BandL Weblog

Where should you locate your next data center?

Geography may have lost its luster among curriculum creators, but it remains an important consideration when planning a data center. After all, no business wants to locate such a critical component to its overall mission in a place where it will be under six feet of water in six months.

 

Threats to a data center from natural disasters can be reduced depending on where the facility is located. When it comes to natural disasters in the United States, nobody has a better handle on them than the Federal Emergency Management Agency, also known as FEMA.

 

Based on presidential disaster declarations over the last eight years, data centers as less likely to suffer service interruptions from natural disasters in the Northwest than in any other part of the country.

 

According to the agency, from 2000 to 2008, FEMA Region X, made up of Washington, Oregon and Idaho, had areas declared disasters by the president only 25 times. Disaster types were fires, floods, earthquakes and severe storms.

 

The Northwest offers other more economic benefits as well. As we have pointed out in past blogs, one of the prevailing concerns with data centers is their massive power consumption to both run the servers and cool the data center as a whole. And this is another reason to favor the Northwest.

 

As Vertatique points out in their blog,

 

"...the popularity of the Columbia River Valley for mega data centers [is] due to its inexpensive and low-carbon hydroelectricity, its climate which reduces cooling costs, and its fiber optic infrastructure. "

 

A close second to the Northwest as a relatively safe place to locate a data center is directly on the other side of the country. During the period tracked by FEMA, Region I, comprised of the New England states, had areas declared disasters by a president only 28 times. FEMA Region II, made up of New York, New Jersey, Puerto Rico and the U.S. Virgin Islands, finished a close third with 29 disaster declarations.

 

However, electricity costs in the Northeast (New York, New Jersey and New England) averaged 0.1628/khr for commercial customers and 0.1131/khr for industrial customers. Compare those rates to the Northwest's 0.0715/khr for commercial customers and 0.0574 for industrial customers. Moreover, most of the electricity in the Northwest comes from renewable, "green" hydroelectric power, which is an added benefit as companies try to be more environmentally conscious. Of course, those looking outside the country for a bargain could look towards Iceland for their data center, but there are concerns of natural disaster and connectivity issues, as Iceland sits atop a giant volcano.

 

So considering the Northwest's cheap and abundant power, as well as the relatively small incidents of natural disasters, there is really no wonder why companies like Microsoft, Google, Yahoo, Ask.com, and Intuit Inc. have been flocking to the Northwest to build out mega datacenters.

Making sense out of turning your data center green.

 
 
At the end of this summer, Prentice Hall published Marty Poniatowski's Foundations of Green IT, a book on designing energy efficient data centers. The volume is a comprehensive discussion of consolidation, virtualization, efficiency and ROI in the data center. If you want to take a free peek at the book, you can check it out at Bitpipe.com.

 

While pressure continues to increase on data center managers to make their facilities greener, many shops don't have a handle on the benefits of doing so. In his preface to the book, Poniatowski, who is chief technology officer managing all pre-sales and consulting technical experts at Computer Design and Integration, offers several compelling reasons for greening the data center:

• For every dollar spent on hardware, 50 cents is spent on power and cooling. This number has increased consistently over the last several years, and this is no doubt one of the primary drivers for virtualization and consolidation.

• The average data center is 52 percent utilized and 1/3 of data centers are approaching maximum floor space capacity.

 

• The average rack is 56 percent utilized and 1/3 of data centers are approaching maximum rack space density.
  

 

Poniatowski argues that the number one challenge for the data center manager is power and cooling. Power and cooling shortages, he writes, cause server or storage downtime in 49 percent of all data centers.

 

A substantial amount of the power consumed by data centers can be reduced —as much as 25 to 30 percent—by virtualization. That's important when one considers that the financial outlay for powering data centers has doubled since 2000 at a cost of some $4.1 billion.

Who you gonna trust? Apparently not Google.

As Google tries to expand beyond the search engine business, it's meeting resistance from some IT quarters. On the surface it may seem that IT managers are worrying about their job security as the King of Search tries to cajole corporations to move their information from onsite data centers into Google's, but IT managers also have concerns about security, privacy and compliance that will take more than Google's good guy reputation to belay.

