Data Center Connect Newsletter

How Deep Can Green Get? Ask Capgemini

Rakesh Dogra, DATACENTER Journal — Tue, 31 Aug 2010 04:00:00 GMT

In February of this year, Douglas Farquahar, Head of Sustainable Outsourcing at Capgemini U.K., wrote on how “intelligent IT can help save the planet, lowland gorillas, and money.” The company has quite literally put money down for each of these reasons.

Capgemini, a consulting, technology, and outsourcing company, operates out of Paris and has clients in more than 30 countries.

The statistics on how much a data center contributes to the world’s carbon footprint are well known. As much as 2% of global carbon emissions are courtesy of the data center industry. With demand for data processing and storage increasing steadily and swiftly, the IT industry is set to rank alongside the oil industry as a “major polluter.”

Green IT and data centers are quickly becoming the norm for companies. It also makes business sense to invest in green IT for the savings and operational efficiencies it brings. According to Forrester Research, the green-IT services sector is set to grow at a rate of 60% every year until 2013. According to Chris Mines, analyst at Forrester, more and more enterprises are going green and are seeking environmentally friendly solutions from data center leaders and service providers like Capgemini. Capgemini’s new data center is being established in Swindon in an old warehouse using 33,000 square feet of space. September 17 is set to be the data center’s start date.

The first phase will have only four modules at 10,000 square feet at an investment of around $18.5 million dollars. Overall investment is approximately $43 million. The data center should achieve 100% occupancy in another year.

The technical floor space is spread across 12 modules. The U.K. Environment Agency will occupy two of these modules—talk about a solid endorsement for the data center’s “greenness”! Each module is 2,500 square meters in size.

The modularity of the data center was a prime requirement for the company according to Paul Anderson, program director of infrastructure outsourcing at Capgemini.

The widely accepted standard to measure data center efficiency is PUE, or Power Usage Efficiency. A PUE of 2.0 is par for the course. The lower the PUE, the greener the data center. The Capgemini data center is operating at a PUE of 1.08, the best (lowest) in Europe.

The Swindon data center is able to operate at this low PUE because of several factors.
Super-efficient cooling is assured through a mix of fresh outside air and evaporative cooling. This combination results in a huge reduction in cooling power usage—92% compared with conventional data centers that use chilled water to cool their infrastructure and 75% compared with data centers that use only free air cooling.

Air is cooled in five stages. The first is a filter, followed by the filter bags and then a direct-expansion chilling system and an adiabatic pre-evaporative cooling system. The last stage is a series of fans. The direct-expansion stage is powered only if the controls detect a higher-than-acceptable temperature.

Electrical efficiency is enhanced because Capgemini has done away with battery-based UPS systems.

A super-refined climate control system constantly scans air temperature and adjusts automatically to ensure that hot air is circulated and cold air is not lost. Each module is cooled by a dedicated air optimizer, climate control, and modular BMS.

The data center is also in an elite club of facilities that have been certified as Tier 3 by Uptime Institute for their security, green operations, and resilience. Tier 3 certification has been awarded to only four data center sites in the U.K.

The PUE has been tested across various loads, ambient operating temperatures, and humidity. By comparison, the data center at Cobalt Park in Newcastle has a PUE of less than 1.3, and another data center in a Wiltshire stone mine is aiming for a PUE of 1.2.

Interestingly, the Capgemini data center is named Merlin. Environmentalists will consider it magical if it delivers on its promise of being the world’s greenest data center.
Article by Rakesh Dogra for .

Infected USB flash drive led to worst U.S. military breach in history

Paul Mah, FierceCIO TechWatch — Tue, 31 Aug 2010 04:00:00 GMT

Malware loaded into USB flash drives resulted in the successful infiltration of the U.S. military's Central Command in 2008, says a new report by The Washington Post. In what was deemed as the U.S. military's "most significant breach," Deputy Defense Secretary William J. Lynn III confirmed that the malicious code was placed in the drive by a foreign intelligence, and managed to spread undetected onto classified systems.

"That code spread undetected on both classified and unclassified systems, establishing what amounted to a digital beachhead," says Lynn in a Foreign Affairs article. The data could then be transferred to servers under the control of foreign governments. The decision by Lynn to declassify the incident reflects the Pentagon's desire to raise congressional and public awareness over the asymmetrical nature of cyber warfare facing U.S. computer systems today, say experts.

Elaborating on the seriousness of the successful infiltration of the command and control system, a former intelligence official who spoke on condition of anonymity told the Post: "This is how we order people to go to war. If you're on the inside, you can change orders. You can say, 'turn left' instead of 'turn right.' You can say 'go up' instead of 'go down.' "

Article by Paul Mah at .

Today's Facility Engineer: Jack of All Data Center Trades

Jeffrey Clark, DATACENTER Journal — Tue, 31 Aug 2010 04:00:00 GMT

Despite the faltering economy, demand for IT services continues to rise. Correspondingly, the need for resources provided by data centers is increasing, which is good news for IT professionals seeking employment in an overall dismal job market. For data center facility engineer positions, the question that many such IT professionals must ask themselves is whether they have the skill set that employers are looking for.

Today’s data center facility engineer must wear a number of hats. Although specialization is the key to success in many jobs, most companies require their facility engineers to perform a number of different types of tasks. Naturally, then, the engineer must have a broad skill set. To get an idea of what companies are looking for, simply scan the job advertisements on any job board, or use an Internet search engine. The job descriptions are often nearly identical.

The facility engineer must, of course, be familiar with the electrical, mechanical, and other data center equipment. In addition to simply having proficiency in this area, many facility engineers must also be able to train other data center staff in related areas. In other words, IT professionals looking to fill this type of position must often be both master and teacher.

Naturally, one person cannot do all the work alone, especially in emergency situations when equipment fails or uptime is threatened. This is particularly true in large data centers. Thus, the facility engineer must be able to supervise both other company employees as well as contractors that may be called in to help resolve the situation. Of course, these skills apply not only to emergency situations, but also to everyday maintenance and repairs that may not be as critical.

At an interview for a facility engineer, a potential employee just might be queried about his social life—and not to determine if the applicant is a well-rounded person. Many data centers simply do not have the option to put off needed repairs for a day or two, or even until first thing the next morning. Mission-critical facilities must be operational at all times—24 hours a day, 365 days a year. Having responsibility for maintaining a data center facility often means being the one receiving that late-night call when something goes wrong. What is often not a part of the job description for facility engineers is the ability to handle the stress that arises in an environment that virtually demands perfection at every moment.

The ideal facility engineer is one who is able to foresee problems and prevent them, rather than simply respond to them at all hours of the night. Although most job advertisements (regardless of occupation) state that the employer desires an employee that demonstrates attention to detail, the facility engineer really must have this characteristic. The ability to recognize minor concerns that could become full-blown problems threatening data center uptime is critical, and it can also save the engineer from those emergency phone calls.

Expert, manager, teacher, and on-call professional: the data center facility engineer must fill a number of roles. Need to know what types of skills are needed to be one? Just look at almost any job advertisement for such a position, and you’ll see listed at least a good portion of the skills that most companies are looking for in this regard.

Article by Jeffrey Clark for .

Hurricane Earl may test IT teleworkers

Patrick Thibodeau, ComputerWorld — Tue, 31 Aug 2010 04:00:00 GMT

If Hurricane Earl, now a major hurricane, hits the East Coast of the U.S. later this week, the top concern for IT executives may not be data center outages but loss of Internet access for telecommuting workers.

Forecasters say the storm could possibly hit land somewhere between the Carolinas and New England sometime before the start of Labor Day weekend.

Critical data centers, with backup generators, facilities and fuel supplies, are now built to continue operating during storms. The same can't be said for the computing setups that telecommuters maintain in their homes, and they may be put to the test this year.

Last year, a lack of hurricanes made it a good one for telecommuters.

There were only three hurricanes in U.S. waters last year, and none of them brought hurricane force winds over land in this country, according to Eaton Corp., a power management company that has been tracking power outages nationally since 2008. The company compiles what it calls a "Blackout Tracker" based on data gleaned from storm stories from news services and in newspapers, and from personal accounts.

In 2009, there were an average of 236 power outages a month in the U.S., said Mike DeCamp, an Eaton spokesman who also works on the Blackout Tracker. Through July, the average had increased to 273 a month for 2010, he added.

DeCamp attributes the increase "to an aging power infrastructure subjected to a daily onslaught of outages." He noted that "weather-related outages have jumped from about 77 per month to 99 per month this year."

Chuck Wilsker, president and CEO of The Telework Coalition and a telecommuter who works from his home in Washington suburb Montgomery County, said he has already had to deal with multiple major storm-related power outages this summer.

Wilsker said he has prepared his home office to cope with power outages, storm-related and otherwise.

