The staff at SearchDataCenter.com recently split our blog Server Specs into three new blogs in order to narrow our focus on specific aspects of the data center industry:

Data Center Facilities Pro: A SearchDataCenter.com blog about data center facility management, engineering and design. This blog covers news, trends and tips on topics like data center cooling, data center backup power and data center energy efficiency.

Mainframe Propeller Head: A SearchDataCenter.com blog about the IBM mainframe and its alternatives. This blog will cover news, trends and tips on topics like the System z hardware, the z/OS, z/VM and z/VSE operating systems, mainframe jobs and software licensing costs.

Server Farming: This blog serves as forum to discuss the latest in server hardware, systems management, Unix-Linux-Wintel operating systems and large distributed computing systems.

Please bookmark the new blogs. Server Specs will remain online in archive mode. Thanks for reading.

Eaton, one of the big data center UPS manufacturers, is offering a free handbook on uninterruptible power supply (UPS) batteries. Most people know that the most vulnerable part of a UPS are the batteries. Knowing how charged they are and how much juice they have left can mean the difference between having enough through time to get the generators started, and leaving your customers without a data center.

I ordered one. I can always learn more about UPS’s and their batteries. According to Eaton, the guide “covers characteristics, performance and maintenance of UPS battery types along with a comprehensive glossary of battery terms.” Sounds good to me. And since I already get plenty of emails from Eaton for all their products, I didn’t care about putting my information in there. They can only send me so many emails, right?

In June, Chris Johnston, Senior VP and Critcal Facilities Chief Engineer at Syska Hennessy Group wrote a Tip for SearchDataCenter.com, “Uninterruptable power supply load bus synchronization: Yes or no?”. We asked for reader feedback, and heard from Brad Walter, Director of Applications and System Development at Active Power, who shared this story:

The one issue that this article did not address is static transfer switch (STS) alarms. I have a customer who claims that even the STS with asynchronous transfer capability alarms when the phase angle between the two sources drifts beyond a certain limit (say 15 degrees), and that a new instance of the alarm is generated every time the two sources drift out of synch. This would not only be annoying to data center operators but significant time would be wasted in re-setting alarms, and even more importantly, frequent false alarms tend to make people ignore alarms that require attention. Have you seen the same issue, or is this customer either not actually using asynchronous transfer STS as he claims, or has he failed to set up alarm features on his switches correctly for asynchronous transfer operation?

Chris provided this feedback for Brad:

Thanks for your question, Brad. I agree that frequent nuisance alarms are annoying and a potential source of unintended downtime. Repetitive nuisance alarms can obscure our perception of a real problem when it occurs. Human nature is to be lulled into being “asleep at the switch.”

Not knowing any specifics about your customer’s situation, I suspect that the alarm features on the STS may not be set correctly to eliminate nuisance alarms. If synchronization between the two STS sources is unnecessary, then the need for an “out of synch” alarm is unnecessary, regardless of the phase angle between the two sources. I suggest that your customer contact the STS manufacturer.

This lead us to wonder, is this a “Picnic” type problem (problem in chair not in computer) or is this a problem with a certain STS?

I called Brad to follow up, and to get a little edification about how these switches are supposed to work so that I could better understand the problem. Brad kindly talked to me for about 30 minutes, giving me a crash course in Electrical Engineering 101 and a refresher on trigonometry.

He explained that in a large data center with many of these switches and alarms, that the headache from them all going off could be enormous. The particular model involved has software in the switch that is supposed to be able to take care of any out of phase problems, keeping the downstream equipment safe. If the software is working as advertised, there would be no need for an out of phase alarm at all. But for whatever reason, it has an alarm, and it’s going off — all of them, at each UPS in the data center.

Brad declined to point a finger the switch manufacturer specifically (a perfectly sane response), but shared that it was one of the big players in this space: Cyberex, Liebert, PDI, or LayerZero.

Without other stories of the same problem, it’s harder to make a case out of this and ask the manufacturers tough questions. So, this is a bit of a fishing expedition: Have you experienced a similar problem with your STS? If so, did you resolve it? How? Which model STS had this problem?

If you represent any of the four STS manufacturers, are you willing to say that this problem could not happen with your equipment? Or, could it, and what should the end-user do about it?

In a response to rising commodity costs, Emerson Network Power announced today that it would increase prices on its Liebert and Knurr brand products.

The increase averages 5% and is the result of “significant cost increases in many of the materials used in the products we manufacture,” according to a statement from Chuck Spears, the president of Liebert North America.

In a downturn economy, the upped prices tend to end up on the user’s balance sheet. Let us know if you’ve received similar notifications from other vendors.

