O.R. ApplicationsUsing formal MS/OR modeling to support disaster recovery planning
Introduction
A disaster recovery plan (DRP) or disaster recovery strategy (DRS) is a system for internal control and security that focuses on quick restoration of service for critical organizational processes when there are operational failures due to natural or man-made disasters. A DRP aims to minimize potential loss by identifying, prioritizing and safeguarding those organizational assets that are most valuable and that need the most protection. In recent years various factors including government regulations in certain industries (e.g. banking, credit unions), the occurrence of natural disasters (e.g. hurricane Andrew of 1993), and the occurrence of social disasters (e.g. L.A. riots of 1992) have led to increasing interest in the development, testing and maintenance of DRPs. This has resulted in intense customer and vendor interest in the development of tools for disaster recovery (e.g. Sclafane, 1996; Violino, 1996; Tevis, 1996; Carson, 1997). For example in just the NT (operating system) recovery area, it is estimated that customers spent $ 102 million in 1996, with the projected amount being $ 324 million for the year 2000 (Torode, 1997).
Although the field of DRP has attracted the interests of information technology practitioners and vendors for a number of years, very little formal management science type research (e.g. Sarker et al., 1996) has been done in this area. Thus, while there is a large volume of publications on DRP in numerous and various practitioner-oriented journals (e.g. Banking Technology, Network World, Communications News), the subject has received scant attention in the management science (including information systems) research journals. It appears that while much attention has been paid by vendors and organization to the development of hardware and software tools for addressing specific aspects of a disaster (e.g. Dryden, 1997; Bucholz, 1997; Merrill, 1997), little attention has been paid to formal modeling of the DRP process. In the instances where the problem has appeared in the academic journals the focus has been on examining the general features of a disaster recovery plan rather than on decision support. But as noted by Jackson (1997), “Creating a good plan is no easy task”. For DRPs are inherently complex, given the various types of possible interactions that could occur between target resources, solution resources and the environment. Also as observed by Howard (1997), “the danger is not over once the fire has been put out”, so that survival is not just a matter of putting out the fire, and there must be a clear plan for recovery that requires various resources including personnel, hardware and infrastructure.
The dearth of models in the MS/OR literature shows quite clearly that the benefit of MS/OR methodology has not yet been brought to bear on this emerging discipline. Although recently Jenkins (2000) has explored the use of integer programming techniques for selecting disaster scenarios, it would appear that the largest body of relevant quantitative analysis exists primarily in the area of risk assessment (Levitt, 1997; Tamura et al., 2000), or has been reactive in focus, being restricted only to the operational post-disaster phase such as the deployment of crews (Pidd et al., 1996; Sarker et al., 1996). Our research, we believe, will be a welcomed addition to the contribution of MS/OR to disaster recovery research. Recognizing that little modeling of disaster recovery issues has taken place to date in the MS/OR literature, our model represents a forward step in providing what is a fast growing discipline with rigorous models for supporting decision making. Consequently, our contribution is prescriptive rather than descriptive in nature. We introduce the application of MS/OR to DRP at the pre-disaster planning/development phase. MS/OR has potential for contributing to the strategic levels of decision-making in DRP to ensure that plans operate as expected/desired when put into operation. There are opportunities for applying MS/OR at all levels of DRP planning, development, implementation and operation, however we focus on the development phase where MS/OR has been most noticeably absent and can have its greatest impact. In this paper we presents a mixed-integer mathematical decision model for selecting subplans for a DRP. We use randomly generated hypothetical problems to demonstrate the feasibility of MS/OR for decision support in DRP development.
The rest of the paper is organized as follows: Section 2 provides an overview of DRP concepts; Section 3 provides a sample of outstanding DRP research problems; Section 4 provides an OR-based solution approach for addressing one of the research problems described in Section 3; and finally in Section 5, the paper terminates with a brief discussion of the contribution of the paper and possible directions for future research.
Section snippets
Some general axiomatic statements about DRPs
We begin this subsection by listing some general axiomatic statements about DRPs:
- •
Several possible disasters can occur.
- •
Each disaster has a set of possible debilitating effects.
