Domain: datacenterknowledge.com
Stories and comments across the archive that link to datacenterknowledge.com.
Comments · 269
-
Earthquake Tested
A related link at the end of the article describes how Sun took one of their Black Box systems to a giant shake table at the seismic research center at UCSD, to see how well it would hold up during an earthquake. Some things pulled loose, and some things will need a little redesign, but it was able to keep functioning during and after the simulated earthquake. Sun produced a slick little video of it.
-
Re:Datacenter????Nowhere in the wikipedia entry does it say that is "criteria" for a data center. In fact, it says things like "they generally include..." or "are usually"
While it doesn't say that those are the "criteria", you even concede that the article says that they "generally include" or "are usually" comprised of the listed criteria, so I don't think I'm that far off the mark in thinking that the items mentioned would be considered "standard".
But that's no reason to go on a rant about what a "data center" vs. a "server room" is.
In fairness, I didn't go on a rant. I made a rather succinct (and ok maybe a little sarcastic) comment about the scope of what is generally considered a datacenter, and subsequently responded to various differing opinions. If I had posted some lengthy diatribe about how "real men work in datacenters of X size" or something to that effect, I'd admit I'd flown off the handle, but I was merely trying to make a point and for some reason it seemed to really piss a few people off.
I have to ask - what is your exact minimum qualifications of square footage, number of racks, kw density, storage density, and kw requirements before you would consider it a "data center"? As I stated in this reply, I don't think there's a minimum of square footage or even necessarily minimum capacity in terms of KW consumption or data throughput. I DO, however, contend that there are certain "benchmark" levels of infrastructure that one must have to call a facility a datacenter (versus a server room, colo room, IT lab, etc etc). Some of those levels are truly redundant utility feeds (from different substations/grids), redundant generators (if you need 1, you have 2, just in case), redundant UPS's (ditto), static switches and associated gear (for moving a "live" critical load), FM-200, real access control (physical security, biometrics, motion detector camera systems, etc).
I guess I'm just surprised that a few people are really bothered by this. I've got guys arguing with me that their buddy's garage with a UPS and an air conditioner is a datacenter for crying out loud!! This isn't about my trying to be egocentric or playing "my datacenter can beat up your datacenter" at all, I'm merely pointing out that it's not really fair to use the same term to refer to places like mentioned in the article and places like these, or this, or this, or this.
You make a good point that it is all relative to a degree, and perhaps people are reacting to what I'm saying because quite frankly, some of them might not have seen a large datacenter before (other than in the movies) and thus are perceiving that I'm trying to dump on where they work/have worked. In truth, I'm just taking issue with everything being lumped together, particularly within the context of a discussion of building one (apparently from start to finish) in 60 days. It's just a ridiculous suggestion. It's no more ridiculous than suggesting that a drive across town in a Yugo is the same thing as a drive across the country in a Ferrari (yes, I even had someone in another thread use the old Slashdot car analogy on me) because they're both taking place in cars. I'm not dumping on the Yugo or claiming the Ferrari's better, I'm just saying it's not really the same thing and one's definitely going to cost more and take longer. -
Button Nearly Drops Calif. Grid, plus some tipsOn April 15 a disgruntled tech hit the Emergency Power Off button at the data center controlling the California power grid, knocking it offline for seven hours. Power industry officials said that if had happened on a weekday instead of late on a Sunday night, much of the Western US might have gone dark.
While it's hard to prevent that kind of deliberate sabotage, a recent session at Data Center World focused on strategies to mitigate the risk of EPO disasters. The bottom line: put a cover on the button, make it well-marked, and have separate buttons or switches for power distribution units, UPS power and HVAC (the code often allows this). And as a deterrent, have video surveillance of the exit area where the button is placed.
-
Button Nearly Drops Calif. Grid, plus some tipsOn April 15 a disgruntled tech hit the Emergency Power Off button at the data center controlling the California power grid, knocking it offline for seven hours. Power industry officials said that if had happened on a weekday instead of late on a Sunday night, much of the Western US might have gone dark.
While it's hard to prevent that kind of deliberate sabotage, a recent session at Data Center World focused on strategies to mitigate the risk of EPO disasters. The bottom line: put a cover on the button, make it well-marked, and have separate buttons or switches for power distribution units, UPS power and HVAC (the code often allows this). And as a deterrent, have video surveillance of the exit area where the button is placed.