 

A case in point is a recent CIO poll taken by TechRepublic. Asked who they trusted more, Google or Microsoft, the dozen IT execs on the panel chose Microsoft by a 2 to 1 margin. For the majority of data center managers on the jury, the choice was largely a case of the-devil-you-know versus the-devil-you-don't.

 

Some key concerns raised about Google by the jurors were:

• It takes forever to get its products out of beta
• It's moving too fast in too many areas at the cost of security and trust
• It pressures companies to consolidate data at remote locations before they're proven safe and secure
• It requires integration of internal IT systems with an insecure environment, i.e.,"the cloud"

Winning the hearts and minds of IT pros won't be easy and Google, which has a reputation for making all the right moves when it comes to building its business model, has uncharacteristically stumbled in its current campaign.  

 

Its "Going Google" initiative targeted users, rather than decision makers, as a means of getting its foot in the door to the corporate data center. It set up a "Spread the Word" Web site to recruit users and Google-fans to pressure IT honchos to adopt Google's corporate offerings. As Dana Oshiro aptly points out in his article, while this may work when evangelizing to family and friends, it doesn't take a genius to predict the outcome of a discussion where a rank-and-file user tries to tell their IT admin how to do his or her job.

 

If Google wants to win its case as an alternative to Microsoft, it's going to have to appeal to the jury and not the audience in the courtroom.

Data Archiving Tools: What Works for You?

Every day I open my inbox to see news feeds of the latest applications and data archiving tools. 

 

I recently read about Sonasofts's SonaSafe for Email Archiving, Red Condor’s Red Condor Archive for message (email) archiving, Nayatek’s Datosphere for content archiving, Sherpa Software’s Discovery Attender, and Iron Mountain’s recent offer of Mimecast-based Total Email Management.

 

While I haven’t seen demos of all these tools yet, they look promising as different ways to manage the avalanche of data created daily by all of us. (And it seems most would work hand-in-hand with our data center automation products). 

 

But, I'm more interested in hearing your thoughts on these data archiving solutions. What do you use and how is it working for you? 

 

Each of these tools also seems to fit a slightly different niche.  What pain points do you feel are unique to your industry or particular business?  Do the available archiving tools address those points? 

 

And most importantly, what else would you like to see to help you manage your datacenter?

Is Your Data eDiscovery-Safe?

I think about eDiscovery data issues on a daily basis and have begun to wonder: Is it just me? Because we are immersed in data storage issues, I forget that perhaps not everyone knows the risks involved with a potential lawsuit. Last week I read an article by Greg Lawn in ComputerWorld titled “How to Avoid an eDiscovery Disaster.”  I thought it was a great piece of writing that outlines some very basic best practices for avoiding legal trouble with data retention.

 

My question is this: do companies feel they are fully prepared to produce all relevant documents should litigation occur? Are you confident that you could find that crucial needle-in-the-haystack memo if and when you needed to?  Perhaps, here at B&L, we should be doing more to educate our customers and potential customers to the risks of improper data retention, as well as the potential cost of data recovery if not properly catalogued. 

 

We created Archived Data Manager (ADM) for the very purpose of saving companies time, money, and effort in finding and managing their archived data.  We don’t want our customers to face fines for an inability to produce relevant documents, nor do we want them indebted to data retrieval companies when finding the information you need is just a click away with ADM. 

 

I am interested to hear what you think – is there enough information out there for companies to make informed decisions on storage and retrieval of archived data? Do you feel confident you can face an eDiscovery challenge should one arise?

What Effects Will Google Wave Have on eDiscovery?

Google Wave set out to reinvent email for the 21st century. However with its ability to post content in different areas simultaneously (e.g., Waves and blogs), it raises more than a few e-Discovery questions. I came across a blog article today that talked about this question of eDiscovery and Google Wave.

 

According to the blog post, Google Wave uses XML databases, so the entire thread (or Wave) can be reviewed easily and backed up. Also, Google Wave is based partially on XMPP, an open, XML-based protocol originally created by the Jabber open-source community back in 1999 that provided near-real-time, extensible instant messaging (IM) and presence information (e.g., buddy lists). According to the post, this adds to the security of the content because XMPP makes it very hard to impersonate someone in a Wave.