For example, Wilsker keeps two batteries each for his laptop and his BlackBerry, and when a storm like Hurricane Earl poses a threat, he makes sure the backups are charged.

During an outage, as the batteries run down, Wilsker plugs a DC-to-AC power inverter into his car and then uses it to recharge his laptop and cell phone. He maintains Internet service via his BlackBerry smartphone device and tethers it to his laptop.

He also has a battery-powered desk light and a wired telephone that operates off the line current. A Xantrex portable battery provides additional power, he said.

"I will be more prepared than most people I know because this is my life -- I work from home," said Wilsker.

The need for teleworkers to be self-sufficient (and less dependent on coffee shops and local libraries for wireless access) is growing.

In a report released last month, the Metropolitan Washington Council of Governments estimated that there as may be as many as 600,000 workers, or about 25% of the region's workforce, who telework at least one day a week.

The council also discovered, via a telephone survey of more than 6,000 area workers, that the number of teleworkers could rise by 500,000 over the next few years.

Teleworkers in the Washington metropolitan area have faced several significant outages already this year even though they have yet to deal with a hurricane.

There were multiple storm related outages in July and two in August. The first resulted in power losses to nearly 300,000 Potomac Electric Power Co. (Pepco) customers. About 75,000 customers lost service in the second storm, and 98,000 in the third. There have also been "complaints of frequent and apparently inexplicable outages occurring outside of storm events," said the Public Service Commission of Maryland, which is now holding hearings on the outages.

Pepco has defended its performance and says that more than 90% of the outages from the storms were caused by trees and limbs falling on lines.

But telecommuting has paid dividends, at least for the federal government.

When blizzards early this year prompted a multiday shutdown of federal offices, "many federal employees rose to the challenge and continued to work, making good use of telework and other work flexibilities," testified John Berry, director of the U.S. Office of Personnel Management, before a congressional committee earlier this year.

Article by Patrick Thibodeau at .

Purdue builds app that slows servers when cooling fails

Joab Jackson, ComputerWorld — Tue, 31 Aug 2010 04:00:00 GMT

While chip manufacturers continue to make their processors ever more powerful, at least one customer has found it useful to slow these chips down, at least long enough to keep them running when the data center air conditioning falters.

Patrick Finnegan, a systems administrator at Purdue University, has developed software that slows the clock speed of server processors, a throttling that reduces the heat they produce.

"Previously our only options were to put in a few large fans and hope that was enough, or start turning servers off," said Mike Shuey, who oversees Purdue's supercomputers. "This software gives us a middle ground that gets us by many outages."

Purdue is now reselling the software for US$250, through FolioDirect, an online e-commerce service for educational institutions.

With most commodity servers, once their ambient temperatures reaches a certain point, usually around 32 Celsius (About 90 degrees Fahrenheit), they will automatically shut off to prevent damage from overheating. Smart administrators will turn them off ahead of that, at least to facilitate a graceful shutdown.

In the world of academic supercomputing these restarts can be deadly, though. Purdue's clusters run many serial jobs that can take days, weeks, or even months to complete. And while some programs have frequent setpoints to which they can return that are close to where they at shutdown, many do not. One Purdue researcher, for instance runs atmospheric climate models that can require four months of continuous computing time.

"If our only recourse to survive an outage is to start turning off machines, we can throw away from two to three million[m] CPU hours of work," Shuey said. "It can take weeks and weeks of run time just to get back to the state we were in the minute before we turned things off."

In contrast, by throttling back the servers, the programs are slowed, but no work is lost.

Finnegan built the software using a clock frequency scaling driver available for the Linux kernel, which can control both Intel and AMD chipsets with frequency scaling capabilities. The software also relies on Altair job scheduling software/a> as well as a set of cluster management tools from the U.S. Department of Energy's Oak Ridge National Laboratory.

As far as Shuey knows, no other software is available to do this task, either open source or commercial, at least for large clusters of servers.

Overall, the Purdue data center runs around 15,000 processors, mostly across two supercomputer clusters.. One, called Coates, supplied by Hewlett-Packard, runs just under 8,000 processors from AMD. The other, a Dell-supplied configuration nicknamed Steele, runs 5,600 Intel processors.

The Purdue team estimates that power usage by processors can be cut by as much as 10 percent on Intel processors and by as much as 30 percent on AMD processors. The amount of power a server uses usually directly correlates to the amount of cooling needed.

Article by Joab Jackson at .

The Rise Of Virtualized IT

Sixto Ortiz Jr., Processor.com — Tue, 31 Aug 2010 04:00:00 GMT

From Humble Beginnings In Mainframe Computers, Virtualization Is Now A Force To Be Reckoned With

Key Points

• Computing has come full circle, from centralized mainframes to decentralized client-server computing back to a centralized computing model enabled by virtualization.

• As virtualization evolves to embrace other hardware elements, new challenges in terms of policies, security, and management will inevitably arise.

• The evolution of cloud computing and virtualization go hand in hand; virtualization is the engine that drives cloud computing.

Today, virtualization is helping companies realize capital and operational expense savings by enabling the most efficient use of current data center hardware. The technology is also having a tremendous influence on all meaningful areas of IT, namely servers, storage, and networking.

Virtualization’s influence will only grow as more and more companies embrace the technology’s promise to enhance the efficiency of their IT operations. Read on to learn more about virtualization’s evolution, its status today, and where experts expect it to go in the future.

Virtualization Yesterday

Most experts agree virtualization had its genesis during the mainframe era; back in the 1960s, IBM successfully applied virtualization technology to split centralized mainframe computers into separate virtual machines, says Julian Box, CTO at Virtustream (www.virtustream.com). But, adds Box, the technology lost steam during the ’80s and ’90s as client-server technologies flowered and gained traction with the advent of inexpensive servers and desktop computing.

Virtualization started to re-emerge as a viable technology in 1998, when VMware patented its virtualization technology for IA-32/x86 systems and then launched its Virtual Platform product in 1999, says Charles King, principal analyst at Pund-IT. At first, virtualization began its comeback in the province of software developers, who needed the capability to code software for multiple operating systems without having to invest in separate hardware platforms for each OS, King explains. Virtualization delivered the ability to host multiple OSes in a single hardware platform, thus delivering significant savings for software developers.

In 2003, EMC’s acquisition of VMware, along with the launch of the open-source Xen hypervisor and Microsoft’s purchase of virtualization assets from Connectrix, launched virtualization into its next phase: maximizing x86 server utilization by consolidating multiple OSes, applications, and workloads onto fewer machines, King says. This issue has become increasingly critical as skyrocketing hardware capabilities have resulted in vast underutilization of servers hosting typical business applications.

Virtualization Today

Today, virtualization continues to gain traction in a growing number of enterprises. In recent years, IT has undergone a behavioral shift as administrators get away from having to know the specific servers and resources used by specific applications and embrace a more flexible model where “hard” configurations are not required and virtualized servers can be created, moved, and decommissioned with ease, says Ken Oestreich, vice president of marketing at Egenera (www.egenera.com). As IT becomes more at ease with this computing model, he adds, uses for virtualization continue to grow.

The operational benefits of virtualization are beginning to take center stage, says Renata Budko, co-founder and marketing vice president at HyTrust (www.hytrust.com). New tools are enabling organizations to reduce server provisioning time, simplify business continuity strategy, dynamically optimize resource allocation, and automate many of the operations that previously required a physical presence, adds Budko. In short, organizations are using virtualization to implement agile, dynamic information systems.

In terms of network virtualization, the latest advances involve the virtualization of networking components, such as switches, routers, or firewalls, says Michael Hamelin, chief security architect for Tufin Technologies (www.tufin.com). Hamelin says some vendors are implementing multiple virtual firewalls within a single chassis, allowing customers to consolidate their physical environment and reduce both administrative overhead and physical cabling. Another trend, he adds, is the complete virtualization of the network or security device by running it on a hypervisor.

Pund-IT’s King says even though broad adoption of virtualization by SMEs lagged behind enterprises due to the technical skills required to implement and support it, the situation has changed over the past couple of years due to increasingly simplified and automated management solutions. Companies such as VMware and Microsoft have been instrumental in pushing significantly broader adoption of virtualization among SMEs. Given virtualization’s essential role in cloud environments, SMEs using SaaS and other cloud-based services are probably reaping the benefits of virtualization whether they’re aware of it or not, King says.

Virtualization Tomorrow

There is no doubt virtualization has a bright future as the technology continues to evolve. But how will this evolution take place? In terms of storage virtualization, the future means the complete removal of provisioning, says Dr. Geoff Barrall, CTO and vice president of engineering at Overland Storage (www.overlandstorage.com). The next wave of storage systems will remove the current state of “array-level dependency” and facilitate the creation of systems that closely resemble VMware virtualized server environments, Barrall says. In other words, virtual disks will be able to be migrated in real time between storage systems without interruption in service, making it possible to create any number of virtual disks or file servers, not bound by the underlying physical hardware.