-
Quincy ain't just about Google and Microsoft.
-
Re:as good as WoW?
google for their current hardware setup, they're doing exactly that (granted not the 72-way servers)
http://www.datacenterknowledge.com/archives/2006/
S ep/13/eve_online_30000_users_on_one_server_shard.h tml could be a good place to start -
Re:Location, Location, Location
PG&E, the utility in California, offers rebates to customers who use outside air to cool their data centers. Some data centers in the Northwest use this as a way to lower data AC costs. The tradeoff is that you have to filter the air for contaminants and watch the humidity.
-
Re:What *is* the reason?
That stuff usually scales quadratically or worse with the number of objects, so it is quite some difference between a 64 player BF2 match and a lazy 2000 player day on a WoW server. If a lot of players have the stupid idea of gathering in a small area, things really get fun. Doing this requires some real hardware.
Like what? Even spec'ing out a Dell Poweredge 6800 with Quad 3.4GHz/800Mhz/16mb Cache, Dual-Core Intel® Xeon 7140M Procssors and 32GB or RAM comes to just $22,780. What kind of hardware is WoW running on? Everquest? Everquest has 1500 servers around the world according to this article They don't make their own quantum computers or anything, so they must be using the same hardware as everyone else. I'm guessing that at worst you could invest 100k in hardware and get a decent size virtual world going on 4 or 5 high end machines. Or take that same 100k and plunk it down on say 30 lower end servers and you have yourself some serious hardware capability with today's processors.
Or don't have a single monolithic game space, but rather have individual developers create a unified distributed gaming space. The web of online gaming. That would distribute the costs over multiple small developers and instead of one company needing thousands of computers to rival a big time publisher, they could invest in just a few to create a small part of a larger virtual world. Yes, the economics are still not as appealing as say running a simple web site, but expectations for an interactive online experience are pushing us in this direction. We can either choose to accept fragmentation and proprietization of the web or we can choose to adapt the open source movement to the new technology. -
Re:Share the PowerDoc - Thanks for the link re the file system. I hadn't seen that.
Google actually has data center facilities all over the place (it's hiring data center staff in nine different locations), and is building more. They are said to be shopping for property in North Carolina, and contemplating a $1 billion facility in India. I think their center network is rapidly becoming more distributed, and given the issues in Silicon Valley, they'll be accelerating that trend.
-
Re:Iceland needs a really big pipe....
Some data centers actually cool their facilities with air pumped in from outside their buildings. There's a study underway at Lawrence Berkeley National Laboratory looking at the use of air economizers at seven data centers that have participated in a PG&E program offering rebates for folks who do this. The study is looking at concerns that the use of outside air will introduce contaminants or excess humidity into the data center. Not for everyone, but seems to work for some folks.
-
Demand Patterns Track Customers, not CostWhile Google, Yahoo and Microsoft can build stand-alone data centers in Washington State to power their search engines, enterprises and Internet companies want space and connectivity in the major business hubs. Right now the strongest demand for data center space is focused on three markets: New York/Northern NJ, northern Virginia, and Silicon Valley. While power is less expensive in nothern Virginia, none of these markets are cheap.
The other hot data center real estate market is Austin, which has benefitted from a regional focus on energy efficiency, a strong technology community, and a skilled IT workforce. Phoenix and Kansas City have become active markets for backup data centers, while a dark horse market is northern New York, which used its abundant supply of cheap hydro power from the Niagara River to snare a $166 million HSBC data center.
-
Demand Patterns Track Customers, not CostWhile Google, Yahoo and Microsoft can build stand-alone data centers in Washington State to power their search engines, enterprises and Internet companies want space and connectivity in the major business hubs. Right now the strongest demand for data center space is focused on three markets: New York/Northern NJ, northern Virginia, and Silicon Valley. While power is less expensive in nothern Virginia, none of these markets are cheap.
The other hot data center real estate market is Austin, which has benefitted from a regional focus on energy efficiency, a strong technology community, and a skilled IT workforce. Phoenix and Kansas City have become active markets for backup data centers, while a dark horse market is northern New York, which used its abundant supply of cheap hydro power from the Niagara River to snare a $166 million HSBC data center.
