Server Specs - A SearchDataCenter.com blog

Server Specs:

 

A SearchDataCenter.com blog


The blog for all things data center, including, design and infrastructure, Unix, Linux, mainframes and x86 servers, power and cooling efficiency, information technology (IT) service management, server consolidation and virtualization and more.

Confessions of ITIL: Where do you stand?

I’m writing a story right now based on a study by the Aperture Research Institute showing the majority of data center managers have poor configuration management, and they don’t trust their infrastructure management tools.

The poll shows less than a third of data center managers are implementing ITIL; 30% are working on introducing ITIL initiatives and 9% are making plans to implement ITIL. Then there were the 20% who said they were merely investigating ITIL and 12% confessed they were not familiar with it at all.

Where do you stand with ITIL adoption?

Data centers on the brink of extinction, Sun CIO says

Will corporate data centers be obsolete in the next ten years?

That’s what Sun Microsystems  CIO Bob Worrall told an earth2tech reporter at Sun’s Eco Innovation event.

The reporter asked, “What does the data center of 2015 look like?”

Worralls response; “In 2015, the question isn’t what a corporate data center looks like, it’s, “Why do you need a data center?’” The data center of the future is nothing but a big black box of service providers. Even corporate enterprises like Sun, who have a traditional IT environment, won’t need that in 2015. As applications become more service-based, the need for our corporate data centers begins to diminish. The notion that every company has to have their own data center — that’s my father’s legacy. That’s not for 2015.”

Interesting thought, but it seems a bit overzealous; 2015 is only eight years away, after all.

I’ve seen reports about the IT professional talent pool drying up by then, and I’ve seen - and written - plenty on the looming power crisis data centers face, but the extinction of corporate data centers is a drastic prediction, don’t you think?

SHARE 2007: New mainframe products

This year’s SHARE conference in San Diego was a busy one for product announcements, with the issues of mainframe security, compliance and modernizing your legacy applications being the most popular. Here’s a look at what came out.

Security and compliance

Most of the security software announced at the SHARE mainframe conference had another goal in mind: compliance. In other words, it’s important to make sure your mainframe is secure, but it doesn’t mean much if you can’t prove it when the auditors come hunting.

NewEra Software released Image Focus 6.1, which includes two compliance-related features. The first allows system programmers to access the IT resources they need while maintaining security for compliance purposes. Another can identify and track changes and anomalies in load libraries and partitioned datasets, so you know when there’s been a problem. Beta Systems Software announced Beta 96 Enterprise Compliance Auditor, which can detect critical conditions and events in your IT systems while continuously documenting everything it’s monitoring. And JME Software released Deadbolt Monitor, which allows you to use a Web browser to access live and historical records generated by RACF, ACF2 or Top Secret security systems.

In other security product news, another feature of NewEra’s Image Focus is Fast DASD Erase for z/OS, which overwrites tracks and cylinders with binary zeroes. Meanwhile, JME also released Deadbolt Password Reset, which allows you to use a Web browser to reset your RACF, ACF2, and Top Secret passwords.
Read more »

The straight story: Solaris on Mainframe

Solaris on the mainframe made big news this week, but we’ve got the straight story from David Boyes, president and chief technologist at Ashburn, Virginia-based engineering firm Sine Nomine.

Solaris on the mainframe is old news. Boyes’ team (which was responsible for bringing Linux to Big Iron) has been working on porting Solaris over for 18 months. Which means it’s that much closer to happening than you might expect — but Boyes won’t say when Solaris on System Z will become available beyond “soon”.

Porting Solaris over to run on z/VM became a possibility when Sun took Solaris open source in 2005. Up to that point, “Solaris was a black box,” Boyes said. Sun has been contributing technical knowledge and hardware, IBM just got on the bandwagon.

According to Boyes, the reasoning the behind moving the Solaris workload over to Big Iron is that Sun customers are sitting on a lot of small to medium-sized pizza box servers. It’s the same problem facing Wintel users — server sprawl.

“How do you address that workload in a way that can take the value of existing infrastructure and deals with the power and cooling issue?” Boyes said. The answer: Big Blue Solaris.

IBM is gung-ho on consolidating Unix workloads onto the mainframe. In a recent announcement, IBM said it would consolidate AIX Unix servers in its own data center, but in this case IBM is switching those AIX workloads over to Linux, not porting AIX to Big Iron.