Jason Dea, product marketing manager at Novell (www.novell.com), says now that the cost savings of virtualization have been realized and acknowledged, the focus is shifting to the promise of cloud computing, enabled by virtualization, as a better way for administrators to run workloads and potentially run new workloads that were not possible with yesterday’s technology platforms.

Egenera’s Oestreich foresees a continuation of the OS virtualization evolution as more functions are virtualized and more operations are automated. In fact, Oestreich believes today’s virtualization management products may eventually evolve into tomorrow’s data center management products. A new phase of virtualization, he adds, involves the virtualization of the “other half” of the data center. In other words, virtualization technology will move beyond virtualizing the OS and into I/O virtualization, network convergence, and virtual switching and storage addressing. Combined with OS virtualization, infrastructure virtualization (i.e., converged infrastructure/unified computing) will help virtualize the entire data center, not just the software, Oestreich says.

“This is the nirvana many IT people have been waiting for—the ability to logically compose an application, from the software down to the network and out to the storage,” Oestreich says.

Article by Sixto Ortiz Jr. for .

Effort to integrate security into heart of computing overdue

Caron Carlson, FierceCIO — Tue, 31 Aug 2010 04:00:00 GMT

Intel's (NASDAQ: INTC) move to acquire McAfee for $7.68 billion last week caught many industry observers by surprise, and left many others--particularly financial analysts--unsure of the real purpose behind the acquisition. (We posted a brief run down of some of the initial reaction.) As the dust settled around the news, though, the possible implications for improved network security started to become clearer.

While some analysts and pundits saw the purchase of the security software vendor as an effort by Intel to make inroads in the mobile device market, others took it more at face value: A much-needed move to integrate security into the heart of computing. As threats grow in volume and sophistication, enterprises need much more powerful security technology integrated into their systems, but there hasn't been anyone in the security market with the size, scope and reach to accomplish this.

Perhaps it shouldn't come as a surprise that Wall Street analysts were (like so many headlines said) scratching their heads over the announcement, while tech veterans tended to see it as a sensible, positive move. Security remains a hard sell among the top brass; if you are more focused on the next quarter's bottom line than the long-term viability of the corporate network and information assets, it can be tough to see the value of investing in something to prevent unknown catastrophes.

But if you are focused on the imperative of securing the network and corporate data, you've probably been waiting for a long time for someone like Intel to buy someone like McAfee. And you probably won't be taken off guard to see other major vendors follow Intel's lead in the not-so-distant future.
Article by Caron Carlson at .

U.S. Cyber-security Leads Week in Security News

Brian Prince, eWeek.com — Tue, 31 Aug 2010 04:00:00 GMT

A recap of some of the major IT security news of the past week: Leading the way is a Pentagon official's discussion of America's cyber-security strategy and a 2008 incident he called the "most significant breach of U.S. military computers ever."

U.S. cyber-security was in the spotlight this past week when a senior Pentagon official confirmed an infected flash drive caused the "most significant breach of U.S. military computers ever."

Deputy Secretary of Defense William Lynn III discussed the 2008 incident in an article for Foreign Affairs magazine.
"That code spread undetected on both classified and unclassified systems, establishing what amounted to a digital beachhead, from which data could be transferred to servers under foreign control," Lynn writes. "It was a network administrator's worst fear: a rogue program operating silently, poised to deliver operational plans into the hands of an unknown adversary."

Lynn also discussed U.S. cyber-security strategy, and called for public and private partnerships to secure cyber-space.
On the subject of threats, the number of exploits circulating for the .DLL loading issue affecting scores of applications boomed during the week. Accompanying the increase in available exploits were revelations about some of the vulnerable programs, which include apps such as Adobe Photoshop and Microsoft Word 2007.

Actually fixing affected applications does not appear to be overly difficult, but the sheer number of vulnerable applications poses a challenge, Marc Fossi, manager of research and development for Symantec Security Response, told eWEEK. Raising awareness of among application developers is going to be a major hurdle, he added.

"According to Microsoft, directly addressing this issue in Windows will result in the loss of some expected functionality," he said. "As a result, they are recommending that the onus be on application developers to fix it. However, we encourage Microsoft to continue to look at ways to address this issue from a higher level."

A new threat report from Symantec named Rustock the most dominant spam botnet on the Web—a title it held despite a significant drop-off in the number of bots under its control.

On the subject of spam, researchers at Sophos and IBM noted an uptick of spam targeting users with malicious zip file attachments. Those who make the mistake of opening the attachments have been hit with everything from Trojan downloaders to variants of the infamous Zeus malware.

Researchers at IBM’s X-Force also released a sweeping report during the week that found the number of vulnerability disclosures during the first half of the year shot up 36 percent compared with the first six months of 2009. The total number of bugs disclosed during the period was 4,396.

Leading the way with the most vulnerabilities was Apple, with Microsoft coming in at No. 2 and Adobe Systems at No. 3.

"The leap in vulnerability disclosures relates to organizations taking a greater interest in exploitable software bugs as well as attackers continuing to develop their own infrastructure," Tom Cross, manager of IBM's X-Force Advanced Research Team, said in an interview with eWEEK. "An area that both whitehat and blackhat security researchers are focusing on is automated vulnerability discovery through approaches such as fuzzing. Predicting disclosure increases into the future is going to be tricky for this reason, and we may see the occasional plateau or decrease."

Article by Brian Prince for

Meeting Future Storage Needs

Dan Heilman, Processor.com — Tue, 31 Aug 2010 04:00:00 GMT

There Are Several Technologies To Consider As You Look Ahead

Key Points

• Make scalability your watchword. Data continues to grow unabated, and its rate of growth is often proportional to the growth of the organization.

• Technologies such as iSCSI, storage tiering, thin provisioning, and virtualization can make it easier for an SME to scale its storage and meet growing needs.

• Avoid one-size-fits-all solutions and instead consider varying vendors. Locking into one can lead to unnecessary expense and fluctuations in quality.

Moore’s Law famously states that the capacity of computing hardware doubles about every two years, and that rule of thumb hasn’t gone wrong yet. For your data center, though, that can mean a headache about every two years: You need to not only meet your enterprise’s current needs, but make sure the storage setup you create lasts well into the future.

How can you best ensure that your storage setup will meet your needs for years to come? In short, the key word is scalability. Data continues to grow unabated, and its rate of growth is often proportional to the growth of the organization.

“Scalability is essential to meet many of the growth and consolidation requirements of today’s data centers,” says Jim Smid, data center practice manager for Iron Bow Technologies (www.ironbow.com). “Luckily, scalable means different things in today’s storage marketplace and with the advent of more modular storage, the ability to scale all resources incrementally is fundamental to prolonging the longevity of the investment.”

Consider Connectivity Options

EExperts say that by using some specific technologies, enterprises can future-proof their storage and give the rest of the data center a fluid and flexible platform on which to grow. Storage-area networks are a given in many data centers, but IT managers should look to have SANs that easily connect to Fibre Channel or iSCSI server interfaces on the front end, says Scott DesBles, director of technical solutions for Compellent Technologies (www.compellent.com).

According to DesBles, iSCSI lets admins use Ethernet-based networks to access and transport storage. And, he adds, less costly storage connectivity options are vital to enterprises that want to add new devices but are short on budget resources and staff. “iSCSI is an ideal interconnect because it utilizes an organization’s existing IP infrastructure and does not require specialized skills or training,” he says.

If a company that currently uses iSCSI eventually wants to move to Fibre Channel, he adds, it should be able to do so without a major upgrade.

Implement Tiering

Automated tiered storage lets enterprises gain the flexibility to not only handle current storage needs, but also adapt to changing requirements as data capacity grows.

“Effective automated tiered storage applications give data center managers the ability to mix and match the drive technologies for an optimal configuration,” DesBles says.

For instance, if a company requires high-performance storage for transaction-heavy applications, it can simply add in SSDs (solid-state drives) to handle mission-critical data while the automated tiered storage technology moves infrequently accessed data to lower tiers of storage, which can be made up of FC, SAS, or SATA drives.

“As more data moves from the high-performance drives to the lower tiers, IT managers will need to purchase fewer expensive, high-performance drives,” DesBles says.

Think Thin

Another key initiative to look into is thin provisioning. Previously, enterprises were forced to guess at what their storage needs would be over time and often ended up purchasing too much or too little storage for their future needs. Thin provisioning lets them eliminate preallocated but unused capacity. As utilization reaches 100%, you can simply add in more drives while avoiding downtime.

Storage thin provisioning presents a certain amount of storage space to the operating system or application, but unlike traditional storage systems, the capacity that is not used by the OS or application can be redistributed.

