ServerTech Blog - Information for the Data Center Equipment Manager

The beginning of the Cybernetic Data Center! Maybe?

This is too good to pass up! Over at smartplanet.com, Janet Fang gives a nice synopsis of an ingestible microchip used to monitor medication use and corresponding vitals. The data is available by wearing an adhesive patch on the skin which can relay information to the patient's smart phone. From there, the information can then be passed on to the doctors.

You must be asking, "How can this type of thing be the beginning of the cybernetic data center? And, what do you even mean by that?" I know this statement might be a bit of a stretch, but hear me out. The term "cybernetic" has been used in a variety of contexts from a simple user interactive monitoring and management system, all the way to the "Terminator". In general, though, it simply refers to a tight interaction between an organic body and some form of engineered technology. I believe the case of an internal microchip powered by the body's chemistry and communicating via connectivity on the skin qualifies as cybernetic.

We all know a generic data center definition. How about this one from Wikipedia, "A data center or data centre (also datacenter or datacentre) is a facility used to house computer systems and associated components, such as telecommunications and storage systems. It generally includes redundant or backup power supplies, redundant data communications connections, environmental controls (e.g., air conditioning, fire suppression) and security devices." Take a close look at this statement and consider the cybernetic organism described above.

"Facility used to house computer systems", check!

"...telecommunications and storage systems", check on communications and the phone adds storage!

"power supplies and environmental controls", check! The body is great at that! True 100% uptime!

Add in that the communication of data from the patient to the doctor is flowing on existing carrier networks, and you get a data center (though a very small one) very much like any other.

This is, of course, just the beginning of the potential for this kind of technology. Also, just the beginning of the risks for this kind of technology, as any number of sci-fi writers have expounded upon. Full cybernetic data center features would probably need to include, muliple sensors and active control devices, some two-way communication, programmability of internal circuitry, serious "un-hackable" security and an ultimate safety kill-switch; but I think it is fair to make the case that we are truly entering the age of the "Cybernetic Data Center".

I can't wait to get my bionic eye so I can upload everything I see to the cloud!... :-)

 

 

Stop asking, "Who will ever need that much storage?"

Seriously. Over at Tech Target, Sonia R. Lelii writes about Cleversafe Inc and their 10 exabyte (10¹⁸ bytes) storage system. Now, I know she isn't actually asking that question. But, the indication that this question should ever be asked just will not go away. Before this, some asked the same question about gigabytes, and then terabytes, and then petabytes. There is no reason to expect this to end. Data storage continues to multiply exponentially. Ok, I give you that it has to end some point before the representation of all atoms in the observable universe, but we are nowhere near that number of 10⁸⁰. Granted, all of the worlds information is only on the order of hundreds of exabytes right now, but it pales in comparison to even the number of atoms in a single person 7x10²⁷ or the number of bacteria on the earth (okay, technically prokaryotes aren't exclusively bacteria, sheesh!) 5x10³⁰. Anyway, the point is that in a few years I don't want to hear anyone say, "Who is ever going to need a zettabyte of storage?" 

The limits of magnetic memory?

IBM has successfully created a magnetic memory bit of a mere 12 atoms in size, whereas the standard disc drive needs roughly a million atoms per bit. The caveat is that it must reside at the very low temperature of 10 Kelvin. The full study has been published in the journal Science. According to an interview by Reuters, a room temperature version is theoretically possible at 150 atoms.

Is this the end of Moore's Law? Can we even find a practical means to get to this design? 

What do you think?...

Know your goals.

Over at eWeek, Chris Preimesberger has some great food for thought on some misconceptions in the data center monitoring arena. For me, the overall takeaway is that you must have a good grasp of the goals for the data center and the overall organization to make the best decision in choosing a monitoring package. Actually, the idea that you should attempt to get everything into one single monitoring package may not be the best way to reach your goals. Remember that each person and job task within the data center is likely to have different goals which may require targetted monitoring and management for most efficient use of their time.

The cloud is definitely in it for the long haul.

I just caught this article in the LA Times by David Sarno comparing cloud storage for the public. It is certainly good for our line of business to see this being pushed more and more into the public's perception. With the ability to move all books, music, movies etc to a "safe" place independent of computer hardware and the ability to then access that data from various devices, how can one not see exponential cloud growth for years to come. 

Personnally, I looking forward to being able to talk to my car and get access to my cloud music. Chris Woodyard with USA Today has a nice article.

White Collar IT

Insights from VP Marketing Mark Harris

---

Since 2005, the world of IT has been under pressure like never before. The traditional role of the IT organization has been to provide a suitable level of services with little if any scrutiny on the costs to do so. It has been quite easy for the annual IT expense line item to exist in corporate budgets year after year as a solitary figure.

The world of IT has been considered so technical within the corporate executive management levels that only those who are part of the IT organization actually understand what is being done and where and how their money is being used with any level of detail. Hence, what has filtered upward to the CEO and CFO is a single line item, a spend number for the year, and a hand-shake assurance that the money is being used wisely. This has been the standard operating practice for years, and there has been little if any requirement for justification or measurement on how the previous monies were being used. IT success was being measured by Availability and Uptime statistics alone. The IT organizations of the corporate world were even encouraged to meet many times stringent service levels by any means, and at any cost. Brute force and "miraculous" ad-hoc processes were commonplace, rewarding those individuals that "saved the day" to assure production IT services remained available across the corporation. Think of this as a blue collar type mentality; with enough tools, enough might and enough time, anything can be delivered if cost is no object.

