問33
頻出度★★★
事業者の基幹となる業務で使用しているシステムにおける障害対策の考え方のうち, 最も適切なものはどれか。
平成21年度 春期 ITパスポート試験 問33
テクノロジ系システム構成要素 › システムの評価指標
This question's explanation and your answers so far.
解説
問 33
Incorrectア
The mistake is putting off the report of the fault until after recovery. The judgment that recovery can be done in a short time may itself turn out to be wrong, and during that time the user departments have their work stopped without knowing either the cause or the outlook. In core business operations the related departments have to make decisions such as grasping the scope of impact, switching to alternative means, and contacting customers, so the principle is to report promptly as soon as a fault is detected and to work on recovery in parallel. Delaying the report actually makes the confusion worse.
Incorrectイ
The mistake is restarting the system and resuming operations before the cause has been investigated. Once the system is restarted, the evidence needed to identify the cause, such as logs and the state of memory, is easily lost, and the same fault keeps recurring. Moreover, if the fault has corrupted data or left processing half finished, resuming operations without checking the cause and the scope of impact lets processing go on with wrong data and expands the damage. Early recovery is important, but it must not unconditionally take priority over investigating the cause.
Incorrectウ
The mistake is stipulating that on-site judgment takes top priority over the predetermined response procedures. When a fault occurs there is no time to spare and the person in charge is easily flustered, so response procedures for each anticipated situation (a fault response manual or a contingency plan) should be prepared beforehand and followed, which prevents inconsistent responses, missed reports, and secondary failures. It is indeed necessary to decide in advance whom to ask for a decision when something unanticipated happens (the escalation path), but that too is defined as part of the procedures; making the priority of on-site judgment the rule turns the matter upside down.
- エ
正解 The correct answer is エ. In a core business system (a system that supports the heart of a company's business), no matter how far quality is improved, faults caused by hardware breakdowns, software defects, or disasters cannot be reduced completely to zero. The basic approach to fault countermeasures is therefore to assume that faults are bound to occur and to prepare in advance measures that keep the damage local when they do. Concretely, such mechanisms include redundancy that duplicates servers, disks, and network routes; fault tolerance, which keeps processing running even after the failed part is separated off; and fail-soft, which keeps at least the important operations running even at the cost of reduced performance or functions. Choice エ states exactly that countermeasures should be prepared so that a partial fault does not stop the whole system or make it impossible to resume operations, which matches the design philosophy of not creating a single point of failure (SPOF). The other three are about operational procedures after a fault has occurred, and each of them breaks one of the basics — prompt reporting, investigating the cause, and following the predetermined procedures — so they are inappropriate.
ここがポイント
- Fault countermeasures for a core business system are based on the premise that faults cannot be reduced to zero, and the basic idea is to prepare so that a partial fault does not spread into a total shutdown. Representative approaches include eliminating single points of failure (SPOF) through redundancy, fault tolerance, which separates the failed part and keeps processing going, and fail-soft, which keeps important operations running even with degraded functions.
- There are three principles for operations when a fault occurs: (1) report promptly to the related departments, (2) recover after checking the scope of impact (such as data corruption) and securing the evidence, such as logs, needed to investigate the cause, and (3) follow the response procedures defined in advance. Any choice that breaks these on the grounds of early recovery or on-site discretion can be judged incorrect.
用語