“I like to think of thin provisioning in the same light as the overbooking concept in an airline,” says Marc Howard, solutions architect at All Covered (www.allcovered.com). “Traditionally, [a company] purchases disks by estimating its capacity over the next several years. Usually the total capacity is over-estimated so that there is plenty of space available, even if never used. It’s that never-used space that often ends up as waste by the time those hard drives are fully depreciated.”

Go Virtual

Quickly growing in popularity among enterprises that are trying to win the guessing game of storage needs is storage virtualization. Server virtualization provides a lot of performance and efficiency benefits to the data center, but to truly take advantage of all that the virtual data center has to offer, experts say enterprises should pair it with a storage virtualization technology.

“To ensure flexibility for future needs, the storage virtualization software implemented must support any server virtualization application without the need for server-side agents,” Compellent’s DesBles says. “Through this strategy, enterprises won’t be locked in to any specific server virtualization technology, and they can move to different applications knowing that their storage will integrate seamlessly no matter what.”

DesBles says enterprises should be looking for a solution that allows them to create any size virtual volume within a short time period that can support most virtual server platforms and optimize the placement of virtual applications. That should keep you from having to allocate a drive to a specific server or deal with issues around capacity planning or performance tuning.

Plan Carefully

Some prudent steps enterprises can take to be ready for future storage needs involve technology less than business acumen and common sense. For instance, before jumping in and designing a storage solution, IT managers should assess access rates, performance, and data protection needs.

“One of the common mistakes made is to treat all data the same and create some sort of holistic one-size-fits-all storage solution that gets them into a nonscalable, vendor lock-in situation,” says Rob Sims, CEO of Crossroads Systems (www.crossroads.com).

Vendor balance is important to keep in mind when shopping for storage. The path of least resistance is going with a single vendor, or the “one throat to choke” approach. But Sims warns that this approach can lead to a vendor lock-in situation, and subsequent purchases" and, more importantly, support and maintenance" can be costly.

"Some of the smaller vendors offer high-quality solutions that many of the main providers actually OEM from the small vendor," he says. “You can get better pricing" and in many cases, better service "from the original vendor."

Article by Dan Heilman for .

Stupid data center tricks

Cara Garretson, ComputerWorld — Tue, 17 Aug 2010 04:00:00 GMT

HTML clipboard

Sure, technology causes its share of headaches, but human error accounts for roughly 70% of all data-center problems.

A university network brought to its knees when someone inadvertently plugged two network cables into the wrong hub. An employee injured after an ill-timed entry into a data center. Overheated systems shut down after a worker changes a data center thermostat setting from Fahrenheit to Celsius.

These are just a few of the data center disasters that have been caused not by technological malfunctions or natural catastrophes, but by human error.

According to the Uptime Institute, a New York-based research and consulting organization that focuses on data-center performance, human error causes roughly 70% of the problems that plague data centers today. The group analyzed 4,500 data-center incidents, including 400 full downtime events, says Julian Kudritzki, a vice president at the Uptime Institute, which recently published a set of guidelines for operational sustainability of data centers.

"I'm not surprised," Kudritzki says of the findings. "The management of operations is your greatest vulnerability, but also is a significant opportunity to avoid downtime. The good news is people can be retrained."

Whether it's due to neglect, insufficient training, end-user interference, tight purse strings or simple mistakes, human error is unavoidable. And these days, thanks to the ever-increasing complexity of IT systems -- and the related problem of increasingly overworked data center staffers -- even the mishaps that can be avoided often aren't, says Charles King, an analyst at Pund-IT Inc.

"Whenever you mix high levels of complexity and overwork, the results are typically ugly," says King. And as companies become more reliant on technology to achieve their business goals, those mistakes become more critical and more costly.

Wrong worker, wrong cable
Take the example of the university data center switch that overloaded because an IT worker mistakenly plugged two network cables into a downstream hub. That happened about four years ago at the Indiana University School of Medicine in Indianapolis, according to Jeramy Jay Bowers, a security analyst at the school.

The problem arose out of less-than-optimal network design, says Bowers, who worked at the school as a system engineer at the time of the incident. The IT department for the school of medicine was split into two locations, with one room in the school of medicine building and another room at the neighboring university hospital -- not an ideal setup to begin with, says Bowers.

TThe department had run fiber -- a purple cable, to be exact -- from a switch in the first building to the second, running it up through the ceiling, through a set of doors and across to the hospital's administrative wing next door. That cable attached to a 12-port switch that sat in the hospital building's IT room, and staffers could easily disconnect from the school of medicine network and connect to the hospital network through a jack in the wall, Bowers explains.

Article Cara Garretson at .

Efficient Equipment Setup and Installation

Chris A. MacKinnon, Processor.com — Tue, 17 Aug 2010 04:00:00 GMT

Track & Manage Equipment Before & After It’s Installed

Key Points

• Data centers operate more efficiently when IT and data center managers understand how the current data center is being run in order to determine the best location for new equipment.

• Data center managers should have the environment professionally cleaned before turning on any equipment in a new or expanded data center space.

• It is a good idea to use management tools and performance metrics to ensure that new equipment is properly installed and monitored in order to meet efficiency objectives and comply with SLAs.

There’s a way around mountains of cable, dust, and data center clutter. Verbs such as tracking, measuring, establishing, and ensuring are the key to a well-organized data center. When you get down to the foundation of such a data center, you find equipment setup and installation among the main pillars. But being set up, installed, and organized all at the same time has its challenges. Here are a few tips to help you achieve your organizational goals.

Track The Data Center’s Configuration

Traci Yarbrough, director of product marketing for data center software solutions at Emerson Network Power (www.emerson.com), says that amid ever-changing business demands, data center managers are constantly dealing with the need to add equipment to improve the networks, servers, power infrastructures, or cooling systems while simultaneously optimizing efficiency. But Yarbrough says that before a data center can operate efficiently, IT and facility managers—both of whom are heavily involved in data center management—must first understand how the current data center is being run in order to determine the best location for new equipment.

“To receive a full, comprehensive view of how the data center operates, IT and facility managers must be able to effectively track the data center’s configuration to analyze the interconnections and relationships between the data center components,” she explains. “This holistic view of the data center is essential to determining the optimal location of the new equipment, so as not to negatively impact systems already in place, and to maximize the efficiency and performance of new deployments.”

Keep It Clean

Kevin Vickery, president of ProSource Mission Critical Services (www.team-prosource.com), says that before turning on any equipment in a new or expanded data center space, data center and IT managers should ensure that the environment is cleaned by a professional data center cleaning contractor that can safely and effectively remove contamination that could adversely affect hardware performance.

Vickery says this includes cleaning below and above the raised floor. “Take care to ensure that the concrete subfloor plenum surface has been encapsulated to provide for a clean environment for your equipment to operate,” he says. “Data center cleaning professionals can provide this service during the construction phase prior to equipment and cabling installation. Dust created by unsealed concrete is one of the major sources of contamination in raised-floor data centers.”

He continues, “Ensure that all openings in the raised floor have been sealed to ensure that all cooling comes through airflow panels to prevent bypass airflow and properly cool your equipment. This will ensure that your cooling system operates as effectively and efficiently as possible, thereby saving energy.”

Vickery says it also helps to ask the cleaning professional to provide an ongoing maintenance plan to help ensure that your critical environment maintains the appropriate level of cleanliness for your hardware to operate effectively and efficiently over time.

Enforce Policies With Management Tools

When it is time to conduct equipment installation, data center management will have already determined whether it is best to implement changes as an independent solution or as a synergistic one that incorporates other systems in the IT service management domain, such as finance, Yarbrough says. “A well-organized data center will have a process management tool in place for both types of solutions,” she says. “Process management tools are generally configurable and directly notify the appropriate people and management staff responsible for specific tasks at appropriate times during the equipment installation.”

Yarbrough says the process management tool should also help to reduce the risks associated with new equipment changes. “By automating the equipment installation process, the amount of time spent per installation will be significantly reduced, saving money,” she says.

Measure & Track Key SLA Metrics

In Yarbrough’s opinion, today’s data centers are evaluated based on their ability to execute company objectives, keep pace with changing technology, and meet the demands for greater efficiency. She says a well-organized data center will have a way to track its progress against the specific SLA metrics outlined for that data center. “Before the installation process begins and immediately after it is complete, present ideal times for the data center manager to note the status of the data center and document the progress, as well as compare it to the SLA metrics,” she says. “The results of the installation should reaffirm the initial goals and aid the organization in their everyday operations.”

Yarbrough says that even without the act of a new installation, data center managers should also have a system in place to look for ways to maintain and improve IT service quality through a constant cycle of aggregating, monitoring, and reporting system information. Equally important, she says, each data center should set its own schedule of evaluation to optimize performance.

Article by Chris A. MacKinnon for .

