Detroit vs. Iowa Data Centers, GM vs. Detroit competing for the IT talent

Chris Crosby has a post on Detroit as a data center location and brings up the folks in Iowa with Facebook.

Detroit. The Data Center Capital of America

Life is good in Altoona, Iowa. With the coming of Facebook, servers will quickly outnumber the community’s 15,000 residents and the city is poised to become one of the country’s leading data center destinations. The citizenry of Altoona are, of course, ecstatic at their good fortune. The economic benefits alone are too numerous to consider. The police department is contemplating adding a second car, the country club might add nine more holes so members can play a full 18, and there’s a rumor going around that Krispy Kreme might be coming to town. Yes indeed, the gentrification of Altoona has begun—and good for them. I have nothing against the good people of Altoona. And yet I have to ask, “Why Altoona?” Why this small oasis in the Hawkeye state as opposed to say, the Motor City? That’s right, why not Detroit? I pose this as a serious question. I realize that a few of you effete data center snobs might snort in derision at the mere mention of this discussion, but really, what has Altoona got that the buckle of the Rust Belt doesn’t?

Coincidentally, GM has been making some news not nearly as widely covered as Facebook on its data centers in Detroit.

Facebook adding to its existing infrastructure is actually less transformative as GM going from its past of outsourced IT to data centers it owns and runs.  GM has recognized that IT infrastructure is critical to its success just like Facebook.

The GM renaissance of IT, canceling outsource contracts, building state of the art data centers I think is actually a cooler story to read about the transformation of IT than just another Facebook data center.  WSJ tells the history of how in 2011 a three day HP mainframe shut down woke up the CIO how ancient their IT infrastructure is. 

Hewlett-Packard Co. ’s outsourcing relationship with General Motors Co. may have been doomed regardless, but the last straw might have occurred in October 2011, early in GM CEO Dan Akerson’s tenure, when a mainframe computer that H-P operated on its  behalf went on the fritz for three days.

“All of a sudden we started having slowdowns in our manufacturing around the globe.  A bell went off in my head,” Mr. Akerson told the Wall Street Journal in an interview earlier this year. “It was so fundamental you just assumed that a company of our stature, our size, our complexity, our global reach, you had to have a 21^st century IT infrastructure. We didn’t.”

General Motors Co.CIO Randy Mott

The outage meant that GM suppliers couldn’t be sure how to fill certain expected orders and had to guess at the details, according to Jeff Liedel, executive director of infrastructure engineering at GM. He said the outage probably “couldn’t have gone on longer, without forcing us to shut plants down.”

Here is a B-roll video with no sound that shows the new data center space.  It's not as sexy as a Facebook data center and the dress shirts with slacks don't look anything like the jeans and t-shirt culture.

The enterprise data center and a companion data center at the Milford Proving Ground are part of a previously announced plan to transform GM’s global IT footprint from 23 facilities to two by 2015. Construction of the $100 million data center expansion in Milford will begin this summer.

GM IT is leveraging the Warren and Milford data centers to create a secure, private cloud that allows super-computer applications, servers and data storage to be efficiently and quickly accessed among multiple users.  

“Our data center consolidation is just one of the initiatives driving the transformation of GM’s business,” said Randy Mott, GM vice president and CIO. “It’s part of an overarching strategy to transform not only information technology but also allow GM’s business operations to be more responsive to our customers, quicker to market and deliver on our objectives to shareholders.”

GM has a 2nd data center to provide an active fail over site where data can be mirrored.

The Milford location was chosen because it is more than 25 and less than 50 miles from Warren, allowing “mirrored” data, so if one facility is off line for any reason, the other will have the same data available without interruption.

"It's all about reducing risk and making sure no one event would affect both centers at the same time," said Curt Loehr, GM Information Technology project manager. "Each Center has its own utility feed using separate paths to provide uninterrupted power.  We even checked weather data going back a half century and Warren and Milford are affected by separate weather patterns." 

GM is saving money by having the data centers on existing campuses, which have negotiated bulk utility rates, existing infrastructure and security.

I guess I kind of felt compelled to write about GM given I have a Cadillac CTS as a rental car.  no comments on the car.  The data center is way more interesting. 

Looking for where to save time in projects, lessons from container revolution

Long before I worked in data centers, information warehouses, I used to work in physical warehouses for HP and Apple as a distribution engineer.  Pallets, forklifts, boxes, bar codes, and processes were what I worked with every day.  When containers arrived to data centers so many people made a big deal of it, but I had seems containers used over 30 years go to move massive quantities of goods and having containers to encapsulate IT white space is interesting, but not even close to as revolutionary containers were to the shipping industry. 

The Economist has an article on Containers where they discuss how innovative containers were.