Oracle too: A reader pointed out to me that Oracle recently increased their software licensing costs, in some cases by as much as 15 percent. Why? According to Forrester analyst Ray Wang, it’s due to the devaluation of the U.S. dollar.

There are plenty of ways to start measuring your data center’s Power Usage Effectiveness, or PUE, which is a comparison of the total facility power to the IT load. But in recent months Microsoft has opened the curtains on its own operations, thanks largely in part to two data center pros there, Michael Manos and Christian Belady.

Back in April, Manos spoke to us about Microsoft’s data center strategy. It was one of the first times the Redmond, Wash. The next month, Manos showed us a demo of Scry (video here), the tool the company uses to measure everything in their data centers.

Now Manos and Belady have written a series of three lengthy blog posts on PUE, why it’s important, and how to go about measuring it. For those of you out there still a little daunted by the task of measuring key analytics in your data center facilities, this could give you a good start.

With the upcoming holiday sure to brighten the skies across the U.S. with colorful flashes of light, I thought it would be a good time to share some of the advancements in Flash-memory based storage devices. Solid-state drives (SSD) can help bridge the server and storage performance gap and decrease data center power requirements. A few weeks ago, at Hewlett-Packard Co.’s Technology Forum, a relatively new entrant into the field, Fusion io announced that is adapting Fusion-io’s ioMemory architecture to HP’s enterprise-class servers, including the HP BladeSystem c-Class system.

“Adapting this technology specifically for HP servers offers radical increases in associated performance for a broad range of applications and workloads and can dramatically improve the effectiveness of data center architectures,” said Fusion-io CEO Don Basile. “With our ioMemory architecture, we’re getting more than 200,000 IOPS [I/O operations per second] within HP BladeSystem c-Class server blades today.”

When I met with Basile for a mere 30 minutes, he shared an intense amount of information, including a variety of statistics that illustrate the power of the technology. The ioMemory technology is still costly enough that it’s not ideal for long-term archival storage, but the speed that it provides makes it valuable in the active data requirement area. Basile explained that the technology is capable of 3.2 GB/s of sustained bandwidth, with extremely low latency (50 microseconds). In terms of I/O, a traditional enterprise application server that is I/O-intensive may cost $15 to $20 per I/O versus $0.15 per I/O using the Fusion-io system.

In addition, Basile explained that all this speed is possible with less energy expenditure and less heat than a typical storage area network (SAN). Heat and power are reduced because of the lack of mechanical heat that would otherwise be generated by spinning disks. In addition to using less than 1% of the power required by a typical SAN, the footprint is minimal: 16 ioMemory cards can fit into 10U without any oversubscription. Basile noted that consumer electronics such as the iPhone have helped lower the cost of silicone chips.

HP is not alone in integrating SSD technology. Earlier this year EMC announced that it had added SSD in its enterprise Symmetrix system, and more recently rumors have circulated that the company will add it to the Clariion storage array. Sun has also announced that it will release a new version of the Solaris operating system designed to integrate flash and traditional disk-drive storage. And Sun’s CEO Jonathan Schwartz says that it’s not just a flash in the pan. Sun’s Adam Leventhal has also produced an informative technical article outlining the technology and its optimized uses.

In the age of increasing energy costs, efficiencies in data centers are welcome, and the Flash technology is likely to play a key role in the data center of the future. If you’re interested in the topic, check out the Flash Memory Summit in August.

I noticed a common theme at the string of computing conferences I’ve attended in the past couple of months: The future of the data center is going to be cloud computing, and resistance is futile.

I heard this from VMware Inc.’s President and Chief Executive Officer Diane Greene during her keynote at the JP Morgan Technology Conference in Boston in May, and the point was driven into the ground during the Enterprise 2.0 Conference there in June. I heard these predictions repeated during the annual Red Hat Summit and the USENIX 08 conferences, also held in Red Sox Nation last month.

Major league players in the data center space like VMware Inc. are putting their efforts into cloud computing because of predictions that it will eventually be the mainstream way information is handled and software vendors are starting to introduce products to manage cloud computing environments.

David Patterson, a professor of computer science at U.C. Berkeley, said during his keynote speech at USENIX that cloud computing is part of the data center evolution already under way.

“In addition to the processor evolution [from single-core to dual- and now quad-core processors], on a larger scale, there are a number of changes happening in the data center; flash memory is replacing mechanical disks, we have software as a service, and utility computing [a.k.a. cloud computing] is being used to outsource the data center,” Patterson said.

The advantages of cloud computing are clear, he said.

“With cloud computing, you put $0 down for your own data center, and pay as you go, and there is no penalty for scale up, which happens instantly. It allows fast scale up with no dead or idle CPUs, and no provisioning is required,” Patterson said.

