How to persuade CEOs to adopt Fault Tolerance

Some CIOs know that fault-tolerant systems have higher reliability than clusters, but what ccios are embarrassed about is how to convince CEOs: A Fault-Tolerant server costs about more than 0.2 million yuan, and a cluster server costs about 20 thousand yuan, however, it is not easy to convince the CEO that the two servers cost more than 40 thousand yuan.

From the perspective of ROI), the fault-tolerant Server is a system, and only one authorization is required for the corresponding operating system, database, and application software. In comparison, the hardware cost of the cluster system is low, but the software cost is high. You need to buy one more software license. In general, fault tolerance is dominant. However, in domestic environments, some vendors tend to give away software through hardware or open-source software, which weakens the overall price advantage of fault tolerance. The problem with using open-source software is that technical support often cannot keep up.

In addition to investment, reliability differences are even more important. For some key business applications, such as steel mill MES Systems (Manufacturing Execution Systems), it is a bridge connecting ERP Systems and basic automation Systems, responsible for factory-level production management and scheduling. Its reliability is indispensable. Once the system goes down, production will pause and the loss will be immeasurable. For such a key business system, the cluster solution cannot meet the requirements. The cluster system is designed for Disaster Tolerance and used for system recovery in the case of disasters, and recovery takes time, there is a risk of interruption.

Some vendors claim that the cluster system can be switched seamlessly, which is not possible in most cases. Some demo environments can indeed be achieved, but in actual situations, seamless switching is difficult to achieve. In contrast, the fault-tolerant system has higher reliability. From the reliability level, the reliability of the fault-tolerant system is 99.999%, that is, the level of 5 9; the cluster has only 4 9. It indicates that the difference is only the area 0.009, but in fact, the difference is huge.

CIOs must be clear about the differences, but it is difficult for CEOs to understand the differences. Five 9th and four 9th are some technical languages, and it is often half the effort to persuade CEOs.

Persuade the CEO not to adopt technical languages, but to clarify one thing: a lightning flash or a control signal error leads to a bullet train rear-end collision. The culprit is that the reliability is not enough. Is Fault Tolerance okay? Of course not, but the fault tolerance solution is the most reliable solution. If the fault tolerance cannot be resisted, no system can resist it.

In a word, the CEO should make multiple choice questions. The CEO may not select the best solution, but must make the CEO aware of possible risks. The CIO only needs to do this. The final decision should not be considered by the CIO.