Bing Bopped By Big Blaze

A FIRE LATE THURSDAY NIGHT IN SEATTLE, WASHINGTON, caused a full power outage in a building that houses several internet server farms including Microsoft’s Bing Travel and Verizon’s DSL service for the Seattle area. Also affected was Authorize.net, a credit card authorization site that services over 250,000 online retail merchants.

Microsoft said that only Bing’s travel feature was affected and that all other Bing sites are operating normally from other locations.

KOMO-TV and radio operated from there too and were off the air for a while. They are now broadcasting from remote locations. Fisher Communications, which owns the complex, also owns and operates KOMO.

KING-TV Ch. 5 has this video report:

This is the second time in just under a year that an electrical fire has knocked out the server farms in the building. Fisher Plaza bills itself as “…the only mission-critical business community in the Northwest combining Class A office, data center, colocation, and retail space with 21st century communications and media services.”

“Pretty frustrating,” one customer told CNET News. ”I understand problems happen, but this the second time in a year that we have had to explain to our customers about an outage. This is supposed to be a ‘world class’ facility. Brings up a lot of questions that are still unanswered from the last outage.”

The Seattle Times is reporting:

The small fire, which broke out around 11 p.m. Thursday at the complex near Seattle Center, apparently began with a failure in KOMO’s equipment, which caused a short, said Seattle City Light spokeswoman Connie McDougall.

It happened in the garage level of one of the buildings in Fisher Plaza, at an electrical vault, where KOMO’s transformers meet the city’s power lines, she said.

Many of the internet customers were able to eventually transfer their service to other servers, but by late Friday night the power was still off in the building. “Fisher is bringing in electrical generators to restore power to the building, at which time it can further assess the situation. The Company is working to restore normal service to its customers as soon as possible,” Fisher said in a statement.

Blogger Kyle Mulka is maintaining a list of websites that are affected by the outage HERE.

Read the full story in the Seattle Times HERE.
The Post-Intelligencer had MORE.

Share and Post on Facebook, Twitter and More:

Featured from across Fire/EMS

Dal90

And for the people whining about power outages...

At least your data center isn't known for it's armed robberies:

http://www.datacenterknowledge.com/archives/200...

Gotta give the thugs credit for style worthy of a Hollywood director...cut hole in wall, when night manager investigates hit him with the taser and tie him up.

And that was their fourth break in.

Dal90

I'm working at a company right now that (primarily) markets into the telecomm industry, and telephone & cable companies predominantly use -48VDC for supplying power to their equipment.

Although I heard of it before, I had never seen it in person. Makes uninterruptible power supply a snap -- everything is fed by a bank of fairly standard batteries that are kept charged by the power grid.

The power supplies for the data center are usually reading in the mid 50s for voltage and a smidge over 500amps on their meters when I walk by.

The commercial off-the-shelf servers we run still are traditional 110VAC, due to the extra expense of buying 48VDC power units in each server.

I reckon like MANY things on the internet and computing, 110VAC is prevelant simply because it got in the door first and the cost of rebuilding the infrastructure to use DC today is too great. Email, IPv4 (the IP addresses we're familiar with), and the current Domain Name Server systems are all the same -- no one in their right mind would chose from scratch today what we actually use. But no one wants to encounter the pain and cost today of fixing it.

Google splits the difference in the servers they custom build. They use 110VAC to each server, but the AC charges an internal 12VDC battery. Unlike most motherboards that use multiple voltages (usually both 5V & 12VDC), the Google boards use just 12V. So Google has no central uniterruptible power supplies for their data centers -- each server has it's own small battery to keep it running through a momentary power glitch, or provide time to shutdown cleanly if it's an extended outage.