But the news about IBM consolidating its own AIX machines and the possibility of Solaris on the mainframe got people thinking — Why not AIX on System Z?

Users on the IBM-Main site asked why doesn’t IBM create a special purpose engine, similar to the IFL, for AIX? Wouldn’t AIX running on a System z special purpose engine generate benefits similar to Linux on IFL?

Turns out those are entirely separate questions. Similar benefits? Maybe for end users — but what’s in it for IBM? Not much.

According to sources at Big Blue, IBM has no financial benefit to support an AIX specialty engine on the mainframe. It’d cost too much money to implement, virtualize and port AIX to System z; plus it would compete with AIX on System p, not to mention create another platform on which IBM and other middleware would have to be ported and certified.

Paul Murphy blogging over at Ziff seems to think that IBM will buy HP-UX. Maybe they’ll port that next. Two out of three ain’t bad.

If you’ve read this far, you should check out David Boyes’ presentation on Solaris on System Z from April 2007 to get the full details on the benefits of Solaris on Big Iron.

Green data center marketing is making me sick

If one more person pitches me on a Green Data Center story, I will barf in my laptop. Unless your server literally sucks carbon dioxide out of the air and sequesters it below the earth, please don’t pitch me on your green angle.

It could just be me — I’ve been living and breathing Green Data Centers for at least the past 10 months. I’m up to my ears in green because I have been writing a book on environmentally responsible data center operations (Chapter 1 will be published next week).

The environmentalists’ worst nightmares are coming true. This whole green wave could be just another fad — like the Macarena. Everybody is going to turn to whatever is next while marketers circle like vultures to suck the last bit of life out of the coolest pheonomenon to happen to IT since the Web.

Maybe I’m just jaded. Where are you at in the hype cycle? Is green IT just getting onto your radar or are you reaching for the Pepto? We’ll do our best to be gatekeepers against greenwashing. But do you want to see more coverage of actual green technologies? And can we blame green marketing for killing the goodness?

Dell to users; our customer service doesn’t suck as bad anymore

Get this; Dell was named number one in customer satisfaction for standards-based servers in the Technical Business Research (TBR) Q2 2007 Corporate IT Buying Behavior & Customer Satisfaction Study: x86-based Servers.

According to TBR, Dell’s overall weighted score improved 1.1% sequentially, while HP and IBM declined by 1.5% and 0.8%, respectively. Dell was the only systems provider to show an improvement in TBR’s latest study, with customer satisfaction scores increasing in eight of the nine attributes measured, including server management, phone support, delivery time, value and ease of doing business.

I wrote a story in March and when I spoke with Dell about their poor customer service, they recognized the problem and seemed to really want to change.

Dell is investing an incremental $150 million this year on its “customer experience” initiatives and quarterly results released by the company in November showed signs of improvement.

“Dell has been in business for 23 years, and there are lessons learned every day,” said Daniel Bounds, a senior product manager for Dell.

Dell increased the number of customer service agents in order to decrease average hold times from nine minutes to three minutes in the past year. The company also reduced call transfers by over 30% and has improved first contact resolution rates by 20%.

Despite this, there are still websites devoted soley to bashing Delland a simple Google search shows plenty of recent customer compliants.

It’s hard to overcome a negative reputation. For instance, my very first vehicle was a little red 1992 Hyundai Excel sedan. I could have run faster than that car accelerated up steep hills and it seemed to break down every other week. Hyundai later made an effort to repair its reputation of poor quality, and according to reports, they have.

The thing is, I’ll never give Hyundai the chance to show me how much they have improved because like many consumers, once I’ve been burned, I don’t put my hand back in the fire.

But, as I sit here typing on my work-issued Dell laptop, I hope Dell’s customer service is improving as much as they say it is.

Will CMDB standards offer a single view of IT infrastructure?

Most data centers out there have a mix of hardware and software vendors, and as a result it can often be tough to get those devices and applications to speak to one another in the same language. Correct me if I’m wrong here.

I’m asking this because the CMDB Federation working group has just announced a specification for sharing information between CMDBs and other management data repositories such as asset management and service desk software. The federation is no joke — it includes heavyweights like BMC, CA, Fujitsu, Hewlett-Packard, IBM and Microsoft.

The group hopes that the specification becomes a standard way for vendors’ software to communicate configuration data with one another so that users know about everything in their IT environments.

