Creativeapril - Fotolia
The disaster recovery strategy at Sunkist Growers Inc. looked the same for many years: taped backups, backup servers and networking equipment sitting in rented space at a data center in Arizona.
This setup enabled the not-for-profit marketing cooperative, based in Valencia, Calif., to fail over 75% of what it had in production.
The system meant regular hardware upgrades and ongoing data center fees -- costs that escalated over the years. The disaster recovery plan also required Sunkist to maintain a level of DR expertise among its 15-member IT staff -- with those trained workers having to travel to Arizona in the aftermath of an incident.
"We were looking at a minimum of eight hours to even start on a recovery -- that's assuming that we could fly or drive out," Sunkist Network Infrastructure Manager Don Conley said.
Conley was left concerned about the pressure put on staff to leave their families in the aftermath of a major earthquake or other natural disaster, especially when travel conditions could be unsafe.
"I'm not going to make them choose work over their families, but at the same time, I have a responsibility to get the business up and running," he said. "For me, it was a personnel issue. I wanted someone who could do most of the disaster recovery so my team could worry about their families and their homes and not have to worry about the company."
A new DR approach
Don Conleynetwork infrastructure manager, Sunkist Growers
In late 2017, Sunkist chose Key Information Systems Inc., which offers advanced networking and hybrid cloud products, to take advantage of their disaster recovery as a service (DRaaS) and backup and recovery as a service (BRaaS) solutions.
Conley opted for the managed DRaaS option because it provided a cost-effective way to ensure that Sunkist could continue operations without overburdening or jeopardizing the safety of his staffers and their families.
"It gives us some flexibility but still allows our company to be up and running in case a major disaster hits," he said.
Conley previously ruled out DRaaS because of cost considerations but said the price of the solution had recently become more competitive and made it a better option than Sunkist's long-standing practice where it managed disaster recovery infrastructure on its own.
Key Info worked with Sunkist to review its environment and requirements, mapping out applications and dependencies, and considering networking design and routing protocols. They then put together recovery scripts to automate the process, focusing on business-critical applications.
"[Key Info] is able to do 98% of the disaster recovery through a scripted system, so we can get the systems up -- some of them running within four hours and others within 24 hours -- so we can be functional within one day," Conley said.
Beyond Sunkist DRaaS
The disaster-recovery-as-a-service market has matured quickly, said Mark Jaggers, Gartner research director. Gartner puts the DRaaS market at $2.40 billion worldwide and estimates it will hit $3.73 billion by 2021.
He said larger organizations generally have the personnel capabilities and infrastructure to develop their own DR capabilities at a cost comparable to DRaaS offerings.
"Smaller and midsize companies don't have that same scale, so having that third-party supplier to help them is an easy victory," Jaggers said.
That doesn't mean vendor-managed disaster recovery is the right option for the small or midsize organization. A smaller company with multiple locations and completely immersed in the cloud, or even those building out their IT departments on IaaS, probably won't need DRaaS solutions that a larger organization like Sunkist would, Jaggers said.
Many companies continue to struggle with how best to plan for disasters and their aftermath, and they sometimes overlook how an incident will impact not only their facilities and infrastructure, but the utilities they rely on and also their employees.
"That's a fairly common subject that comes up, not just in terms of IT recovery, but overall business recovery: the question of whether you will have staff available," Jaggers said. "You can't count on having a subject matter expert on a component of the infrastructure being available in times of a disaster."
Meanwhile, Jaggers added that some organizations that might be good candidates for the DRaaS option based on their technology stack opt against it because they're not comfortable outsourcing the function for a variety of reasons. Similarly, organizations committed to their mainframes and other legacy systems won't see strong incentives to move to DRaaS.
"Your disaster recovery experience is really reflective of what your IT infrastructure looks like," Jaggers said, noting that hybrid environments and the complexity of modern IT departments have similarly complicated DR strategies.