Supercomputing in the cloud gains momentum

Todd R. Weiss, ComputerWorld — Tue, 17 Aug 2010 04:00:00 GMT

HTML clipboard

In a flagging economy, companies are slowly turning to on-demand services to get high-performance compute cycles without having to pay for high-end hardware.

An economy that continues to stagnate could prove a boon to an increasing number of providers of on-demand supercomputing capacity.
The market for such services has so far grown slowly, said Charles King, an analyst at Pund-IT Inc. "But a weak, rocky economy makes on-demand services a lot more affordable than purchasing and maintaining a dedicated supercomputing cluster," he added.

Dan Olds, an analyst at Gabriel Consulting Group Inc., said on-demand supercomputer services can prove particularly useful to smaller businesses that need powerful processing but can't afford to buy high-performance computing systems.

More than 25 such businesses nationwide have directly participated in the Ohio Supercomputer Center's (OSC) Blue Collar Computing initiative, which sells supercomputing services to companies that have never used such high-end hardware. More than 250 additional companies have also used the services by accessing them through OSC partners, like the Edison Welding Institute (EWI), said Ashok Krishnamurthy, interim co-director of the Columbus, Ohio-based center.

EWI, a nonprofit industry organization, offers members access to E-Weld Predictor, a Web portal that lets users access the OSC's 1,650-node IBM supercomputer cluster, dubbed "Glenn," for tasks like simulating complex welds, Krishnamurthy said. "This simulates a whole bunch of prototypes, [cutting] the time it takes to create a welding process from six months to two weeks," he said. Typical workstations couldn't handle all of the calculations that are involved in such processes, he added.

The OSC supercomputer runs AMD Opteron multicore processors and IBM cell processors; it has a peak performance of more than 75 TFLOPS, according to OSC.

Woodward Governor Co., a maker of jet engine parts, signed a contract with a consulting firm this spring to gain access to supercomputers housed at IBM's Computing on Demand facility in Poughkeepsie, N.Y.

The contract follows Woodward's yearlong participation in a project funded by the Defense Advanced Research Projects Agency that was designed to show how small manufacturers can use high-performance computing technology to significantly speed industrial design and modeling projects.

During the pilot project, which was overseen by the University of Southern California Information Sciences Institute, Woodward used cloud-based supercomputing cycles to run complex software that quickly created and evaluated designs that could be incorporated into the production cycle without delay. The company estimated that the speedier process could lead to a savings of about $500,000 annually.

In addition to IBM, other companies that offer on-demand supercomputing include Silicon Graphics International Corp., which in February launched a service called Cyclone. And Hewlett-Packard Co., Intel Corp., Yahoo Inc. and others are in the process of testing a new public cloud dubbed Open Cirrus.

Article by Todd R. Weiss at .

Advanced Threats Are A Growing Problem

Patrick Kean, Processor.com — Tue, 17 Aug 2010 04:00:00 GMT

HTML clipboard Advanced threats against enterprises are widespread and on the increase, and enterprises are having a difficult time dealing with them. That’s one of the major findings in a recent study conducted by Ponemon Institute and sponsored by NetWitness.

The institute surveyed 591 IT security managers in the United States and found that 83% of respondents believe their organizations have recently been the targets of advanced threats, with 71% saying they have seen an increase in such threats over the past year.

Mike Spinney, senior privacy analyst at Ponemon, cites several reasons for the increase in threats, including the possibility of carelessness on the part of employees when it comes to using portable media, using their workplace assets in an unprotected environment, or downloading applications to a computer at work.

Attack Uncertainty

Uncertainty about the attacks among study participants was evident, with 41% reporting that they were unable to determine how frequently they were targeted by advanced threats.

Less than a quarter of the respondents feel that prevention or quick detection of advanced threats is a top priority in their organizations. Further, only 19% feel their IT leaders are fully aware of the situation. “If advanced threats are not recognized as a priority,” Spinney says, upper-level executives “won’t assign the appropriate resources to address the threat.”

About 58% of respondents feel they have the proper procedures in place to address the threat situation, and a third say that security-enabling technologies are adequate. When it comes to security training, 26% of survey participants report that security personnel at their enterprises are adequately trained. This could be remedied in part, Spinney says, by emphasizing the human element in threat prevention.

One of the biggest problems enterprises face is detecting advanced threats, with 80% saying it takes a day or longer for their enterprise to detect such threats. About 46% report that it takes a month or longer, leaving a major security gap. This is a problem that “adequate and up-to-date security systems” could help to prevent, Spinney says.

What can enterprises learn from all this? They need to train new and existing security teams in advanced security techniques, senior management must become educated on the seriousness of these threats, and current antivirus and intrusion detection system solutions need to be improved, according to Spinney. And new, more intelligent solutions need to be found. The keys are “awareness, education, and leadership,” Spinney says. “There is no excuse for a lack of attention to this issue.”

Article by Patrick Kean for .

Top Tips For Energy Conservation

Bruce Gain, Processor.com — Tue, 17 Aug 2010 04:00:00 GMT

HTML clipboard Use Technology To Reduce Power Consumption

Key Points

• Tracking and monitoring data center temperatures, power consumption, server loads, and other metrics can lead to power reductions by knowing how to more efficiently allocate resources.

• ASHRAE has significantly raised recommended server temperatures, so if the thermostat is constantly set at 70 degrees Fahrenheit, you are probably wasting a lot of energy.

• Data center lights don’t need to be on during most hours of the day, and using motion sensors to turn them on and off can translate into significant electricity cost savings.

The power-consumption problem is well known in the IT industry: Server loads are increasing on an industry-wide scale, and the energy needed to cool and operate the infrastructure is rising at an exponential rate. However, there is no reason to be overwhelmed or to accept that rising energy costs come with the territory. Technologies and power management processes exist that will help you not only reduce data center energy consumption but also slash costs. A few things you can put in place do not even require a major investment, but they offer major returns.

Smart Monitoring

Using sensors and other tools to tightly track temperatures, power consumption, workloads, and other metrics throughout the data center and software that analyzes the data can go a long way to reduce energy consumption. “You can use your management system to look back at the historical trends and then use it as a proactive control,” says Bill Kosik, energy and sustainability director for critical facilities services at HP (www.hp.com).

Monitoring tools, for example, can be used to locate and track applications that draw significant power resources but can be scheduled to run during off-peak hours when electricity costs are lower. When temperature hot spots are discovered in the data center, work loads can be shifted to other servers so that the air conditioners do not have to use more power to compensate for the clusters of higher temperatures, Kosik says. “You are not going to see 80% reductions [by using these methods], but you are really going to optimize your work load, and it is possible to see up to 20% in power-use reductions from better power management and workload shifting,” Kosik says. “It is not insignificant.”

Turn Down The AC

When servers began to fill up rooms in enterprises and heat levels first became an issue, the initial reaction was to just crank up the air conditioners and to keep the rooms at 70 degrees Fahrenheit or even cooler. However, blindly keeping data centers cool is not only passé, but represents a huge waste of energy. “When I first got started in this business, you pretty much had to have a jacket on every time you walked into a data center, which was more than likely going to be 55 degrees,” says Jeff Lowenberg, vice president of facilities for The Planet (www.theplanet.com).

During the past few years, engineers have discovered that servers can operate at temperatures much higher than previously thought, which means enterprises can reap significant electricity savings by simply turning down the air-cooling systems. “The new ASHRAE (American Society of Heating, Refrigerating and Air-Conditioning Engineers) standard will let you take it all the way to 81 degrees, but we try to get ours to 75 to 76 degrees,” Lowenberg says.

At the same time, worrying about the ambient temperatures of the data center is not nearly as important as the inlet server temperatures, Lowenberg says. “The only temperature that you are concerned about in the data center is the inlet air to the computers,” Lowenberg says. “You don’t care how hot your hot aisle gets, since as long as you can keep the air inlet temperatures to your computers within the ASHRAE and the manufacturers’ standards, that is all you need to do. Everything else will take care of itself.”

Water Cooling

Outside air can be used to chill water for data center cooling purposes. Even in Dallas, where temperatures are relatively hot much of the year compared to many areas of the United States, Lowenberg estimates that temperatures are cool enough 2,000 hours out of the year so that outside air can be used for low-power cooling. "All I need is a 4-degree temperature delta so that if it is 56 degrees outside, I can provide 60 degrees water to the data center,” he says. “You still have to run the pumps for the water, but that huge chiller motor, even though we use high-efficiency chillers, still consumes a lot of power, so I am taking that out of the mix.”

Though still in the planning stages, Lowenberg says he hopes his data center will see a 20% energy savings to cool the data center by using outside air. “We can use this method only about 25 to 30% of the year, but during the entire year I am going to realize close to a 20% energy savings,” he says. “The chiller probably [accounts for] 85 to 90% of the energy consumption in the cooling system when it is running.”

