains, however: the technical feat of prepar- ing for and, when necessary, executing the switch from a primary data center to a backup data center. "Building and operating a DIY DR environ- ment requires specialized knowledge and experience that's oen unavail- able in the organization," said Mike Fuhrman, chief product officer for Flexential. "Many times, the required staffing and capital expense associated with colocation-based DR operations prevent the organization from build- ing robust and effective DR sites. On top of this, many organizations are subject to regulations such as PCI DSS or HIPAA, an additional hurdle that's difficult to clear with an in-house DR site." e next natural step is then to outsource the IT side as well—a ser- vice that cloud or hybrid-IT providers can tackle. Fuhrman added, "Hy- brid-IT providers not only have the required secure data center footprint, but they also offer the technology stack and services customers need to successfully consume DR as a service. is kind of DRaaS allows customers to move from a primarily capex-based, DIY model to a completely predict- able opex model." And with the hybrid approach, customers have the option of mixing cloud-based and company- owned equipment to satisfy a broad range of requirements. Although either option eases the burden on a company implementing an off-site DR data center, the cloud in particular offers—at least in theory— the easiest route. Fuhrman noted, that "in-house DR sites require the company to take the lead in operat- ing and maintain all aspects of the DR environment, including space, power, connectivity, storage, compute and network infrastructure—clearly a lot to handle. ose in charge of the site must regularly conduct business-level continuity analysis to meet BCP (busi- ness-continuity plan) requirements, not to mention the process requires extensive planning and specialized knowledge in DR operations." ese requirements and skills mean expenses that companies—par- ticularly those whose focus isn't IT related—may be unable to afford. And even for those that can afford the costs, budgeting for more-central busi- ness activities probably offers greater returns as long as a less expensive DR option (i.e., the cloud) is available. e next natural question: for whom is cloud-based DR best suited? CLOUD-BASED DR: WHY NOT? e relevant question may be less about which companies can benefit from DRaaS and more about which companies can't benefit from DRaaS (perhaps better said, which companies benefit more from an in-house DR site). e costs of building and main- taining a company-owned DR facility, including the requisite personnel and the distraction from the company's primary business, make the in-house approach seem unthinkable when third parties specialize in this task. But as with most rules, exceptions arise. Fuhrman identified at least one exception in particular: companies with a geographically dispersed data center presence. "If companies (e.g., Fortune 100) own and operate their own data centers in various geo- graphic locations (not colocation), they most likely already have the operational efficiencies and required knowledge to successfully build their own DR sites." Although even here DRaaS may remain a profitable alter- native, a geographically diverse data center deployment means the com- pany always has a facility that satisfies one major requirement for a good DR site: it's outside the disaster zone. A complicating factor is latency. e business model of some compa- nies—for example, some financial- services providers—precludes the use of distant data centers. In a disaster, therefore, using a DR site beyond a certain geographical limit could be almost as costly as a complete out- age. In this sense, target market can