Story image

Opinion: How to capture “Lessons Learned” with cloud-based DCIM

09 Oct 2018

Article by Schneider Electric Digital Services and Data Centre Software vice president and general manager Kim Povlsen

“Those who cannot remember the past are condemned to repeat it.” – George Santayana

From a distributed IT and data centre monitoring perspective, what if cloud-based data centre infrastructure management (DCIM) software could help you determine whether current events are critical or unimportant based on historical data?

Knowing the difference between a non-event, an emergency, and an outright crisis is vital to continuous and efficient operations. IT and data centre professionals must be able to differentiate between the thousands of alarms that can be activated at any time in a normal working day.

Different level events must drive different actions

Operators can face many challenging scenarios and must make quick decisions based on real-time data. They must learn to classify the impact of alarms to determine a proper response.

Software with predictive analytics could take routine events – like an alarm firestorm – and prioritise which events are true emergencies.

Other events are unique – so extreme they are impossible to predict based on historical information.

Then there are those events that are difficult, but not impossible to predict.

In the European heat wave of 2018, one data centre operator reported getting 20,000-30,000 alarms a day because the outside air temperature was regularly rising above 35°C (95°F).

Operators quickly recognised a pattern.  An uptick in temperature routinely happened before the cooling system kicked in and reduced temperature.

This repeating cycle caused temperatures within the data centre to regularly rise above, and then fall below, target thresholds as often as 30 times a day. Monitoring systems responded by generating thousands of alarms on the rack PDUs as the thresholds were crossed.

Analysis helped operators understand which alarms needed attention and which could be ignored.

A software system that can recognise the characteristics of an event based on historical data, and grade the effects and responses to previous alarms, can inform the personnel about how critical or not such an event is likely to be.

Stop reacting, start predicting

New cloud-based data centre infrastructure management (DCIM) tools are constantly evolving to improve operations and mitigate potential issues.

By capturing endless, anonymised data points, cloud-based DCIM software is designed to help operators move from being reactive to being predictive.

The software system learns from operator’s reactions to different events and becomes “smarter” by establishing alarm priorities based on previously seen conditions.

By automatically building ratings of criticality each time an alarm occurs, the software can guide operators’ reactions with appropriate warnings.

If a crisis does occur, the system stands ready to highlight what is important and push to the background anything that is not. Through this support, operators can more quickly perceive what constitutes the real cause of a problem and respond appropriately.

The software system can become personalised to a particular IT environment and those who operate it in much the same way that a search engine, online retailer, or digital content provider learns your preferences and serves items that relate directly to your primary interest.

Discover today’s cutting-edge cloud-based DCIM tools

In a distributed IT or data centre environment, we are always learning more about real problems.

Improving today’s alarm systems will increase productivity and help deal with major hazards.

As systems scale and more data is captured, predictive analytics become feasible. This has the potential to raise the bar on what is possible through state-of-the-art, cloud-based DCIM software.

GitHub launches fund to sponsor open source developers
In addition to GitHub Sponsors, GitHub is launching the GitHub Sponsors, GitHub will match all contributions up to $5,000 during a developer’s first year in GitHub Sponsors.
Check Point announces integration with Microsoft Azure
The integration of Check Point’s advanced policy enforcement capabilities with Microsoft AIP’s file classification and protection features enables enterprises to keep their business data and IP secure, irrespective of how it is shared. 
Why AI will be procurement’s greatest ally
"AI can help identify emerging suppliers, technologies and products in specific categories."
Are AI assistants teaching girls to be servants?
Have you ever interacted with a virtual assistant that has a female-based voice or look, and wondered whether there are implicitly harmful gender biases built into its code?
Google 'will do better' after G Suite passwords exposed since 2005
Fourteen years is a long time for sensitive information like usernames and passwords to be sitting ducks, unencrypted and at risk of theft and corruption.
Hackbusters! Reviewing 90 days of cybersecurity incident response cases
While there are occasionally very advanced new threats, these are massively outnumbered by common-or-garden email fraud, ransomware attacks and well-worn old exploits.
Data#3 to exclusively provide MS licences to WA Government
The technology services provider has won two contracts with the Western Australia Government, becoming its sole Microsoft licence provider.
Why cash is no longer king in Australia
Australia is leading the way in APAC for granting credit on B2B transactions.