Data Centers

The seven types of power problems

One of the basic parts of your network infrastructure is, of course, your power supply. It's as basic as you can get, but you can't afford not to understand how your systems are powered and what kinds of problems to look out for.

One of the basic parts of your network infrastructure is, of course, your power supply. It's as basic as you can get, but you can't afford not to understand how your systems are powered and what kinds of problems to look out for. American Power Conversion (APC) has provided a very detailed white paper on the seven kinds of power problems as defined by the IEEE:

1. Transients

2. Interruptions

3. Sag / Undervoltage

4. Swell / Overvoltage

5. Waveform distortion

6. Voltage fluctuations

7. Frequency variations

According to the executive summary, here is what you'll find:

"Many of the mysteries of equipment failure, downtime, software and data corruption are often the result of a problematic supply of power. There is also a common problem with describing power problems in a standard way. This white paper will describe the most common types of power disturbances, what can cause them, what they can do to your critical equipment, and how to safeguard your equipment, using the IEEE standards for describing power quality problems."
If you need some remedial work in power supplies, check out the download.

About

Selena has been at TechRepublic since 2002. She is currently a Senior Editor with a background in technical writing, editing, and research. She edits Data Center, Linux and Open Source, Apple in the Enterprise, The Enterprise Cloud, Web Designer, and...

7 comments
naseri
naseri

i think it was great for me ! thank u very much . Naseri from Iran

mjd420nova
mjd420nova

A tractor trailer missed a corner and stuck a utility pole, knocking down a 69KV line into a 13 KV feeder to the data center. The surge destroyed every power supply faster that the protectors could respond and open the circuit. Besides the power supplies, the main transformer and breaker cabinets had to be replaced. A new system of fast acting relays and breakers was added once all the equipment was repaired and it only took three weeks. A long time for over 25,000 users to wait and caused a rush to mainframe terminals that could access the same data bases in an attempt to keep things current.

bandman
bandman

I'm pretty sure that the time someone unplugged our one of our two main power feeds at the colocation, it counted as an interruption. What would you label it as when they plugged the line back into the secondary circuit, tripping the breaker and bringing them both down again? /true story

robo_dev
robo_dev

sent the whole fire alarm system into full alarm, which caused an EPO for 1/3 the data center. We're talking about 35 Oracle database servers shut down cold. Took about two weeks to recover....

HAL 9000
HAL 9000

Where the system is fully tested to prevent further damage. So I'm assuming that changes to prevent this occurring again where put in place? Something along the lines of a Controlled Shut Down as apposed to a Crash Shutdown. :D Col

robo_dev
robo_dev

Blame the hardware guy and create some bogus story about what really happened....seriously. glad I don't work there anymore....

Editor's Picks