Use Lights & AC Only As Needed

Data center lights are usually left on, which represents a tremendous waste of energy. But why not just turn the lights on only when they are needed? One way to do that is by installing motion sensors that activate the lights whenever someone enters the data center, says Ruben Peijnenborgh, managing director of Shields Environmental (www.shields-e.com). “Some data centers keep their lights on 24/7 when they are only occupied three to four hours a day,” Peijnenborgh says. “You can achieve 40% savings in lighting electricity costs with motion sensors.”

Air-cooling system power needs fluctuate, yet the power draw usually remains constant regardless of the workloads. One way to reduce cooling power consumption is to invest in systems that consume only the amount of power that is required on an as-needed basis, Shields Environmental’s Peijnenborgh says. “The result is a 10 to 30% reduction in electricity consumption for the air conditioning,” Peijnenborgh says.

If possible, reduce the need for cooling from the beginning. Data center managers more often than not are unable to pick the location where a data center will be built. However, the differences between energy costs depending on where a data center is located in the United States can be dramatic. So if it is possible to pick a cool and not-so-humid location for your data center build instead of the balmy and wet southeast, for example, significant costs can be saved. “It has always been the case that cooler and less-humid climates [require less power] to operate [in] than hotter and more humid climates,” Kosik says.

Article by Bruce Gain for .

Computational Fluid Dynamics (CFD): Wise Investment or Unnecessary Expense?

Jeffrey Clark, DATACENTER Journal — Tue, 17 Aug 2010 04:00:00 GMT

HTML clipboard One of the keys to adequately cooling IT equipment in the data center is maintaining proper air flow throughout the facility. If heat is not removed from the vicinity of the computer equipment that produces it, that heat can build up to the point that equipment failure becomes a distinct possibility. Depending on the configuration of the data center, maintaining proper air flow may require isolation of hot and cold air in designated aisles, or any number of other approaches. Regardless of the details of the facility’s layout, accurately determining how air moves throughout the data center usually cannot be determined by simply “eyeballing” the configuration. One increasingly popular method of evaluating air flow in the data center is computational fluid dynamics—CFD. But is CFD really a necessary tool that can pay real dividends in terms of reduced cooling costs and fewer equipment failures, or is it an over-hyped expense that is full of hot air?

The answer to this question may vary from company to company and data center to data center. A small data center with a limited budget may not be able to gain the types of returns in operational expenses to justify the cost of a CFD analysis; a large multi-million-dollar data center, on the other hand, might benefit greatly from a detailed CFD analysis, despite the initial associated costs. Thus, the process of determining whether CFD is needed for a particular company’s data center may boil down to an analysis of the pros and cons of CFD.

Performing CFD analysis is not simply a matter of punching a few numbers into a calculator—it is a complex numerical process that requires fairly extensive computing power. In addition to having or developing an accurate model of the data center in question, either a company must license and learn how to use a particular CFD software package, or it must hire a firm capable of performing the analysis. Either of these options may involve significant expense. A blog by Mark Fontecchio at SearchDataCenter.com (“Can you afford data center CFD?”) raises the question of the cost of CFD modeling. As with most products, CFD software ranges in price, but lower cost may mean lower accuracy. “There are different tiers of affordability out there, however. The cheapest I’ve seen is CoolSim, which at its most basic level, costs $7,500 a year...But the most important aspect of CFD modeling is accuracy, and with a product like CoolSim, there’s a question around whether you get what you pay for.”

Needless to say, the process of increasing air flow efficiency, and thereby reducing cooling costs, is one that has a point of diminishing returns. Some companies may benefit from extremely fine details regarding the air flow in their facilities, but for others, this may be an added expense with no real potential for a significant return on investment. Fontecchio states, however, that companies that are willing to research CFD products “can quickly find...the right product for the right price.”

CFD can pose some danger to the unwary. Companies can easily fall into the trap of looking for a high-tech solution to problems that have easy-to-identify, low-tech fixes. CFD should not replace common sense and reliance on best practices when it comes to data center design and maintenance. In another SearchDataCenter.com blog (“CFD isn’t a tool to tell you to block cable cutouts, consultant says”), Fontecchio cites Terry Rodgers of data center consulting firm Syska Hennessy as noting that “CFD modeling is too expensive to be used by data center managers simply to tell them that they should be following best practices.” Paying tens of thousands of dollars to model a data center to find hot spots that can be easily measured in seconds by placing cheap thermometers in the facility is not a wise approach. Furthermore, data center operators should consult the best practices for operation of their facilities before even considering CFD modeling. That is to say, CFD should not be used to reinvent the wheel: unused cable holes that allow hot air into cold aisles, for instance, should be plugged; data center operators should not turn to a complicated and expensive computer analysis when taking well-known, industry-standard steps is enough. CFD should be an option for companies only after they have implemented industry best practices.

So, is CFD just a bunch of hot air? No, probably not. But is it a glamorous expense that some companies cannot afford? To be sure. CFD may pay off for a company looking to add an expensive new cooling system to a large data center, but it may be an extravagant waste of money for a small data center simply looking to shave some cost off its cooling bill. In either case, companies should make an effort to evaluate their needs, the potential returns of a computational analysis, and the accuracy needed in the analysis before committing to CFD and a particular CFD product.

Article by Jeffrey Clark for .

Cable Organization and Management

Robyn Weisman, Processor.com — Tue, 17 Aug 2010 04:00:00 GMT

Though Easily Overlooked, Cables Are Integral To A Well-Organized Data Center

Key Points

• Determine where your cables originate and where they run to.

• Provide detailed labels for each cable and color code groups of cable.

• A good wire management scheme keeps cables neat and out of the way of your equipment.

Cables in the data center lend themselves to a variety of fairly unflattering analogies. “The stuff piles up like books on a coffee table,” says Rudy Rangel, sales manager at IT equipment vendor Rackmount Solutions (www.rackmountsolutions.net). “It takes discipline to organize them, but it’s easier said than done.”

haphazard cables are, at best, a nuisance and, at worst, a danger. “Mismanagement of cables can lead to significant lost revenue from unnecessary downtime,” says Nick Schmidt, senior manager of IT infrastructure at CDW (www.cdw.com). “Cable management is the foundation for space maximization, overall performance, safety, and cooling efficiency.”

Putting together an organization and management plan for your cabling infrastructure is tedious and takes a lot of work, particularly if you’ve inherited several nests of tangled, lint-covered wiring from your predecessor. Block out some time during your business’ off-hours and put together a plan that keeps some of the following tips in mind.

Determine Your Sources & Isolate Them

Start out by going to your various sources, most commonly your servers, and determining what cables are running from them, Rangel says. Check out your switches and patch panels and anything else that leads to wires sticking out of your racks. Once you’ve done that, trace each cable and figure out where it runs.

Then, isolate and consolidate your sources. “If you have stuff spread over three racks and can fit them into one, that would be ideal. If you don’t need [certain] ports, and they have cables plugged into them, take them out,” Rangel says.

CDW’s Schmidt adds that you should plan on replacing up to 30% of your current cabling as you reorganize your cable infrastructure. “Cables will often stretch, kink, cut, or break during the removal phase of getting organized,” he says.

Rangel recommends that you make sure your ports are running properly and that your existing wire management can handle additional equipment. “If you add 10 people to your company, you probably are going to run at least 10 more lines. Make a note to go back to check everything,” Rangel says. “You may be running and gunning when you first set everything up, but without gradual maintenance, in six months time, you’re right back where you were.”

Label & Color-Code Each Cable

Generate a detailed label using a label printer for each patch cord and network cable so that you can easily tell where the wires are going and to what devices they connect, says Christina Hansen, product specialist at CableOrganizer.com (www.cableorganizer.com). “It’s a time investment when you’re doing it, but compared to having to dig through cables when trying to fix a problem, it’s a huge difference. If you’re walking into a situation [that’s a mess] right up front, taking care of this will make your job easier,” she says.

In addition, developing a cable color-coding plan in conjunction with labeling will aid you in quickly identifying the grouping and routing of cables for troubleshooting purposes, Schmidt says. “Color coding allows for the IT staff to identify and route cables based on purpose. For example, cables that are connecting IT security devices may be red to heighten attention during installation, while virtual server hosts cables may be blue to identify the increased importance of the host server connections,” he says.

Design A Smart Wire Management Scheme

Cables can be managed in a number of ways so that they don’t crisscross and crowd your equipment. You risk nullifying your cooling, power consumption, and overall data center performance plans without a cable management strategy. For example, cables that accumulate in the back of racks prevent servers from properly dispersing heat, while those hanging around in the front of racks complicate your ability to access servers.

