Accident confidentiality affects data center O & M Security

Accident confidentiality affects data center O & M Security

A few days ago, the data center industry experts obtained a simple suggestion to improve the reliability of the data center, that is, to increase the redundant facilities of the data center by an order of magnitude, if the data center industry can be persuaded and accepted. Experts think this is unrealistic.

Reliability is not a green energy-saving problem, but it must have an influence on energy efficiency. These two requirements are usually opposite to each other, and the reliability is getting higher and higher, which usually means that the data center stores more diesel fuel in the reserve area and even deploys more equipment in the entire hot backup facility, this will increase the carbon footprint of enterprises.

However, all these efforts in reliability may result in carbon waste, which may be a waste of time.

Many may be confused about the reliability of their data centers, and key facility expert Andy anceor gave a thought-provoking speech.

According to the report, the hardware may sometimes become invalid, which is not common, but data center faults are usually caused by human errors, or because the design, testing, maintenance, or operation details of the data center facilities are neglected.

Then he listed some typical data center accidents. For example, if a data center, its backup diesel storage tank forgot to be sealed, the consequence was that the diesel engine was not started in time, the reason is that the part with only 20 cent coin size is missing. When a fault occurs, the power system switches from flywheel to the generator set, but the output frequency begins to drift. This creates an "out-of-phase" condition for a static switch, and its transformer coils are saturated.

In another example, a data center uses two flywheel UPS systems to back up power redundancy, but the fault still occurs. When Vicente was called for an investigation, he immediately knew where the problem was, because he had encountered this situation in other data centers.

Such faults can be predicted and prevented, provided that the previous fault investigation results have been published. Because the failure investigation of the data center requires the signing of a confidentiality agreement (NDA), the disclosure of this information will benefit enterprises with such problems, but in fact, due to confidentiality, this information is usually not disclosed.

Today, due to technological advances, complex systems can be very reliable. Mature industries, such as the aviation industry, can achieve a fault-free 0.1 million to 20 hours, and this number is a shame for the data center industry, because it did not.

"There is still room for improvement in data center reliability, just because people don't agree to share it ." . Airlines are forced to share the results of the accident investigation through law. They are strictly controlled because in the event of a fault, passengers will be suspended.

This does not apply to data centers, although this may change because more important personal security protection functions are required in the data center infrastructure. Industry insiders do not want strict supervision and control.

However, at this meeting, ancent proposed an idea. That is to say, on the basis of sharing an anonymous database, data centers that have experienced accidents can share their own failures. Through actual surveys, people in the data center industry can learn from the cases. If this idea is implemented, it may lead to a real breakthrough in data center reliability.

It will be a tough struggle to make this happen, but it should win the support of the industry, because no matter who wants to cut the cost of data centers, whether it is a waste of electricity, capital or energy.