This is especially appealing to data centers that have maxed out their power resources, but need to increase their infrastructure.

Though cloud computing is considered an immature technology, it really isn’t. The chief architect of the Xen project, Ian Pratt, said during his session at USENIX, called Xen and the Art if Virtualization, that the folks at Cambridge University who started the XenoServer project with him back in 1999 were architecting it under the cloud computing concept.

Though their ideas about what the cloud would look like differ from what we see today, the concept was similar: Develop a public infrastructure for wide-area distributed computing that can be used by people across the world.

“We originally thought there would be data centers all over the world, and clients would be able to choose a location, perhaps close to another IP address they wanted to interact with,” Pratt said. “The other difference is, we thought the machines would be owned by many different merchants, and there would be a broker acting as a third party recommending the different vendors, and those brokers would take a fee.”

Instead, we have companies like Amazon.com, Google and Salesforce.com offering the complete cloud computing environments , but Pratt expects this to change.

“I think we will see cloud computing move in the direction where it will become more open instead of all of the hardware, software and networking being located at and owned by a Google or Amazon.”

Today, most cloud computing providers host x86-compatible applications on virtualized servers, and most support only the Linux OS, according to Cambridge, Mass.-based Forrester Research Inc. To keep costs low, many cloud providers use a Xen-based hypervisor. Charges for usage are usually based on CPU hours, gigabits consumed and gigabits per second transferred rather than on a monthly service fee.

Specifically, Amazon charges 10 cents per compute hour used and 15 cents per gigabyte of storage. According to Forrester, that translates into about $70 to $150 per month for a fully utilized Amazon server, versus the average $400 a month that it costs an enterprise to run a server.

The benefits aside, IT pros are apprehensive about taking their mission critical apps out of their secure data centers and putting them into something as translucent sounding as cloud computing. This fear was quite evident during the Enterprise 2.0 conference event called “An Evening in the Clouds.” A panel of IT pros sat and listed to Google, Amazon and Salesforce as they fluffed cloud computing, and then they voiced their many concerns.

Is it secure? Is it reliable? Does it perform better than my existing data center?

The answer from all the cloud computing providers was, of course, a resounding “yes.”

But not all applications are available in the cloud, so it isn’t for every company. The cloud computing environment also lacks government standards, which makes some users nervous.

“I wouldn’t suggest moving all of your apps over to the cloud today, but hopefully one day all will be right in the world,” said Jeff Keltner, the business development manager at Google Apps.

When it comes to data center metrics the one most often talked about is square footage. Nobody ever announces that they’ve built a facility with Y-tons of cooling, or Z-Megawatts. The first metric quoted is X-square feet. Talk to any data center manager however and they’ll tell you that floor space is completely irrelevant these days. It only matters to the real estate people. All that matters to the rest of us is power and cooling - Watts per square foot. How much space you have available is nowhere near as important as what you can actually do with it.

If you look at your data center with a fresh eye, where is the waste really happening?

Since liquid-cooled servers are at the far right-hand side of the bell curve, achieving electrical density for the majority of us is usually a matter of effectively moving air. So what is REALLY preventing the air from moving in your data center? I won’t rehash the raised floor vs. solid floor debate (since we all know that solid floors are better) but even I know that the perforated tiles, or the overhead duct work is not the REAL constraint. A lot of folks have focused a lot of energy on containment; hot aisle containment systems, cold aisle containment systems, and even in-row supplemental cooling systems.

In reality however, all of these solutions are addressing the environment around the servers, not the servers themselves which are after all, the source of all the heat. Why attack symptoms? Let’s go after the problem directly: The server.

First of all, the whole concept of a “rack unit” needs to be discarded. I’ve ranted before on the absurdity of 1U servers, and how they actually decrease data center density when deployed as they are currently built. I’d like to take this a step further and just get rid of the whole idea of a server case. Wrapping a computer in a steel and plastic box, a constrained space, a bottleneck for efficient airflow is a patently absurd thing. It was a good idea in the day of 66 Mhz CPUs and hard drives that were bigger than your head, but in today’s reality of multi-core power hogs burning like magnesium flares it is just asking for trouble. Trouble is what we’ve got right now. Trouble in the form of hot little boxes, be they 1U or blade servers. They are just too much heat in too constrained spaces.

Virtualization won’t solve this problem. If anything it will just make it worse by increasing the efficiency of the individual CPUs making them run hotter more of the time. Virtualization might lower the power bills of the users inside the server, but it won’t really change anything for the facility that surrounds the servers in question. The watts per square foot impact won’t be as big as we hoped and we’ll still be faced with cooling a hot box within a constrained space.

So here is my challenge to the server manufactures: Think outside of the case.

