Disasters in the oil and gas industry that impact business operations, such as drilling platforms and gas pipelines, are often nature-related. Recent hurricanes, floods and forest fire bring to mind this fact. Accidents from leaks, spills and explosions contribute to the constant threat of disaster. Combined, these threats contribute greatly to the industry’s ongoing advances toward reducing risk and managing through the recovery process when disaster strikes.

But a particular aspect of business continuity planning (BCP) has been getting more attention

Figure 1. The data center should include necessary space, security, power, climate, fire suppression, and flexible and affordable network access. (Image courtesy of Alpheus Data Services)
recently — the impact of disaster on data operations. Like traditional revenue-generating operations, data operations are vulnerable to natural disaster and human error with equally devastating results for the enterprise. Disaster recovery planning (DRP) for information technology (IT) has become mission critical, and IT departments are being tasked with developing increasingly more resilient DR solutions to mitigate risk.

DRP for IT is the process of restoring access to data, software and systems required to maintain critical business operations after an event. This is complicated by rapidly evolving requirements driven by data growth, regulatory concerns and a need to access data ever more quickly. Data restoration must also be prioritized relative to recovery time objective (RTO), the acceptable time interval required to restore access to the data, and recovery point objective (RPO), the acceptable point of data roll-back.

All this demands a continuous and integrated process of planning and implementation aimed at prevention and mitigation, preparedness, response, and recovery from disaster. With increased spending on DRP the demand for additional data center space has increased. This article considers the growing use of third-party facilities, or “managed data centers”
to address DRP demands.

Growth in managed centers
Unfortunately, the recent data center boom has created significant backlogs for fundamental data center resources. For example, many new deployments have been delayed due to a recent shortage of generators for emergency power. Demand drivers include the phenomenal growth in Internet content hosting facilities, the growing attention to enterprise business continuity planning, and mounting power and cooling issues for many existing single-user data centers. A growing number of facilities operated by data center service providers offer an attractive alternative to in-sourcing the construction or expansion of company-owned facilities and are playing an increasing role in BCP.

So what exactly should one consider in evaluating a managed data center alternative? Managed data centers, like all enterprise data centers, are first and foremost charged with addressing the environmental needs of the data operations and IT staff working at the center. However, unlike a typical “primary” data center located near corporate headquarters (sometimes within corporate headquarters), a DR site placement benefits from geographic separation from headquarters and beyond the impact of weather and other factors that directly impact the primary data center.

Therefore, the ideal DR site must include adequate rack space for immediate and long-term requirements, a high degree of security, properly conditioned power with adequate battery and generator back-up, redundant cooling and redundant fire suppression. The ideal space should also be geographically separate. Network requirements are often an afterthought, but proper planning dictates that network connectivity options be included in any site selection process. It makes little sense to provide redundant data if the network connectivity to users is “down.”

Network requirements
A network service designed for a remote DR site typically includes a redundant, high capacity data connection between the primary data center and the DR site. This critical element is used to back-up or even “mirror” corporate data on a periodic or real-time basis. Additional network requirements might also include redundant Internet access and redundant wide area network connections from the DR site to support a regional, national or global enterprise operation during an event.

Note that emerging packet-based network technologies like MPLS allow some service providers to simplify redundant network design through the use of virtual circuit technology. An advanced design can help contain costs and improve the network reliability and availability with service level agreements defining minimum levels of availability and performance.

Financial considerations
A DRP process should begin by calculating risk. What is the likely long term financial impact of an indefinitely unavailable data system? Based on this determination, management should allocate a budget for its specific DR needs, as no two companies are alike. For example, what data storage capacity, RTO and RPO are sought for each mission critical data application? These can vary considerably even between enterprises of similar size.

Ultimately, each DRP is designed from the ground up; the critical infrastructure components of space, power and network often accounting for a majority of the overall DRP cost. These factors serve to frame the budget, while cost containment considerations impact the resulting site selection decisions. Again, managed data centers that control costs and deliver superior service provide valuable alternatives to building or buying.

El Paso’s disaster recovery solution
El Paso Natural Gas Corp. owns the continent’s largest pipeline system and is one of North America’s largest independent natural gas producers.

Considering that El Paso’s dependability in the marketplace is significantly tied to IT dependability, its internal IT disaster recovery site selection recommendations were closely scrutinized. This rigorous process illustrates the critical criteria that any energy company should evaluate in solving their own disaster recovery site selection.

Problem. In 2005, the company had concerns about its backup data center located in San Antonio, Texas. With the lease space renewal pending, the company concluded this site no longer met necessary space and environmental requirements for its critical disaster recovery needs. As its needs spiraled, rack space was becoming limited and the other building tenant (an important El Paso operating unit) was also placing more demand on space. It had become apparent that the San Antonio site might need to be abandoned and that several alternative DR site options should be sought out and evaluated.

Solution. The company compiled critical, high priority attributes for selecting a new mission-critical disaster recovery site. These included cost containment, own-versus-lease, site readiness, network connectivity, remoteness from primary site weather conditions (Houston-based companies face severe threats during hurricane season that impact the entire Gulf Coast region), proximity to the center for ease of travel, staff relocation impact, high/low rise considerations, and environmental factors including space, power, fire suppression and security.

An inland data center in either Austin, Texas, or Dallas, Texas, with completely redundant fiber network capabilities back to Houston made the most sense. The Alpheus Data Services managed data center in Austin was selected primarily because of its geographical, data center, and network services attributes: 2000AMP@480 volts AC power, Alpheus Engineering & Construction team, readily available CRAC units, standby power system, preaction fire suppression system, Alpheus-owned regional fiber network, security fencing, ample secure parking, and expansion capabilities. By strategically planning ahead for critical data back-up, El Paso achieved its objective of being well-positioned for safety and security of all its valuable data.

Comments