A conversation with Mike Frost about intelligent energy management

Jon Udell: Hi, this is Jon Udell. My guest for this week's interview is Mike Frost, who is the CEO of Site Controls. His company is bringing Linux, Java, and a modern approach to networking to the world of energy management. One of the things that ought to be a national priority, but isn't, is the idea of the energy web, a power grid that's as intelligent, reliable and adaptive as it could be and should be. There hasn't been much hopeful progress on this front, but my conversation with Mike convinced me that there are going to be entrepreneurial approaches that can work.

Mike Frost: Hey Jon, this is Mike Frost of Site Controls. How are you?

Jon: I'm good. Thanks for calling, Mike. So, I've just been reading up on what there is to see at sitecontrols.com.

Mike: Okay.

Jon: This is a fascinating project. When I go back and look at what I've read and written on the subject, I am again astounded by the degree to which is this not an issue on very many people's radar screens.

Mike: It's interesting, isn't it?

Jon: It's more than interesting. Given current geopolitical circumstances, it's just nbelievable. But in any event, let's set some context here.

Mike: If you look at it from the highest macro level, energy

[pause] You know, I just saw two fighter jets go by. You don't see that every day. [laughter]

Jon: What airport are you in?

Jon: Oh.

[laughter]

Mike: Whoa, those are pretty bad-ass planes. [laughter]

Mike: Anyway ... the idea is that, if you try to pull power in the middle of a hot afternoon out of a bunch of residences, you're cutting off people's electricity. If you try to pull power off of the grid in large industrial applications, you can't shut down wafer fabs. You end up shutting down rock-crushing plants, and large industries, and they send people home, and it's all very expensive and it all has lots of high consequences. But the real goal for us to be able to consolidate and aggregate the load in all these little buildings. They're convenience stores, they're restaurants, they're retail outlets, and they're fitness centers, and they're all controllable via software.

Jon: So the idea is that, before we get to the point where we have an intelligent energy grid, or an energy web in the large, we can start to see some of these network effects happening. To the extent that your network has an interesting number of sites plugged into it, it enables you to have a macro effect.

Mike: That's exactly right. It depends on the business rules, and the business value to that particular customer. You know, Petco really cares about aquariums. They care about off-hour temperatures, all kinds of things, because they have inventory that's alive all the time.

Jon: Right.

Mike: A cheesecake factory cares about a whole different set of things. They care about how cool is it inside of my walk-in, what is the dishwasher temperature, is it exactly 71 degrees because I always want my guests to be comfortable. They have a completely different set of things that are important to them.

Jon: Sure.

Mike: Then when you talk about Michaels stores, it's a completely different set of things that are important to them. So you're exactly right. Our vision here is to be able to create an intelligent grid of our customers within the grid. If you add them all up, it's about a third of what's in most areas in terms of demand on the grid during peak periods, so it's a significant amount.

Jon: I had been thinking about this more in terms of the intelligent home. And obviously, an aggregate, that would be enormous. But the build-out is going to be incredibly slow, because you have to reach all of those endpoints and instrument them, and that's just going to take forever. So I can see where...

Mike: There's a second key issue there. In a home, if you think about being doing this on-the-fly, using software that reduces a significant amount of load, it's going to entail high consequences. It's going to happen when there's extreme weather, it's hot or cold, and it's going to be at the most extreme time. Whereas if you take a Michaels store or a Party City, those things are great big huge cubic boxes full of conditioned air.

Jon: Right.

Mike: If we go in and push the set points five degrees, nobody's going to notice, Jon. We've run test after test, in fact we're going to be doing an event at one site today, we did one a couple of weeks ago in Texas. We can do the same thing that you could accomplish with hundreds of thousands of homes. But you can do it in these retail environments with low consequence. People don't notice.

Jon: And many, many fewer points of intervention.

Mike: Exactly. We can pull 20 or 30K out of a Michaels or a Party City or a Phillips 66 convenience store.

Jon: Yeah.

Mike: But you're only going to get three quarters of a kilowatt or one kilowatt at best out of a home.

Jon: Right.

Mike: And every one of those retail establishments is equivalent to

Jon: Right. Well, these are parallel initiatives, but they're on different schedules. On the home front, years ago, a guy from Intel was visiting Byte Magazine, and he asked me a question that I've never forgotten, which was, "Why is it that your telephone bill is exquisitely itemized, and your power bill is basically one number?" And I've been thinking about that ever since. To this day it's unclear to me what the payback is going to be for replacing a refrigerator or replacing a water heater. Multiplied by a million homes, that's enormous, and it's obviously gated by not just the energy web in the network, but also by the advent of appliances that are prepared to plug into it.

Now that also applies to what you're doing. And it's one of the things I wanted to ask about. I can easily understand how you can drop a Linux- and Java-equipped Pentium box into a Michaels store, or into a bunch of them, and hook it up to the Internet. But it seems like the tricky piece is getting the sensors and the actuators deployed out to the control points, right? And you're dealing with a variety of legacy and newer equipment, which people aren't going to just wholesale upgrade that, so what's the story there?