Rangel recommends vertical and horizontal wire management schemes, which typically consist of channel rack mounts attached to the sides or tops of racks and cable ladders that wend their way overhead. Both methods work to keep cables away from equipment while keeping the cables themselves protected and safe. Nevertheless, there are issues with this strategy, particularly if your mass of cables is as big around as a boa constrictor.

Schmidt prefers using overhead cabling when adding new cables or accessing existing cables. “To maximize efficiency, copper, fiber, and power cabling should all be done overhead because this reduces installation costs and can accommodate growth without excessive disruption to existing equipment,” he says. “Overall, overhead cabling is easier to install and therefore easier to keep organized.”

Rackmount Solutions sells a fairly new product called Neat-Patch (www.neatpatch.com), a 2U rackmount unit that sits between patch panels in a server rack and offers an innovative way of managing and protecting cables. Neat-Patch solutions make use of patch cords that are slightly longer than necessary, providing just enough room to connect two adjacent patch panels in a rack and have a loop that protects the bend radius of the cable, eliminating attenuation. “When you’re done making connections and loops, there’s a panel that snaps on to cover the actual patch cord compartment,” Hansen explains. “You see where each connection goes, so it’s easy to connect and disconnect servers because [the cables] are laid right out there for you.”

Article by Robyn Weisman for .

Government Googled

Rakesh Dogra, DATACENTER Journal — Tue, 03 Aug 2010 04:00:00 GMT

HTML clipboard Google Inc, a behemoth in the world of cloud computing, has recently made the U.S. government one of its clients for its cloud-computing services.

The Google Apps for Government announcement came at a press event at Google headquarters after the company received FISMA (Federal Information Security Management Act) certification. To be precise, it has received the FISMA-Moderate rating, which implies that Google now has authority to store unclassified but sensitive data, which means up to 80% of government data. As Kripa Krishnan, Technical Program Manager at Google Apps for Government, clarifies, the FISMA review and certification makes for easier comparison between Google’s security features and those of Federal agencies. The company also goes on to say that most agencies have found Google’s security at least equal to their own, thus bolstering the government’s confidence in moving to the cloud. Google Apps was designed with the framework provided by Federal agencies and is a suite of cloud offerings for word processing, special email, and collaboration applications. The cloud bundle consists of Gmail, Calendar, Google Sites, Groups, Postini, Video, and Google Docs.

Google Apps for Government is a modification of Google Apps Premier Edition and goes an extra step in ensuring a separation between government mail and calendar data and that of the general populace. Such a separation is necessary, since the product is aimed at widespread usage in the face of fears following the Google Street View incident.

The new edition also offers the promise of helping the government “redeploy” resources for public use, since the company’s community cloud does not necessitate installation and maintenance of hardware and software. Making its case even stronger, Google provides assurance that the information will be stored within U.S. borders. Gmail and calendar data will be stored in a separate storage system designed specially by Google’s government contractors.

The Google cloud can also deliver better operations than the existing systems, resulting in cost savings. By pushing a portion of their IT systems into the cloud, the government can take advantage of such cost savings. Reports suggest that the current U.S. administration is looking at annual savings of $76 billion for computing and software by way of moving into the cloud. Incidentally, Federal spending on hardware and software is approximately $120 billion every year.

Google is also emphasizing to the government the potential benefits of paperless offices; these benefits include not only cost savings, but also a sleeker IT DNA structure.

The near future will also see the in-the-pipeline Google Chrome OS being updated for government use. Eric Schmidt, CEO of Google, says FISMA certification and government support is a banner of trust that has gone a long way in breaking down another barrier to seeking new and more effective data-management practices. He also said that Google would also seek FISMA certification for its Chrome OS, because all the aforementioned apps will run incredibly “well and securely” on that OS.

The U.S. government has been a client of Google in the past. Consider, for instance, Berkeley Lab’s usage of Google Apps to share emails in the cloud. Berkeley Lab is a national laboratory of the U.S. Department of Energy. Business products like Google Earth and Google Apps are also widely used by Federal agencies and state governments. New Mexico, Kansas, and Florida are just some of the states that are Google Apps clients.

Interestingly, Microsoft is hot on Google’s heels in the FISMA race. Even though it is the biggest player for the government IT industry, its Microsoft Exchange email system has not yet received FISMA endorsement. The company is seeking certification for its Business Productivity Online Suite. Microsoft also has a few aces up its sleeve, one of which is its government-only cloud facility.

The FISMA seal should also give SMBs some confidence in Google’s services. Of course, they will still have to consider the nature of the data they will entrust to Google, but a FISMA certification does not hurt

Article by Rakesh Dogra for .

6 cool innovations for the data center

John Brandon, ComputerWorld — Tue, 03 Aug 2010 04:00:00 GMT

Some of these new technologies are already on the market, and some have yet to hit -- but they all promise to make your data center operation run more smoothly.

Sure, consumer gadgets are getting most of the attention these days, but data centers are getting some love too. These new products and technologies promise to solve real data center problems or are already working to make enterprise operations run more smoothly. How many are on your wish list?

1. Fiber optics with a twist
The success of the HDMI cable in consumer electronics has proved that having a common cable that works with Blu-ray players, HDTV sets and just about any set-top box helps remove clutter and confusion. Intel has developed Light Peak following this same approach. It's fiber-optic cable that will first be used with laptop and desktop computers to reduce clutter and to speed transmission, but it could also make its way to the data center as a way to connect servers and switches.

The 3.2mm cable, which is about as thin as a USB cable, can be up to 100 feet long. Intel has designed a controller that will sit inside a computer, and cables are currently in production. Third parties, including Hewlett-Packard and Dell, will start making computers with Light Peak fiber-optic cables in them by 2011, according to Intel.

For data centers, Light Peak presents a few interesting possibilities. Fiber optics have been in the data center since the early 1990s, when IBM introduced its Escon (Enterprise Systems Connection) product line; it connects mainframes at 200Mbit/sec. Light Peak differs in that it runs at 10GB/sec., and Intel claims that the components will be less expensive and lighter-weight than existing fiber-optic products.

"Intel claims Light Peak will be less complex and easier to manage by eliminating unnecessary ports, and deliver the higher throughput required by high performance e-SATA and DisplayPort systems," says Charles King, an analyst at Pund-IT in Concord, Mass. "If the company delivers on these promises, Light Peak could simplify life for data center managers plagued by installing, managing and troubleshooting miles of unruly optical cables."

Success here will depend on "how willingly developers and vendors" embrace Light Peak and build products around it, King explains.

2. Submerged liquid cooling and horizontal racks
Liquid cooling for data centers is not a new concept, of course, but Green Revolution Cooling has added a new twist. For starters, the rack is turned on its side, which helps with cable management and makes it easier for administrators to access equipment, and the horizontal rack is surrounded by liquid. A new coolant, called GreenDEF, is made from mineral oil that is nontoxic, costs less than other liquid-cooling methods and is not electrically conductive like water, according to a GR Cooling spokesman.

"The liquid actually moves through the floor and circulates up through all of the computing nodes," says Tommy Minyard, director of advanced computing systems at the Texas Advanced Computing Center, part of the University of Texas at Austin. This means more-effective cooling because heat is moved away from the processors via cables on the sides and under the rack, he explains. Minyard is installing GR Cooling systems in his data center and expects a 30% to 40% savings compared to traditional air-cooled systems.

Green Revolution uses horizontal devices for racks, along with a new type of coolant, to reduce energy costs in a data center.
Minyard says liquid cooling has made a rebound lately, recalling the days when Cray offered submerged cooling systems, and he notes that even IBM is moving back into chilled-liquid-cooling some compute nodes.

Pund-IT's King says a major issue is that enterprises have fought the return of liquid cooling in the data center because of the high costs of implementing the technology and because it is unproven as a widespread option.

"Liquid cooling usually costs much more to install upfront than air cooling," says Mark Tlapak, GR Cooling's co-founder. "Compared to air, every liquid cooling system has some added nuance, such as electric conductivity with water-based cooling systems. " But, he says, "spring a leak in the water systems, and you lose electrical equipment." Still, for Minyard, GR Cooling is an ideal fit: His data center gravitates toward dense, powerful systems that pack intense power into small spaces, such as IBM blade servers and the latest Intel processors. The Ranger supercomputer, for example, uses 30kw of power per rack.

Next Page

Article by John Brandon at .

The Real Story Behind Storage Costs

Christian Perry, Processor.com — Tue, 03 Aug 2010 04:00:00 GMT

HTML clipboard

Shelling Out For Storage Requires More Than A Passing Glance At The Price Tag

Key Points

• Storage costs can spiral well beyond the initial acquisition as power, support, space, and maintenance add significantly to the long-term bottom line.

• Notoriously high maintenance costs for storage arrays have led many IT managers to vigilantly seek long warranties and other service options that help justify the cost of hardware.