 

The part that concerns me is the use of dynamic content in a Wave. This is content that is generated on the fly, such as interactive maps and automatic content from robots. This information changes every time it is called upon (potentially), so tracking down the exact content of a Wave may be near impossible.

 

Another point to make is the fact that Google has opened up Wave to developers, enabling them to create gadgets, widgets and other add-ons to Google Wave. If they start creating objects that don't live within the Wave (like Wave Gadgets), then tracking down the actual content may be even worse than culling data from archived emails.

 

As I am just delving into this, I'd love to find out what others think about these issues and any others Google Wave may present to eDiscovery and data center operations in general.

Top 10 eDiscovery Resources

If you read my blog regularly, you know that we spend a lot of time concentrating on eDiscovery issues and ways to stay compliant with eDiscovery. With Archived Data Manager (ADM) helping companies comply with eDiscovery regulations, it behooves us to stay on top of the latest issues.

Well, I thought I'd pass on my top 10 eDiscovery information resources I turn to on the Web. So without further ado (although those who like their ado, please feel free to add more...), I give my Top Ten eDiscovery Resources:

10. The Amendments to the Federal Rules of Civil Procedure

Before you engage in a discussion about eDiscovery, it is best to understand what it is.

 

9. e-Discovery Insights

Written by an IT executive turned California e-Discovery attorney and consultant, the blog offers eDiscovery information particularly as it applies to California companies. However, he also offers great insight into the general topic of eDiscovery.

  

8. DiscoveryResources.org

An aggregator of sorts for eDiscovery information, resources and news.

7. Rule26Blog

A great blog that provides insight into eDiscovery issues as well as the eDiscovery implication of new technology (e.g., Twitter).

6. Google Alerts

Yes, Google. With their Google alerts, I get all the latest information regarding eDiscovery sent right to my Inbox.

5. EDD Update

A joint project of Law Technology News and Law.com Legal Technology, this blog offer great information from a law technology perspective.

4. Processor.com

Not a site solely dedicated to eDiscovery, it does however provide valuable articles on how eDiscovery impacts the data center.

3. The Data Center Journal

Like Processor.com, the Data Center Journal does not focus only on eDiscovery. However, it also provides great articles on how eDiscovery impacts the data center.

2. The Electronic Discovery Blog

A lawyer and electronic discovery consultant, W. Lawrence Wescott II is a senior consultant with Kahn Consulting, Inc., a firm specializing in the legal, compliance and policy issues of information technology and information management.

1. Christine Taylor's Blog

Although not all her blogs concern eDiscovery, she provides an insightful blog aimed at storage professionals (rather than legal professionals) that is always on-target and provides invaluable information.

Sites to watch:

The Organization of Legal Professionals

This newly-formed organization of legal professionals aims to help create guidelines and standards for eDiscovery for legal professionals.

 

I'm always on the lookout for new sources of information about eDiscovery (and data center management, for that matter), so feel free to let me know about your favorite resources.

Why Cloud Computing Makes Sense...for Some.

One of the key selling points of cloud computing is that a company can pick and choose IT services from a provider (or, ideally, several providers) as needed, but not have to invest in building up its own IT infrastructure. Need a database host? Want software to audit your inventory? Need more bandwidth? No problem. They're all just a click away.

 

In an example from our own business, some of our customers prefer not to incur the operational costs of owning a tape management system (e.g., maintenance and upgrading of the system). They have chosen to use B&L's Vaultledger OnDemand product and receive the full functionality of VaultLedger over the Web, or cloud.

 

Our customers can now control their backup tapes from any computer—anywhere in the world—with an Internet connection. All maintenance and upgrades to servers and software are performed by B&L from our facility, and are covered in the subscription fee.

Cloud computing is probably most advantageous for small to medium-sized businesses. It allows them a level of technology and services that in previous years would have taken a large capital investment.  

 

“True” cloud computing should allow for ultimate scalability.  And a payment structure that is often likened to that of a utility bill—only paying for as much or as little as is used. 