Innovation

Big bills left in the shipping container

May 20th 2013, 21:16 by R.A. | WASHINGTON

I LOVE the history of the shipping container. Nothing could be more confounding to our usual ideas about innovation, stagnation, and technology.

And Contaners were a simple idea created by the execs not some whiz kid.

Except that's not how it works out. And not because an inventor came up with a revolutionary new technology. Instead, a few savvy shipping magnates figured out a better way to do things. A much, much better way.

The new system was simple. Customers or aggregating shipping firms would pack their cargo into giant, purpose-designed metal boxes. The boxes could be loaded on truck or rail trailers for transport to port, where purpose-built cranes would swing them onto purpose-built boats that don't carry anything but containers. Cargo could travel from factory to destination without ever being handled by a human. 

When you feel like there aren't any interesting opportunities turn around take a look behind you.  There may be some really good opportunities in your past that you didn't see.

Obviously container shipping revolutions don't come along every day. But I find this history to be a powerful antidote to economic pessimism. It's as if humanity faced a stand of trees stripped of low-hanging fruit and despaired of further economic gain, only to have someone shout, "Hey, there are also a bunch of trees behind us!"

What am I working on in a Data Center, thinking about how work can be improved

I am in a data center this week and one of my friends asked what I am up to. I could write him an e-mail or throw up a blog post that describes what I am doing in a more interesting way.  

Being at a data center many people bring their lunch in (hint: this means I am not in a Google data center where there is catered lunches.).  Given I am from out of town I don't have my lunch, so I took myself out to a local place and watched this Ted Video by Dan Ariel "What makes us feel good about our work?" while I ate and watched this video I found Dan makes points that I can use to describe what I am working on in a data center.

Data Center operations can always be improved.  One technique I like to use is talking to the people who do the work and find out what their pain points are.  Dan Ariely's question of what makes us feel good about our work? is another way to view what are the pain points that cause you to feel bad about our work.  it is not just the money. 

I want to talk a little bit today about labor and work. When we think about how people work,the naive intuition we have is that people are like rats in a maze -- that all people care about is money, and the moment we give people money, we can direct them to work one way, we can direct them to work another way. This is why we give bonuses to bankers and pay in all kinds of ways. And we really have this incredibly simplistic view of why people work and what the labor market looks like.

One point that resonates for me is the one on a cancelled software project described at the 7:45 mark of the video.

I went to talk to a big software company in Seattle. I can't tell you who they were, but they were a big company in Seattle. And this was a group within this software company that was put in a different building. And they asked them to innovate and create the next big product for this company. And the week before I showed up, the CEO of this big software company went to that group, 200 engineers, and canceled the project. And I stood there in front of 200 of the most depressed people I've ever talked to. And I described to them some of these Lego experiments, and they said they felt like they had just been through that experiment. And I asked them, I said, "How many of you now show up to work later than you used to?" And everybody raised their hand. I said, "How many of you now go home earlier than you used to?" And everybody raised their hand. I asked them, "How many of you now add not-so-kosher things to your expense reports?" And they didn't really raise their hands, but they took me out to dinner and showed me what they could do with expense reports. And then I asked them, I said, "What could the CEO have done to make you not as depressed?" And they came up with all kinds of ideas. They said the CEO could have asked them to present to the whole company about their journey over the last two years and what they decided to do. He could have asked them to think about which aspect of their technology could fit with other parts of the organization. He could have asked them to build some prototypes, some next-generation prototypes, and seen how they would work. But the thing is that any one of those would require some effort and motivation. And I think the CEO basically did not understand the importance of meaning. If the CEO, just like our participants, thought the essence of meaning is unimportant, then he [wouldn't] care. And he would tell them, "At the moment I directed you in this way, and now that I am directing you in this way, everything will be okay." But if you understood how important meaning is, then you would figure out that it's actually important to spend some time, energy and effort in getting people to care more about what they're doing.

You can of course guess who the big unnamed software company is in Seattle and which CEO this was.  To give you more information on support your conclusion here is when Dan Ariel was in Seattle and series sponsor in July 2012.

Presented as part of Seattle Science Lectures, with Pacific Science Center and University Book Store. Series sponsored by Microsoft.

Dan points out the issue of ignoring performance and how it affects what people think of their work.

Now there's good news and bad news here. The bad news is that ignoring the performance of people is almost as bad as shredding their effort in front of their eyes. Ignoring gets you a whole way out there. The good news is that by simply looking at something that somebody has done, scanning it and saying "uh huh," that seems to be quite sufficient to dramatically improve people's motivations. So the good news is that adding motivation doesn't seem to be so difficult. The bad news is that eliminating motivations seems to be incredibly easy, and if we don't think about it carefully, we might overdo it. So this is all in terms of negative motivation or eliminating negative motivation.

So much of what goes in the data center can be many small processes that are required to complete an overall task.  Think of all the steps required to receive a new server and get server used by customers.  Think about this example Dan uses.