Mike: There are two sides to that story. There's being able to build in the ability to deal with old and new equipment on the site side. What we chose to do there was use Java-based open system development running on Linux, and essentially turn every site into a web server.

Jon: Right.

Mike: Once we accomplished that, it's really easy to interface because you're just using standard open systems architecture to do it. For example, The other day, we were talking to guy who runs a really large chain of convenience stores, and his big-ticket item is the Frostee machine when you buy the little frozen Coke products. Ninety percent gross margins for those guys. Whenever that's broken, they really want to know.

Jon: Yeah.

Mike: And so, sure enough it's got a little digital I/O, we were able to figure out some of the error codes so we added that online, we read the error codes, and we send email to the district manager whenever that thing is broken.

Jon: Right.

Mike: Since we've already got a gateway into the site, we've already got a computer there, we've already turned it into a web server, adding that one component is rather simplistic.

Jon: So on the device side of this, rather than waiting for some grandiose standards to coalesce, it's more like three yards and a cloud of dust, right? You take a given class of devices, which have fairly interesting distribution in the world, and you just reverse-engineer how to talk to them.

Mike: Exactly. And it's the same thing whether it's a Frostee machine, a walk-in cooler, or a swimming pool monitoring system, anything that's got an analog or a digital I/O -- it's fairly manageable for us to be able to interface to it.

Jon: How much of this interface work is green field, and how much of it is prior art?

Mike: 50/50.

Jon: Okay.

Mike: The other side of that question, which I think is equally as interesting given the conversation, is that the business rules for every utility are different. Not only is every site different, you've got old stuff and new stuff, but every utility in every region calculates the energy bill a little bit differently. In the future, they'll all be implementing time-of-day, time-of-use pricing a little bit differently. They'll all be issuing demand response ... are you familiar with that term?

Jon: Yeah.

Mike: They'll be issuing demand response signals and pricing signals in a different way, so it all makes sense to us, we just turn it into a web server that understands the business rules there. You go to one place and set the business rules for a bunch of places. Each location can very simplistically modify the way it interfaces with a given utility.

Jon: Right.

Mike: And if the price of power goes sky high, then you apply a different strategy to that site, without a human having to intervene, you know, by pushing a bunch of buttons and telling it what to do.

Jon: So if I'm Petco's operations guy, I come into the management console of your system through the web, and have some representation of those business rules that I can tweak in a relatively nonprogrammer way?

Mike: That's right. And you can select that stuff by state, you can select it by city, because regions have to be very, very flexible. You can aggregate that demand. If I'm the Petco guy in Nebraska, I can see every Petco in Nebraska and exactly how many kW I'm pulling right now. How many kWh I've pulled this month, this week, whatever you want to see.

Our architecture is really different. Since we're a bunch of computer guys from Austin, as opposed to a bunch of controls guys from Michigan [laughter] I don't know if I should say that, but controls guys from wherever they come from, right? The way we designed the architecture, instead of it being a one-on-one building manager that's dialing into a building and looking at an HVAC, we designed it to be persistently connected to the Internet. Then there are decentralized servers that constantly polling those sites. So it goes into every one of those Petcos, and it gets the temperature of every fishtank every 15 minutes, and puts that in the database. So whenever that manager in Nebraska wants to know something about his sites in Nebraska, he's not logging onto each one. That would be incredibly inefficient. He's just going to something like an online banking website, right? Request a couple of reports and boom, there's the data.

Jon: So you're staging the data off of the various devices onto your unit on the site. What amount of pre-existing networking are you leveraging there, and what amount needs to go in new?

Mike: Well, what we are doing is a combination of things. If the network is there, that just makes installation that much easier, we just log on to it. We create our own low-voltage controls network within the site. That is what we hang all of our sensors off of, and we're on the verge of introducing a new wireless technology, based on ZigBee, where everything is going to be wireless within the site from a central perspective if it goes back onto the internet connection.

Jon: Okay. So I have a couple of questions here. But first of all, I noticed that you guys have joined the GridWise Alliance.

Mike: Yeah.

Jon: I did some reading on that, and I find that it is the vision, or part of the vision, that was outlined in that amazing Electric Power Research Institute document that I found a couple of years ago.

Mike: Right.

Jon: Which is an extraordinary piece of work, and again, the fact that I don't think I've heard any politician make any reference to that thing ever is just astonishing.

Mike: Very.

Jon: In any event, a key near-term priority, as laid out by that EPRI document, was the modernization of the grid. The assertion is that simply by leveraging what we have now in terms of communications and computing capabilities, we could make the grid -- never mind all the fancy energy web stuff about demand-response and real time negotiation -- but if we simply made the thing as reliable as it could be, then that alone would cost-justify any upgrades that need to happen. Because the cost of outages is so enormous. We would not be in a situation where every large IT shop feels like it has to have its own redundant power back-up strategy, which is incredibly expensive and wasteful. So if we could rely more on the grid...

...but, well, you can tell me better than I can sort this out for myself, but just poking around I found a link that said "Success Stories" at a DOE site [laughs]. It says "Success Stories--This page is under construction! Please check back soon!" Which doesn't give me a warm fuzzy feeling that there's been a whole lot of progress on this front. So what's your view of that?

