Slashdot Mirror


The New Data Center Capital of America

crimeandpunishment writes "Move over Silicon Valley, here comes... Buffalo. Where the weather might actually be a big advantage. The recent opening of Yahoo's state-of-the-art data center, which uses the region's cooler climate and a high-tech 'chicken coop' design to dramatically lower energy costs is getting a lot of attention in the industry."

4 of 162 comments (clear)

  1. Silicon valley.... by catbutt · · Score: 5, Insightful

    was never really known for being a "data center", it's more known for where engineering and development happen.

    Data centers don't really need that many highly skilled employees working on site. In the future data centers might have no one employed but security guards and (relatively unskilled) maintainance. In that case it doesn't really matter where they are located, at least in terms of helping the economics of the region.

    1. Re:Silicon valley.... by cgenman · · Score: 3, Insightful

      Compared to hurricanes, mudslides, snowstorms, and other natural disasters, earthquakes are pretty tame. They happen once every few years, and rarely knock out the power. The snowstorms in the Pacific Northwest caused much more extensive computer outages than the occasional earthquake in California.

      Really, the only problem is that you're shaking active hard drives for about 30 seconds, which is never good. But most are good enough to park their heads, and it rarely causes a real head crash.

  2. Re:Not as cool as it used to be by FuckingNickName · · Score: 4, Insightful

    These days, though, with advances in lights out management, you can build a huge data center and only need a few low-pay button pushers

    This explains why 10 years ago the admin helped you out, and today you help out the admin.

    Remind me not to host any nontrivial systems where your philosophy manages the data centre. I want skilled people working quickly where the problem is going to happen, not slowly by trying to troubleshoot 1000 miles away.

  3. Re:Not as cool as it used to be by coryking · · Score: 4, Insightful

    If you are running a massive data center that hosts a webfarm, cloud cluster, or some other large horizontally scaled computing project and require highly technical staff troubleshooting individual machines onsite, your process and application is completely screwed up. A well designed, horizontally scaled app should not fail if multiple machines go down.

    At the scale of Yahoo, Google or facebook, they probably dont even bother to troubleshoot a machine that is even hinting at questionable behavior. They just yank it off the load balancer and have some unskilled dude take the machine, dump it, and put in a new one.

    If you have a massive failure of your system, short of a natural disaster it ain't a hardware issue or a server issue. It is an application bug that require software engineers to fix. They don't have to be at the datacenter, they just create a patch from the comfort of their normal office (or home) and push it out to production.