Let me say one last comment. If you think about Adam Smith versus Karl Marx, Adam Smith had the very important notion of efficiency. He gave an example of a pin factory. He said pins have 12 different steps, and if one person does all 12 steps, production is very low. But if you get one person to do step one and one person to do step two and step three and so on, production can increase tremendously. And indeed, this is a great example and the reason for the Industrial Revolution and efficiency. Karl Marx, on the other hand, said that the alienation of labor is incredibly important in how people think about the connection to what they are doing. And if you make all 12 steps, you care about the pin. But if you make one step every time, maybe you don't care as much.

The types of work I am studying are everything from design, construction, operations of the building and IT equipment.  Dan thinks some tasks should be change to create more meaning.  But, the problem is how far do you go?  One person cannot have the skill to design, build, construct, and operate a data  center.  Well, there are very few people who can do all of these things, but their capabilities cannot scale to have enough resources to run a data center.

And I think that in the Industrial Revolution, Adam Smith was more correct than Karl Marx,but the reality is that we've switched and now we're in the knowledge economy. And you can ask yourself, what happens in a knowledge economy? Is efficiency still more important than meaning? I think the answer is no. I think that as we move to situations in which people have to decide on their own about how much effort, attention, caring, how connected they feel to it, are they thinking about labor on the way to work and in the shower and so on, all of a sudden Marx has more things to say to us. So when we think about labor, we usually think about motivation and payment as the same thing, but the reality is that we should probably add all kinds of things to it -- meaning, creation, challenges, ownership, identity, pride, etc. And the good news is that if we added all of those components and thought about them, how do we create our own meaning, pride, motivation, and how do we do it in our workplace and for the employees, I think we could get people to both be more productive and happier.

What I do in a data center is not public, but Dan Ariely's talk is public so I can reuse his presentation to illustrate the concepts I am working on.  The approaches I am using are way beyond what I originally did when I started working over 30 years ago.  Using mobile, social, and big data concepts allow new ways to improve work.  And guess what.  What I am working on makes me feel good about my work, because I am helping others feel about their work.  

As Faceboook Lulea powers up, makes sure its neighbors don't put infrastructure at risk

Part of what a big data center company like Google and Microsoft have learned to expect is that neighbors show up after they are the first to build.  Facebook going to Lulea was one of the first data center operators in the area.  And now that there are others looking at Lulea and Facebook is going through commissioning, there are concerns there is enough infrastructure.

NyTeknik reports in Swedish on the local event.

Soon it's time for the commissioning of the first of the giant data centers like Facebook is based in Luleå. Availability of green electricity from hydropower and a stable electricity grid were two of the factors that got Facebook to choose Luleå.

While trying municipality to attract more IT companies and a large computer giant is about to establish itself around Facebook in Porsön in Luleå.

But now reacts Facebook and through its agents Pinnacle Sweden, asked to Luleå further investigate the availability of energy, water and sanitation, the newspaper reported Norrländska Socialdemokraten.

Here is a picture from the same publication.

What's next for GreenM3 to focus on? Thinking 5 years out what will the data center industry be like

I started out in the data center industry working with some smart people who asked some tough questions

1. What is the energy consumption of an application in a data center?
2. What is a way for our company to be a thought leader in the data center industry?
3. How should I apply product lifecycle management principles to the data center?

All these are questions in the past, and I have been getting kind of bored.  I left HP, Apple, and Microsoft because I got bored.  Boredom has bad effects.

Boredom is generally an unpleasant emotion, composed of a lack of stimulation and physiological arousal. Unfortunately, boredom is often associated with the academic setting of attending class and studying. That said, while academic boredom may be present in higher education, do we care?

...

As stated by the authors, “the evidence in our studies suggests that boredom typically impairs attention, motivation, behavioral strategies, and performance in achievement settings. The pervasiveness of the boredom experienced by many students, coupled with its deleterious effects, clearly implies that educators, administrators, and policy makers responsible for the design of academic settings should pay more attention to this emotion.”

Last week another smart data center guy said in response to an e-mail.

It sure is an interesting time....I don't think anyone knows how it will turn out in 5 years.

Yeh, it is impossible to know what data centers will be like in 5 years.  But, then I thought about the question more.  

Five years from the data centers built are going to look different than what is currently industry standard.  So, what is known there will be change.  Who is the most aggressive and innovative?  There is media coverage for some data centers that are not leading others to be innovative.

Data Centers 5 years from now will be greener - more energy efficient and more renewable energy.  The cost of water will most likely grow faster than the cost of power.

There are actually many things that can be reasonably certain will occur 5 years from now.  Of the top data center operators who will be in the top 5 years from now. Who are the new companies?

What I like about thinking 5 years out is it is hard to get bored. :-)