The world of IT is getting more complicated today. With the world's economy under pressure, and the price of energy skyrocketing, a significant focus on IT has emerged. The costs associated with IT have risen sharply over the past five years based primarily on the costs of energy. Today, it is becoming obvious that the energy costs consumed by most IT devices over their useful lifespan actually exceed the acquisition costs of those same devices. So while historically there have been entire departments of purchasing agents squeezing vendors for a few extra discount points on equipment purchases, there has been little if any effort employed to optimized the amount and efficiency of energy used within the IT infrastructure and it's data centers. These highly visible rising IT costs are now drawing attention at all levels. A tremendous amount of oversight has arisen. Many previously autonomous organizations are being forced to participate in this overall challenge, and visibility into just how IT operates including its core structure, its strategies, and its executions is required.

A new approach to IT is beginning to emerge. While the days of brute force service delivery are far from gone entirely, there is an enormous desire for IT to be run as any other business unit within a company. There must be a plan, a set of metrics, and a set of ongoing reviews to establish both quantitative and qualitative performance. In simple terms, IT is being asked to do the paperwork which supports its business. IT is now painfully aware that their very bane of existence is contingent upon other organizations (e.g. Facilities) which were previously considered subordinate or independent. All in all, business analysts are now defining the world of IT. These "White Collar" financial analysts within the IT organization now command significant power, and it is not uncommon to see new proposed projects that would have previously been executed blindly, instead being scrutinized and delayed or cancelled due to the inability for these proposed solutions to be deployed with defendable business justification. White Collar IT professionals are tasks to run their business in a justifiable and defendable manner.

So what's happening to enable the new style of IT as a business unit? Data Center Infrastructure Management or "DCIM." It is a term coined by the analyst community which has now stuck as a White Collar solution category. DCIM refers to the ability to manage at a highly discrete level every asset involved in the delivery of IT. Every aspect, over the entire lifetime of the asset. The physical asset ownership, its deployed location, the resources it consumes and the overall context of where it fits relative to all other assets. Tactically it also includes power and environmental management, visual data center floor-planning and rack/cabinet configurations, connectivity, etc.

The category of DCIM has a ton of potential, but like virtualization was half a dozen years ago, it is following the same hype-versus-reality path as all other revolutionary innovations. While some impressive point capabilities exist across many of the 200+ vendors that claim some form of DCIM value, it has yet to produce the coordinated results sought by White Collar IT. That said, DCIM is here to stay and those companies that begin down this path will be rewarded. It is for this very reason that Sentry Power Manager (SPM) has been engineered to provide specific linkages to external DCIM applications, building management systems and IT management frameworks from IBM (News - Alert), CA, HP and BMC. Server Tech is ready for the DCIM revolution when you are.

Hidden threats to business, case in point - the Thailand floods.

Over at CNN Money, David Goldman reports on Intel slashing its sales forecast due to circumstantial disaster. The Thailand floods earlier this autumn have wreaked havoc on hard drive manufacturing. Set aside the annoyingly obvious question of why it was considered a good idea to build such factories in a flood plane and consider the ripple effect. Could Intel have predicted that their sales would be contingent upon where someone else built their factory? I suppose they might have. 

More importantly, is anyone taking note of this circumstantial effect on a whole industry due to one sector's questionable decisions? Certainly, many within the data center industry are taking note, though not many companies can do anything to prevent such ripple effects. It makes me wonder which other industries we rely heavily upon might have such built-in weakness due to clustering of manufacturing facilities. 

Collection of cool data centers

There is a neat little collection of interesting data centers over at Colo & Cloud. My favorite has to be the James Bond style of the Bahnhof’s Pionen “White Mountain” data center. But, I do respect the idea of re-purposing other buildings like mills, mines, shopping centers, or anything else sitting dormant. Hmm!, now that I think about it, an old mill could be a cool project to create a home for my retirement. I better get on that!... :-)

 

 

Sandbagging or just playing it safe?

Peter Judge, over at green data center news, has a provocative piece alleging that some hardware manufacturers (specifically server manufacturers) are holding back on providing data showing the safe operation under the higher temperature allowances from ASHRAE. He claims that this supposed reluctance to release information is slowing the greening of the data center.

I suggest he has unreasonable expectations for these hardware manufacturers that must, in the end, answer to their customers' requirements for reliability first. I see no benefit to the manufacturer to hold back on data other than it being incomplete. He claims that it is out of a desire for secrecy and competitive benefit, but I see no such benefit. Indeed, it would seem that, as he claims, Dell should have the benefit by releasing data showing the capability of withstanding high heat conditions. Furthermore, what secret is being kept. All one has to do is open the box to see the electronic and mechanical design.

The data center has always been a world of conservative claims and will continue to be for some time as the requirement for reliability and up-time are still tops for most applications and businesses. Until that changes, I think we should expect that hardware manufacturers will continue to answer the customers' expectations by playing it safe.

But, then again, I could be wrong. What do you think?...

Which weighs more - the internet or a strawberry?

Seriously, someone figured this out. Anyone up for checking the math?