-
Yes, but the phone company HOSTS the MMOsWhile there's a certain logic to the scenario presented by Ramprate in TFA regarding phone companies and ISPs, it's also true that the largest online games are actually hosted by a phone company. AT&T hosts World of Warcraft and Sony Online Entertainment's major games. At this year's E3, AT&T announced the expansion of its online gaming operation. Given the hosting fees coming in from Blizzard and Sony, AT&T/SBC has a vested interest in their success. Does the nation's largest phone company have leverage in dealing with ISPs who might be tempted to mess with MMO traffic? I suspect they do. Food for thought.
If Net Neutrality did squeeze online gaming, it might create an opportunity for someone like GameRail, a high speed network that directly connects online game players to the servers that host popular FPS titles. GameRail peers directly with ISPs, universities and game server providers (GSPs). The question is whether game server hosts see usefulness in that type of middleman. The answer to that question might change in some of the scenarios imagined int eh article.
-
Why Gilder Is Telecosmically WrongEverything is getting cheaper but power, which for some data centers now costs more than hardware. Nicholas Carr explains why Gilder's assumptions are problematic:
"What Gilder calls 'petascale computing' is anything but free. The marginal cost of supplying a dose of processing power or a chunk of storage may be infinitesimal, but the fixed costs of petascale computing are very, very high. Led by web-computing giants like Google, Microsoft, Amazon, and Ask.com, companies are dumping billions of dollars of capital into constructing utility-class computing centers. And keeping those centers running requires, as Gilder himself notes, the "awesome consumption" of electricity"
As I noted in our commentary at Data Center Knowledge, the power issues with high-density blade server computing has been understood for years. Back in 2002, Liebert and APC and other equipment vendors were developing products that could address huge heat loads. They saw it coming, and sensed a market opportunity. So where were the chip makers? Even as cooling vendors prepared for the results of the huge power and heat loads, little was done to address their source.
-
Working Models and Cost Issues
This issue has a been a hot topic at conferences for data center professionals, with a lot of debate about timetables. Several facility designers are advocating DC distribution as the solution to the current power/cooling challenges. Corporate data center managers like the cost savings projections, but want to see it work in someone else's facility before they put their neck on the line and pitch a DC conversion to their bosses. That's the real value of the Livermore project discussed in TFA - it provides a working model.
Right now the cost of power is remaking the landscape of the data center industry. Yesterday there was another announcement of a huge data center in central Washington State. Sabey will invest $100 million in a facility right up the street from where Microsoft and Yahoo have data centers under construction. It's all about cheap hydro power. Both Microsoft and Yahoo have contracted for more than 40 megawatts of power from the local utility. That's why DC is one of the solutions that will begin to get serious consideration. -
Working Models and Cost Issues
This issue has a been a hot topic at conferences for data center professionals, with a lot of debate about timetables. Several facility designers are advocating DC distribution as the solution to the current power/cooling challenges. Corporate data center managers like the cost savings projections, but want to see it work in someone else's facility before they put their neck on the line and pitch a DC conversion to their bosses. That's the real value of the Livermore project discussed in TFA - it provides a working model.
Right now the cost of power is remaking the landscape of the data center industry. Yesterday there was another announcement of a huge data center in central Washington State. Sabey will invest $100 million in a facility right up the street from where Microsoft and Yahoo have data centers under construction. It's all about cheap hydro power. Both Microsoft and Yahoo have contracted for more than 40 megawatts of power from the local utility. That's why DC is one of the solutions that will begin to get serious consideration. -
BitTorrent, of courseHaven't you been reading Slashdot recently? There was an article about Bittorrent getting gigabit worth of capacity.
...
Besides, even old people in Korea get gigabit access in their cybercafes these days
:-) -
Data Center Secrecy and the Fight Club Rule"The first rule of the Data Center is - you do not talk about the Data Center. The second rule of the Data Center is - you DO NOT talk about the Data Center."
Like Google, Wal-Mart is crazy secretive about a mega-data center in Joplin, Mo., which caught the attention of the local media. The Wal-Mart rep: "This is not something that we discuss publicly. We have no comment. And that's off the record."
-
Big Blue Chillin' With Water
IBM today unveiled a chilled water cooling system to manage heat loads for high-density cabinets using its eServer cluster products. It's been dubbed "Cool Blue." The debate about water cooling in the data center is a heated one (for a conversation about cooling) and has been discussed here before. Amid all the other strategies mentioned here, it's interesting to see Big Blue touting water cooling.