• When negotiating with storage vendors, it’s crucial to obtain long-term estimations that provide an accurate glimpse of costs across all storage-related factors.

As hard drive costs continue their plummet to all-time lows, buying more storage to solve capacity problems has become a common practice for many enterprises. Although the low upfront cost of drives and other storage hardware might seem insignificant in the big picture, hidden costs such as power, support, space, and maintenance can easily turn the tables on what initially seemed like a smart solution.

“These factors—operational expense over time—form the majority of cost of storage, as much as 70 to 80%, according to some analysts. Therefore, these elements are a very important factor. In fact, they should be given higher weight than cost of acquisition,” says Rob Peglar, vice president of technology at Xiotech (www.xiotech.com).

Cost Considerations

Although storage might boast fewer operational headaches than other technologies, make no mistake that you’ll likely pay more for storage hardware than just its purchase cost. For example, Peglar explains that maintenance on storage arrays is notoriously expensive, leading smart IT managers to seek arrays with long warranties (such as five years) and low- or no-cost warranty service over that period and beyond. It’s also critical to remember that these costs can vary widely among vendors.

“We recently went through an RFQ [request for quotation] for new storage arrays and sent the requirements out to three vendors,” says Jason Burnett, director of network services at NeoSpire (www.neospire.net). “All three vendors came back within 10% of each other on price, but the support, maintenance options, and facility requirements—power and space—were so vastly different that we actually chose one of the more expensive manufacturers.”

For production environments that require 100% availability, it’s important to keep in mind that some storage array vendors offer only 24-hour turnaround on parts delivery and onsite technician assistance, Burnett warns. For his part, Burnett wasn’t willing to accept the risk of maintaining large amounts of data that customers depend on around the clock. Instead, he seeks storage vendors that provide regional warehousing of parts and service and a four-hour turnaround for part replacement or onsite assistance.

Intel, which manages about 18PB of primary and backup storage in its design computing, office, and enterprise environments, looks at a wide range of factors when considering storage costs, including hardware, software, power, space, management, network, maintenance (including amount and frequency of failures), and four-year hardware maintenance and software licensing costs, says Shesha Krishnapura, senior principal engineer with Intel IT Engineering (www.intel.com).

After four years, new storage purchases can be more easily justified due to the higher maintenance costs stemming from expired vendor warranties, Krishnapura says. He adds that cooling and power do factor into Intel’s total storage cost, but only at about 10%, and as the company refreshes with new technology, it benefits from more efficient power and space factors provided by those new technologies.

Scrutinize The Storage

Gauging storage costs is a simple but crucial exercise, Peglar says, because the various hidden elements can greatly affect the overall costs. For managers entering a negotiation for storage acquisitions, he recommends insisting that the vendor provides a true five-year (or greater) cost of operation with all factors included, such as growth of capacity, growth of required performance (as application workloads grow and change), and all the required “inputs” (dollars, watts, BTUs, rackspace, support/service/maintenance costs, and any software licensing and software maintenance over time).

“Evaluation is a joint exercise between the [customer] and the supplier; it is the latter that best knows the detail of the true cost of operation. Consumers should insist on open and honest disclosure of all operational costs from their vendors, over at least five years,” Peglar says.

When considering storage costs, it’s also important to consider other wide-ranging concepts and strategies outside of the straight hardware infrastructure realm, adds Cindy LaChapelle, a lead consultant with Compass (www.compassmc.com). These include disaster recovery, business continuity, data management, data availability, lifecycle management, and data migration. Although other technologies might help to reduce storage costs, keep in mind that the actual savings and overall benefit to the business (in terms of storage) can be difficult to gauge.

“Emerging technologies like virtualization, deduplication, thin provisioning, etc., help optimize storage utilization, but the tools to measure real occupancy levels on shared storage systems are limited, so the effectiveness of these technologies in really optimizing the existing storage and aligning technology performance and capacity to the business needs is still relatively unproven,” LaChapelle says.

Article by Christian Perry for

Greening Data Center Cabling

Bridget Mintz Testa, Processor.com — Tue, 03 Aug 2010 04:00:00 GMT

HTML clipboard

Save Money, Energy & The Planet

Key Points

• The key to green cabling in the data center is planning enough space for proper airflow for both data and power cables, eliminating excess airflow, and laying cable properly.

• In purchasing the cable plant, think long term and get the longest-lived cable possible. Not only will this save the cost of frequent complete cable plant change-outs, but it also reduces energy use and thereby minimizes heat in the air.

• Recycling cable goes beyond copper and includes finding proper homes for old cable jackets.
A data center is crucial to the enterprise, so running it as efficiently as possible can only benefit the enterprise. Although going “green” sometimes seems to involve spending a great deal of money with few business benefits, greening the cable plant is different.

A green cabling infrastructure can make a data center more efficient. That saves money on power and cooling, which are, of course, directly related. But, if planned properly, green cabling also helps save capital expenditures for the lifetime of the facility. And when cable must at last be replaced, some expenses can be recouped from recycling. Of course, at the same time, using less energy, keeping air temperatures cooler, and keeping toxic materials out of landfills all benefit the environment.

For both power and data cables, going green in the beginning is about physically maximizing airflow and air space, as well as getting the most robust cabling you can with the longest possible life cycle. As for recycling, don’t just think copper. It’s possible to do a lot more.

More Air Space, Less Heat

Maximizing airflow depends initially on good data center design. Later, it depends on proper maintenance of the cabling plant. “Historically, data centers were built to accommodate as much power or data cable under the floor as possible,” says Duncan Lee, CEO of PDU Cables (www.pducables.com), a supplier of UL-listed power distribution cable assemblies for data centers. “But as servers got bigger, they required more power and more power cable.” With only 6 inches of raised flooring in older data centers, there just wasn’t enough space for proper airflow.

Today, data centers built with hot/cold aisles should be designed with bigger plenums for better airflow. Hot aisles should sit right over precisely measured power cable paths. “We want the cold aisle free from power cables,” Lee says. “Lay the power cables along each hot aisle under the floor in the exact length necessary with no extra cable coiled up.” This maximizes airflow.

PDU Cables also recommends placing low-voltage data cables overhead, with high-voltage power cables under the raised floor. Sufficient air space must be planned for both types of cable so they don’t heat up, which wastes power and raises air temperature.

The best planning can be defeated by poor maintenance. “In renovations, abandoned cable is the issue,” says Brian Duval, global marketing communications manager for Siemon (www.siemon.com). “Often, people run new cable over old. The abandoned cable constricts airflow. For renovation, get the old cable out.”

Green Purchases

During the data center design phase, don’t limit plans to what’s needed right now. “Think about the technology and the physical plant as the company grows in five, 10, or 15 years,” Duval says. “Put in the longest-lived data cable you can. People need to understand how fast cabling applications progress. If you want cabling for 10 to 15 years, you should look at 10Gb core cabling.” That’s for both copper (CAT 6A or even 7 or 7A) and optical fiber (50-micrometer optical multimode 3 or 4 fiber).

It’s the same for power cables. Not only does more robust power cabling support new IT equipment demands, but it yields direct energy and money savings. Eric Bulington, chief engineer of wire and cable for Anixter (www.anixter.com), says, “One of the simplest and low-tech ways to reduce the amount of power consumed, increase sustainability, and reduce the cost in a data center is simply by increasing the size of your power conductors.”

Bulington says that switching from four-gauge to two-gauge power cables in a 240-volt 100-foot run to a high-density server cabinet can reduce power consumption by $51 per run per year. That “may not sound like much, but if you have 100 cabinets, that’s over $5,000 per year in energy savings,” he says.

Switching to the more robust cable gauge also reduces waste heat. “Look at four server cabinets running at a continuous full 100% load,” Bulington says. “By increasing from four-gauge to two-gauge, you can save 1,500 watts per hour . . . the same amount of energy that a 5,000 BTU space heater would put into the environment. That’s heat you don’t have to remove from your facility with your cooling system.”

Recycle The Cable Plant

Buying long-lived cable, whether data or power, also means there’s less to throw away. “Regardless of whether you dispose of it properly, disposing of it isn’t green,” Duval says. “It’s better to get cable with a long life cycle.”

When it is time to replace the cable plant, though, go beyond recycling copper, which is easy. “The last frontier is making sure the jacketing material is properly disposed of,” Duval says. “The jacket has a lot of value for those who can use it.”

Plastic manufacturers typically have the most use for ground jacket materials. They can also use all the individual components, such as patches and connectors. So don’t just dump these items in a landfill. “Find cable reclamation services,” Duval says. “Either have the contractor find one or ask the scrapyard that buys the copper what they’ll do with all the plastic.” If it sends the plastic to a landfill, Duval recommends finding another recycling organization. “The concept of ‘trash’ is dying,” he says. “Almost any waste can be monetized.”

Article by Bridget Mintz Testa for .