Read a sample: 1. A Framework for Recovery Decisions
At 8:15 on a Monday morning, a regional distributor loses access to its order-processing application. The warehouse can still pick products, customer-service representatives can still answer calls, and the payment provider reports normal service. Yet new orders are not reaching the warehouse, and staff cannot see whether several urgent orders were already entered.
The first useful question is not “Which server failed?” It is “What does the business need to keep doing while the application is unavailable?” The technical cause matters, but the answer must ultimately support an operational decision: what to preserve, what to restore, how quickly to act, and what risks are acceptable along the way.
Continuity and recovery questions often contain several plausible answers. A backup may exist, an alternate system may be available, and a manual workaround may be possible. Each addresses a different part of the problem. The strongest choice is the one that fits the stated impact, requirements, and constraints—not simply the one that sounds most technical or comprehensive.
This chapter introduces a repeatable way to make that choice. The framework is useful both for exam-style scenarios and for the realistic messiness of an actual disruption, where the facts may be incomplete and the first plan may not work.
Business continuity concerns keeping essential operations functioning during a disruption. That might mean using a redundant system, moving work to another location, changing staff assignments, or temporarily recording transactions on paper.
Disaster recovery focuses on restoring technology services and data after a disruptive event. It can include rebuilding systems, restoring backups, validating data, and returning applications to an operational state.
These activities overlap. A successful failover, for example, can help maintain an essential service and may also be part of a recovery strategy. The distinction is most useful when it clarifies the immediate objective:
- Continuity question: How can the organization continue the essential work?
- Recovery question: How can the organization restore the affected service and its data?
- Combined question: How can the organization operate acceptably until restoration is complete?
In the distributor’s outage, taking orders through an approved manual process is a continuity measure. Restoring the order application and validating its database are recovery measures. Neither automatically replaces the other. A restored application may arrive too late for a critical customer request, while a manual process may become unmanageable if the outage lasts for days.
A scenario may ask for the first action, the best long-term strategy, or the most appropriate way to meet a specific objective. Those are different questions. Before choosing an answer, identify which one you have been asked.
Use the following five steps to organize the facts. They are not a rigid incident-response procedure; they are a way to reason from business need to a defensible choice.
Name the service that users cannot rely on, then identify the business process it supports. “The database is down” describes a technical condition. “Customer orders cannot be accepted or checked for duplicates” describes an operational impact.
Look beyond the component named in the scenario. An application may depend on authentication, network connectivity, encryption keys, a database, or a third-party service. Restoring the visible application alone may not restore the business capability.
At the same time, do not assume that every dependency has failed. If the scenario states that the network is functioning, avoid choosing a network replacement merely because networks are important in general. Separate known facts from possibilities that require investigation.
Ask how long the essential service can be unavailable and how much data loss the organization can tolerate. These are related but distinct requirements.
The recovery time objective, or RTO, is the target duration for restoring a service after disruption. The recovery point objective, or RPO, expresses the maximum acceptable age of data at the recovery point—in practical terms, the amount of recent data loss the organization is prepared to tolerate.
For example, an RTO of four hours addresses restoration time. An RPO of one hour addresses the acceptable gap between the recovered data and the point of disruption. Restoring service in two hours does not satisfy a one-hour RPO if the recovered data is a day old. Conversely, recovering current data does not meet a four-hour RTO if the service remains unavailable for two days.
These objectives are targets, not proof that a particular technology can meet them. A plan must account for the work needed to locate, restore, connect, and validate the service. Chapter 2 examines recovery objectives in greater depth.
A technically attractive option may be unusable because the organization lacks staff, access, time, connectivity, or an authorized way to handle sensitive information. Consider both positive constraints, such as a stated budget limit, and missing prerequisites, such as an encryption key needed to read a backup.
Useful questions include:
- What systems, people, facilities, and outside services must be available?
- How much capacity does the temporary or restored solution need?
- Who can authorize the action, and who can perform it?
- Are the required data, credentials, equipment, and communications accessible?
- What security and accuracy risks does the workaround introduce?
Avoid filling gaps with optimistic assumptions. A backup stored elsewhere is not necessarily accessible; an alternate site is not necessarily ready; a cloud service is not necessarily independent of the same identity provider that has failed.
Start with the requirement the question emphasizes. If the immediate problem is that essential calls must still be…