This isn’t a new idea really, nor is it mine. We’ve all seen how Google has abandoned cases for their servers. Conventional wisdom says that only a monolithic deployment such as a Google data center can really make use of this innovation. Baloney. How often does anyone deploy single servers anymore? Hardly ever. If server manufacturers would think outside of the case, they could design and sell servers in 10 or 20 rack unit scale enclosures. They could even sell entire racks. By shedding cases altogether, both server cases and blade chassis, they could create dense, electrically simple, easy to maintain, and most importantly easy to cool servers. The front could be made of I/O ports, fans, and drives. Big fans for quiet efficiency. The backs could be left open, with electrical down one side and network connections down the other. Minimize the case itself to as little as possible… think of Colin Chapman’s famous directive about building a better race car: “Just add lightness.” The case of a server should serve one purpose only: To anchor it to the rack. Everything else is a superfluous obstruction of airflow. No need for steel, as plenty of lighter weight materials exist that can do the job with less mass.

Go look in your data center with this new eye and envision all those server cases and chassis removed. No more artificial restriction of airflow. Your racks also weigh less than half of what they do today. You could pack twice the computing horsepower into the same amount of space and cool it more effectively than what you have installed.

Ten years from now we’ll look back at servers of this era and ask ourselves “what were we thinking??” The case as we know it will vanish from the data center, much like the horse and buggy a century before. We’ll be so much better without them.

Last week I attended a data center tour with Sun Microsystems’ Dean Nelson, Sr. Director of Global Data Center Design Services.

Nelson’s team consolidated four large Sun Microsystems campuses in California – consolidating over 200,000 square feet of data center down to 80,000 square feet, while still allowing capacity to grow.

This new data center, built into Sun’s existing office space in Santa Clara, is designed for modular growth and is very energy efficient (1.28 PUE in one of the rooms, according to Nelson)

In this first video, Nelson demonstrates the Starline Busway, a power component Nelson Describes as track lighting on steroids. These products allow Sun data center staffers to plug in anything from 120 single phase power to 100 amp 3 phase without an electrician, allowing for modular growth and flexibility. The busways also have IP connections that track power usage in real time.

In this second video, Nelson shows us the hot-aisle cold-aisle containment strategy Sun is using. It’s essentially a ceiling on the hot aisle that prevents hot return air from mixing with the cool intake air.

During the USENIX ‘08 Annual Technical Conference in Boston this week I attended a session titled “Playing Fast and Loose with the Sysadmin Space-Time Continuum” led by a jokester named David Blank-Edelman, the director of technology at Northeastern University.

The interactive session was designed to help the sysadmins in attendance solve their most pressing data center challenges, and some creative solutions were thrown around.

But the best part of the 90 minute session had nothing to do with problem solving; it was the debauchery that Blank-Edelman coached attendees to employ in their data centers.

Blank-Edeleman warmed up the room of about 40 sysadmins with some critical bonus interface ideas, like how to get bird chirp sounds into server rooms. He directed attendees to the site Peep, which lets sysadmins monitor their networks with bird sounds instead of the traditional beeping.

“It is quite lovely, as long as there aren’t any issues, in which case the server rooms becomes a scene from Alfred Hitchcock’s movie ‘The Birds,’” he said.

Blank-Edelman, who also authored the book “Perl for System Administration; Managing Multi-Platform Environments with Perl” also showed his session attendees how to have some fun with Proxies. He told the story of a fellow sysadmin, Peter Stevens, who got sick of his next door neighbors using his wifi, and instead of encrypting it, decided to have some fun. Stevens arranged it so that any unauthorized users would be sent through a web browser that flipped the user’s webpage images upside down.

But of course, the purpose of introducing these antics wasn’t to inspire mayhem in data centers across the country - well, maybe a little - but mostly, it was to get the wheels of creativity turning.

“Being in the upper echelon of sysadmin society, you have to be able to improvise, and to do that you have to talk to other creative sysadmins and think outside the box,” Blank-Edelman said.

For instance, a creative firewall idea involving port knocking, which is used to keep external traffic - and hackers - out of systems. In general, when data gets transmitted to closed ports, it is received by a monitoring daemon that only opens ports when the correct port sequence is sent.

Blank-Edelman suggested starting out with a firewall that does not include any ports at all. Clients then attempt to open a random set of ports –say, 3, 7, 9, 12 - and only the clients that knock on the right set of ports are let in, he said.

“It’s a cool idea. When have you heard of starting with no access at all? People have taken this idea in all different directions,” Blank-Edelman said.

There were plenty of these little tips and tricks mixed in with funny antics during the session, and after a morning of technical whitepapers, this afternoon session was a sigh of relief.