Right now cloud computing is still in its infancy. As bandwidth becomes cheaper and easier to access, the technology and its acceptance should grow.

But how does anyone provide security for all this data? And, if documents or files are lost in the cloud, who is legally responsible for that missing data? The vendor? The client?

Something to think about.

 

How to Decommissions Servers, Increase Data Center Space and Remain Compliant.

Just finished reading an article on practical steps data center managers can take to help free up much needed space in data centers.

 

The article is good and offers many practical solutions concerning consolidating data, or eliminating unused or under-used servers. However, many companies (especially in the financial service industry) often acquire, through mergers or acquisitions, servers that contain large amounts of online and offline data.  This data needs to be vetted, managed, and in some cases, retained for regulatory reasons.

 

Moreover, keeping servers that contain archived or offline data wastes resources and can be very costly. In fact, according to a study by McKinsey and Company, data centers produce 80 metric megatons of CO2 per year. And since data centers are no longer estimating costs by square footage of their data centers, rather the cost of a data center is being driven by the IT power consumption, reductions in the number of servers you run can have a great impact on your bottom line as well as the environment.

 

Luckily, B&L has a solution to help companies decommission servers while still keeping the data you need for regulatory compliance. Archived Data Manager™ (ADM) allows you to rapidly extract and index data from historical backup catalogs and safely decommission those servers while remaining compliant.

 

Think of it this way: if you were to keep every file you have on your computer in paper form, you'd soon run out of space in your office to house it all. However, by keeping everything in electronic form, you can get rid of the paper, but still have access to the information you need.

 

ADM works the same way to help decommission servers in the data center, free-up space, reduce energy consumption and still have the information you need to satisfy regulatory audits.

A Cloud by Any Other Name...

Cloud computing: is it the best new IT idea today or just a shiny repackaging of familiar services? The answer likely depends on who you talk to. Debate over the virtues of cloud computing has received a good deal of media coverage for the past two years as start-ups, computing giants, and even the Federal government engage in the conversation either as providers or consumers of computing services. There are both enthusiastic advocates of cloud computing and those who are more cautious about jumping on the bandwagon.

There are even, as Bernard Golden points out in his CIO article " Cloud Backlash: You Can't Call The Whole Thing Off", people starting anti-cloud movements because of the hype and disillusionment over the whole issue.

 

Yet, as with most debates, there is a middle ground that best encapsulates the pros and cons and the truth may be that cloud computing is an exciting concept that does incorporate some traditional (albeit rebranded) IT ideas with new abilities to provide scalable, pay-as-you-go services.

 

Cloud computing grew from clustering and grid computing concepts.  Connecting multiple computers to create a cluster allows them to share the data processing load and thus increase capacity.  Connecting multiple independent clusters creates a “grid” of computing capacity that vendors can use to offer increased services and storage capacity to customers. From these initial concepts, providers moved to offering IaaS (Infrastructure as a Service), PaaS (Platform as a Service), and SaaS (Software as a Service).   Cloud computing differs from these offerings in its flexibility.  Whereas SaaS and IaaS are standard outsourcing services, cloud computing, as stated above, allows businesses a new level of scalability.

 

At B&L Associates, we offer our VaultLedger product over the Internet. We call it OnDemand, others call is SaaS, and now we are told we should call it "cloud computing." So, I guess we are also a cloud computing provider. But no matter what you call it, I think most people would tell you that mass adoption of most enterprise-wide cloud services has been slow to materialize. Why is that? On the surface, it seems like an ideal solution. I think it may have to do with the slow general adoption of any technology.

 

Would we have been better off sticking with a single acronym or phrase to describe this process? And by doing so, giving it some history and validation, as opposed to switching names every few years?  Enterprise clients are traditionally conservative in adopting new technology until they basically have no other choice. The switch from open system to mainframes didn't happen overnight. How much longer would it have taken if we kept changing the name of the same technology?

 

So what should we do to stem the anti-cloud tide and help cloud computing reach its full potential? Honestly, I don't know. It could be several factors.

 

In my next blog post I'll talk about one that is close to my heart: securing the information in the cloud.

 

But I'm interested in what you think. What has to happen for cloud computing to reach the tipping point?