Inside Microsoft's living, breathing building

Inside Microsoft's living, breathing building

This is the second of a three-part series. Part 1 can be found here and Part 3 here.

Tearle Whitson and Jonathan Grove stroll across the rooftop of Microsoft’s Building 88. It’s one of the first sunny days of spring, and singing birds accompany the picture-postcard view of Washington state’s Cascade Mountains.

“This is one of the perks of the job,” says Whitson, taking a moment to survey the scene before climbing into the whirring, dark interior of one of the roof’s large, white air handlers. As nice as they can be, now rooftop field trips like this are few and far between.

Whitson and Grove have experienced a seismic shift in their workday since helping to develop Microsoft’s smart buildings tool. Two years ago, the two spent a lot of time climbing over rooftops, inspecting pump rooms and peering above ceiling tiles at variable air volume boxes.

“I used to spend 70 percent of my time gathering and compiling data and only about 30 percent of my time doing engineering,” Grove says. “Our smart buildings work serves up data for me in easily consumable formats, so now I get to spend 95 percent of my time doing engineering, which is great.”

Before Microsoft’s buildings leapt up the IQ curve, the duo’s home was on the so-called range. They’d move from building to building, camping out in each for two weeks at a time to inspect and tune it top to bottom before moving on to the next. It would take them five years to tune up all of the buildings on campus, and then they’d start the process all over again. Their tune-ups were making the buildings run more efficiently, saving the company around $250,000 annually – but the new data gold rush will help them save six times that much.

The duo now spend most of their time at the ROC, chewing on building data. Although they’re no longer camping out together tuning up Microsoft’s campus the old-fashioned way, the two have maintained the comfortable rapport and geeky banter they established working in the field.

Facilities engineers such as Whitson and Grove think of the buildings they care for as living, breathing things. Just like the human body, buildings have a wealth of indicators that things are going well – or, in some cases, not so well. Also like the human body, small ailments can lead to much larger failures, and an ounce of prevention can lead to many pounds of cure.

And now, with the new data-driven software solution that the team built, they can do an even better job of managing the health of Microsoft’s buildings – they’ve gone from country doctors, tapping a patient’s knees with a rubber hammer, to specialists with an MRI machine who can examine every layer of the knee inside and out.

Still, Whitson says it can be a tad unnerving to take a building and its network of sensors online and watch as the software immediately discovers a host of inefficiencies.

“There is a little bit of a mindset among facilities engineers. Everybody is prideful, and they take such ownership of their buildings, that it’s hard for them to find out that there was a lot that they were missing. It was a humbling lesson I had to learn early on while doing this,” Whitson says. “We have to get the old-school technicians out there to understand that this is going to help them. This is not to say you’ve been doing it wrong – you’re doing fine. But you can go farther.”

Both are quick to tell you that Microsoft’s smart buildings solution has revolutionized the way they and their fellow engineers work.

“We had the perfect environment and people to put all the pieces together. Our solution is a little unique to Microsoft, but very applicable industry-wide,” Grove says.

Image courtesy of Microsoft

A bevy of large, wall-mounted displays wrap around the interior of the room, and the same monitors sit atop every desk. The smart buildings tool dashboard is splashed across many of the screens, showing off a colorful collection of maps, dials, lists and tickers.

Engineers can get big-picture information at a glance, such as how many kilowatts of energy are being consumed across Microsoft headquarters at any one moment. With a few clicks they also can zoom in on one building, one floor or office in that building, or one piece of equipment.

“Let’s see how City Center building is doing today,” Whitson says, and within seconds he’s clicking through a wealth of information about the building – the number of employees who work there, the outside air temperature, the thermostat, what time the lights come on and go off, even a list of mechanical inefficiencies the software has detected and how much each of those faults is costing the company per year.

Once Whitson sits down in front of a terminal, he is suddenly connected to Microsoft’s buildings. Those 500 million data transactions per day can be accessed from the dashboard at his desk rather than from crawling through pump rooms or across rooftops to get data.

Now that the smart buildings software cooks up a chuck wagon of data every day, what to do with all of that tasty information?

The software identifies issues large and small, and even puts them in prioritized order according to how much the problem is costing the company. A majority of problems they can fix right from their desks, and for the rest, the engineers issue work orders (about 32,300 per business quarter).

Apart from efficiency, the surge of data also has made for some eye-popping analytics. These are mechanical engineer Trevor Sodorff’s specialty.

“We have good people, but without good software there are limits to what you can do,” Sodorff says. “Everything lives within the context of the bigger picture.”

One of Sodorff’s party tricks, if you will, is whipping out algorithms to detect new mechanical faults. So during meetings that wander through stretches that don’t pertain directly to him, rather than discreetly checking his email or letting his mind wander, Sodorff writes algorithms.

At one such meeting, Sodorff announces that he’s just written a new algorithm for detecting when the air in a given building is being overcooled. He projects the algorithm on a screen, and then launches into a deeply technical explanation about when a discharge air pressure set point is something-something, then the air is being overcooled by something-something for a duration of 900,000 milliseconds.

“That’s 15 minutes,” says Grove, his fellow engineer, translating on the fly.

Later in the meeting, Grove is talking about how the smart buildings software helps the engineers measure and validate that the energy reduction they’re seeing is due to reduced consumption and not because it was 5 degrees cooler than yesterday. It’s an important distinction for companies to make, especially when seeking a utility rebate.

“We may do an audit, and find we’ve done something that saves 200,000 kilowatt hours, which works out to … uh …” Grove says.

“$16,000,” Sodorff says without missing a beat. (“I studied math for two years before I decided to be an engineer,” he adds. “It’s what butters my bread, if you will.”)

Darrell Smith beams at his team.

“Now you see why developing this software at this scale was a once-in-a-lifetime opportunity,” Smith says. “It affords you the ability to work on some very large-scale, world-changing projects with some very smart people. It’s Microsoft University.”

If the smart buildings tool was developed at Microsoft University, today is graduation day. Where much of Smith’s time the last few years has been spent developing the software, he now spends hours with visiting business, government and industry leaders offering enthusiastic show-and-tells. He’s presented to hospitals, oil companies, automobile manufacturers, cities and federal government agencies – even at the Pentagon and very soon, this same solution Microsoft has deployed will be available to any business.

“Never in my wildest dreams did I think I’d be presenting at the Pentagon. It was a thrill,” Smith says. “It’s been interesting, because I don’t see myself as a salesperson. I see myself as an evangelist for the smart building industry, and what can be achieved with smarter buildings.”

Office buildings, hotels, stores, schools, hospitals, malls and other such commercial buildings are responsible for up to 40 percent of the world’s total energy consumption. In the U.S. alone, businesses spend about $100 billion on energy every year.

“Buildings have been built and run the same way for the last 30 to 50 years,” Smith says. “This isn’t a Microsoft problem, it’s an industry problem.”