Would that be helpful? I’d like to hear from users out there about whether it’s difficult to keep track of everything they have, and if so, whether a standard way for those applications and hardware to speak to one another would be a good thing.

Fujitsu installs fuel cell for its data center chiller plant

Fujitsu has a fuel cell from UTC Power up and running at its Sunnyvale, Calif. campus. The 200-kilowatt fuel cell is helping to run the company’s chiller plant, which in turn is cooling the data center.

Though Fujitsu is talking about how it is the first high-tech company in California to put the fuel cells to use, there are other companies who have been using them. Verizon in New York and the First National Bank of Omaha are no strangers to the technology. Fujitsu said it’s considering buying more fuel cells to directly power the data center, and is working with UTC Power to get one that can operate at variable loads instead of the steady 200 kilowatts all the time.

So what is fuel cell technology? Basically, the fuel cell module takes natural gas and strips the hydrogen from it. The hydrogen then runs through a fuel stack, and when it is combined with oxygen, it generates electricity. The byproduct is mostly water vapor with a little carbon dioxide mixed in.

Meanwhile, UTC Power’s fuel cell module, the PureCell, takes up 180 square feet and creates 925,000 BTUs of heat per hour. At Fujitsu, that heat will go into water to be piped back into the boiler system and used in the office environment for heat.

Fujitsu wouldn’t say how much one of these modules costs, but it must be big bucks because they’re getting a ton of rebates on it. Pacific Gas & Electric is giving them $2,500 per kilowatt and the federal Department of Energy is giving them a tax credit of $1,000 per kilowatt. For a 200 kilowatt module, that adds up to $350,000 in rebates and tax credits. Not too shabby.

Still, there is the persisting fear out there, sarcastic or not, that hydrogen power is not safe. What do you think? Would you ever consider using fuel cells to power your data center? If you need fuel to ponder it, check out our two-part series on hydrogen fuel cells in the data center.

IBM to become Solaris OEM

Recently saw this piece of news about IBM joining forces with Sun. Apparently, these big boys think that they’re going to put the squeeze, according to this USA Today article, on HP. The idea is that Sun applications will be able to run on IBM servers (Series X at first, but eventually mainframes) in hopes of getting Sun back in the spotlight in regards to software.

Jonathan Swchartz made some interesting statements on his blog about this news. First, he referred to IBM as “the first Tier 1 x 86 system vendor” to take part in this strategy. Does that mean that he expects others to do the same? Who? Dell? He also seems to really believe that this is really about better serving IT departments, which I think is true. Nothing increases productivity like standardization, modularity and uniformity - at least in the mid-term.

Though this move is being billed as a collaboration, but it seems to me that the real winner will be IBM, who stands to make Sun’s hardware irrelevant. If this is not the case, it is at least an interesting strategy. Who knows, maybe the Solaris subscriptions will go through the roof if you’re able to run it on a BladeCenter and Sun will be able to switch missions.

Mainframe training programs rebounding

SAN DIEGO — Panelists from five colleges and universities started Tuesday morning at the Share user group conference by talking about their mainframe course curriculum.

The panelists — from North Carolina Central University, Estrella Mountain Community College, Illinois State University, Widener University, andMarist College — are all part of the IBM Academic Initiative, which encourages universities to teach mainframe skills to students.

The panelists were just five of 325 schools on the IBM Academic Initiative list that are either investigating or currently teaching mainframe-related courses. That’s 325. That’s a lot. Do you think there are that many mainframe job vacancies out there? Not everyone thinks the mainframe skills shortage is as dire as some make it out to be.

From the discussion that ensued during the panel session, it seems like the skills shortage is in information technology overall, not just the mainframe. There were a series of actions and consequences over the last seven years that led to meager computer technology programs in colleges and universities: the technology bubble burst, which caused jobs to dry up, which caused enrollment drops in computer-related college majors, which caused a dearth of courses and course literature.

But it also seemed like the schools were starting to build their curriculum back up again. The panelists were a good example of that.

One thing that was interesting was the structural differences between the mainframe curricula from school to school. Some have certificate programs, others have mainframe course tracks that are rolled into existing computer technology degrees.

Cameron Seay from the North Carolina Central University, for example, explained how its mainframe courses are really just a subset of business classes but that they’re looking “forward to expanding in the future.”

“We’re in the School of Business and don’t have as many technology courses as I’d like,” he said.