- •
Each effect has the ability to affect a number of business functions.
- •
To protect against each effect, a set of (solution) resources are necessary.
- •
Two strategies are possible: prevention (risk mitigation) and recovery.
- •
Some recovery resources may be substitutable.
- •
Substitutable resources do not necessarily possess the same
A sample of outstanding DRP research problems
In this section we present an overview of three DRP problems that could be suitably addressed by mathematical modeling techniques. In the following section we present a mathematical model for addressing the subplan selection problem.
A mathematical model of the subplan selection problem
In this section, we present a mixed-integer model for the subplan selection problem (i.e. selecting a group of subplans from a number of alternatives in order to establish the disaster recovery capability for the organization).
- K
set of business functions
- R1
set of solution resources that occur in real quantities (e.g., time)
- R2
set of solution resources that occur in integer quantities (e.g., equipment)
- R
set of solution resources: R=R1∪R2
- J
set of disaster effects
- I
set of disaster types
- S
set of recovery
Notation
Conclusion
In this paper, we have shown how MS/OR modeling can be used to provide strategic decision support for DRP selection. We have presented a new model for selecting subplans in building a DRP/DRS that addresses the concerns of feasibility, consistency and completeness. The objective of the model is to maximize the total value of the coverage provided by the set of selected subplans subject to several constraints. The model, which contains, real, binary, and general integer variables can be solved
References (16)
Selecting scenarios for environmental disaster planning
European Journal of Operational Research
(2000)- et al.
A simulation study for emergency evacuation
European Journal of Operational Research
(1996) - et al.
Power restoration in emergency situations
Computers and Industrial Engineering
(1996) - et al.
Modeling and analysis of decision making problem for mitigating natural disaster risks
European Journal of Operational Research
(2000) Safeguarding data from disaster
Telephony
(1997)Ten vital steps to save your C/S system from disaster
Datamation
(1997)- CPLEX Callable Library, CPLEX Optimization Inc., Incline Village, NV,...
Cheyenne rides to windows NT's rescue
Computerworld
(1997)
Cited by (95)
Risk and resilience-based optimal post-disruption restoration for critical infrastructures under uncertainty
2022, European Journal of Operational ResearchTree optimization based heuristics and metaheuristics in network construction problems
2021, Computers and Operations ResearchCitation Excerpt :Much research work in this area falls under the network design paradigm, addressing the questions of what capacities/edges/facilities to add to the network to ensure effective restoration of functionality in case of some limited destruction (e.g., Ambs et al., 2000; Balakrishnan et al., 2001; Balakrishnan et al., 2002), or which temporary arcs to add to restore service after a disaster (e.g., Lee et al., 2007). Other streams of research address optimizing the choice of subplans in developing effective disaster recovery plan for an organization (e.g., Bryson et al., 2002) and network vulnerability assessment (e.g., Murray et al., 2007). We refer the reader to Matisziw et al. (2010) for an overview of research directions in network restoration.
A two-stage approach to agile pharmaceutical supply chain management with product substitutability in crises
2019, Computers and Chemical EngineeringCitation Excerpt :Like many researchers, El-Agizy also considered the demand uncertainty and looked for the minimum cost. The innovation by Bryson et al. (2002) was the investigation of both natural and human disasters in shaping the issue under study. These crises included natural crises like drought, famine, disease outbreaks, extreme temperatures, floods, earthquakes and ferocious fires, and also human disasters like industrial accidents, road accidents, terrorism, political disasters and crime.
Optimum post-disruption restoration under uncertainty for enhancing critical infrastructure resilience
2019, Reliability Engineering and System SafetyCitation Excerpt :Various studies have been proposed in the literature in the context of post-disruption CI restoration into a mathematical programming framework. Bryson et al. [17] applied a mixed integer programming (MIP) approach for selecting a set of recovery subplans giving the greatest benefit to business operation. Casari and Wilkie [18] discussed restoration when multiple infrastructures, operated by different firms, are involved.
Resource allocation when planning for simultaneous disasters
2019, European Journal of Operational ResearchA decision support methodology for a disaster-caused business continuity management
2019, Decision Support Systems