What you need to do to prevent data center downtime

Recently, according to a survey report on data center operations conducted by a research organization, some data center outages should not have happened. The main reason for the outages was not due to bad weather, unplanned maintenance, or even power grid failures. Instead, it was caused by poor planning and improper maintenance by some organizations. From airlines to Internet giants, these major users and enterprises have become victims of downtime and have deeply felt the impact of these preventable data center outages. Unfortunately, this situation is not uncommon.

Data center industry vendors conducted a survey of IT and data center managers in various European countries and found that 27% of respondents said they had experienced long-term outages in the past three months, which had an adverse impact on their organization's business. The vast majority of respondents (82%) believe that most key business processes rely on IT services, and 74% of respondents said that the health of the data center directly affects the quality of IT services. Organizational business depends mainly on IT services and IT equipment relies on the function of the data center. The fact that more than a quarter of data centers suffered long-term outages shows that some methods and measures at the industry level are wrong.

[[208813]]

(1) Advance planning to prevent power supply interruptions

Just as critical business processes rely on IT services, the data center itself must provide resilience to keep the business running. It is a core asset in any enterprise risk management strategy.

Staff errors, backup generators not starting, panicked decisions, etc. are all mistakes that can be prevented with the right processes and sound power system design. However, organizations often fail to follow the golden rule of data center power management: actions get results and consequences require actions.

Organizations need a disaster recovery process that clearly defines what steps should be taken when restarting a data center. In the event of a data center outage, staff may be overwhelmed and under pressure to restore normal service. After all, the staff's main goal is to resume normal operations as quickly as possible. Therefore, a disaster recovery process can help avoid prolonged outages.

(2) Improving staff skills

In fact, lack of awareness and understanding of power management in data centers is a common problem. Two-thirds of data center professionals participating in the survey indicated that they do not have full confidence in power assurance. Before organizations can master power management (from UPS maintenance to battery inspection), more power-related outages may occur.

However, workforce skills are critical to improving power availability, and many organizations find it difficult to recruit and retain the right professionals or talent, whether it is designing for energy efficiency, managing consumption on an ongoing basis, or handling power-related faults quickly and effectively to avoid and mitigate outages.

(3) Update and upgrade infrastructure

In addition to improving skills and securing power, data center infrastructure itself often needs to be updated and upgraded to meet corporate expectations for efficiency, reliability and flexibility. In this survey, about half of the respondents said that their core IT infrastructure needs to be strengthened, while nearly two-thirds of the respondents said that facilities such as power and cooling need to be strengthened.

Power management is increasingly becoming a software-defined activity. Given the gap in staff skills, software can bridge the gap between IT departments and power by providing power management options through dashboards that are familiar to IT staff, making management easier and automating the management of power infrastructure. Adopting this technology can reduce or eliminate downtime in the data center.

The industry has already moved toward a virtualized environment in the data center. IT and data center professionals are now very familiar with using virtualization technology to maintain hardware. So why not use the same principles? All power distribution designs and related resiliency software tools must be compatible with the products of major virtualization vendors to promote the future development of infrastructure. This approach will enable data center professionals to continuously maintain the system, thereby reducing the risk of infrastructure obsolescence.

Better preparation and disaster recovery processes could have prevented many outages from happening. Data center industry users must learn from these incidents and take necessary measures. Therefore, in order to reduce outages in data centers, effective power management is a must.