Mike: Well, as you know it is complex. My view is this. In order for grid-wide initiatives and the intelligent grid to work in most areas, it will require government subsidy, because these are not things that have strong ROI. And that's why we have chosen to address the specific small segment that we're in, as a way to really get things started. Because the ROI for the customer on our end is pretty fast. In other words, if Petco can pay to build the railroad, and it makes sense for Pecto -- because they use less inventory, they run their buildings smarter, they save energy, and they actually have a less than 20-month ROI -- then they're paying the freight to build the railroad. Whereas in all the other segments, like residential, or many of big commercial bids, there's nobody to build the railroad, everybody's waiting for the government to do it.

Jon: Yep. This is an excellent strategy. It's a way to move forward without having to wait for any big clunky machinery to kick into gear, which may or may not happen in our lifetime.

Mike: Yeah, exactly. And if we're successful -- and we really hope we are -- then the retail stores, the kinds of of national chain brands that exist everywhere, they really end up leading the way for developing the GridWise Initiative, the national intelligent gird, whatever you want to call it. And doesn't it make sense? Because, if you're like me, in every city you get off a plane, you go to the mall, you don't know where you are, this place looks the same as every other place.

Jon: It will benefit us all benefit to have systems hooked up to the Internet in this way. It will also magnify the vulnerability that exists in our critical infrastructure. Not a pleasant subject to discuss but, nonetheless, something that needs to be addressed.

Mike: Well, you know, as a small company, we're trying to be experts in energy management, we're experts in facilities management, we are to a large extent relying on state of the art from a security perspective. I can say that we're doing what we can to work closely with Cisco. We're working with the IT departments -- and the IT groups with these large national brands are pretty good -- to get things as secure as possible. It's very, very difficult to hack into these VPNs that are being built behind the firewalls of national customers. It makes it kind of hard to manage sometimes, because the stuff is firewalled up pretty hard. Doesn't mean things aren't going to happen but, knock on wood, we haven't had anything hacked into just yet.

Jon: From a defense in-depth perspective, though, do you think there are things you could or should add?

Mike: Well, here's what we've done. We chose to deal with it architecturally. Take for example the Petco manager in Nebraska. He's got access to the data center, and he can see data, but he can't really get in and change things at the site level, because he's just looking at that database of information. The national energy manager, based in San Diego, now he's got the higher-level password, but he's behind the corporate firewall when he exercises that, and there's a little bit more security there. So, architecturally, we don't let everybody have access to stuff that can actually go in and change data and change strategies. So our user base is limited, far more than branch banking. You know, I would think the bankers are the ones with the big issues to deal with. We don't have that many users of critical data.

Jon: Oh, yeah. I mean, let's not even go there in terms of banks. That's truly frightening. [laughs] So, fair enough.

One of the things that strikes me here is that there's an opportunity to collaborate with the energy markets in new ways.

Mike: Oh, absolutely. For example, there are certain places where when demand is extremely high, there's not enough distribution capacity. So what they may need to do is reduce demand in a geographic area that is really specific, but their instruments are so blunt that they just do the best they can to reduce demand across their entire service area. That's what most utilities do now. That's what Cal ISO does, that's what New England ISO does, they all do basically the same thing. Our vision is -- and we're not shipping this project to the utility side, but that's where we're going -- is to produce a software product that allows them to essentially dial it down in very specific areas. That gives them the ability to really improve the quality of the grid, if you will. Does that make sense?

Jon: Yes.

Mike: Because now, in specific areas, where there's a lot of congestion, you can say, "What I would like to have happen from two to four this afternoon is I would like every restaurant, convenience store and gas station to shed a third of their load." We can do that. Add all that up, it could be a significant amount of power.

Jon: And if a request like that flows to you, and you then have a variety of customers hanging off of your network, including the Petcos and the Michaels, and a bunch of others, what's the communication between you and your customers in terms of what you're allowed to dial down and how far you can go?

Mike: The web server at each location has its business rules pre-determined. And so when that event happens, we've already thought about what do we do at Level A, what do we do at Level B, what do we do at Level C. Depending on the word that comes down from Cal ISO, that strategy has already been pre-determined. Now, we'll notify customers that's happening. For example, at Petco, when things get really bad and California's doing rolling brown-outs, they've discovered in the past that by putting a sign at the front door that says "We're doing our best to conserve energy today, please pardon any inconvenience if it's a little bit warm and a little bit dark" -- because they lower the lights, they push the set points up to 78 degrees. As a result, they shed 30 or 40 percent of the load that they would normally be using. But they find that from a customer awareness perspective, man, it's great PR!

Jon: Oh, it should be. Absolutely. They should be rewarded and acknowledged for doing that. I agree, that would make them look like they're on the ball.

Mike: Well, at some point there's going to be a new level of social consciousness in these large retail brands. You already see it at Starbucks, you already see it at Petco. We're going to start seeing that out of all of them. It's just going to become the way to behave.

Jon: And probably a good role for government here, instead of -- or rather, in addition to -- the much longer-term solution of the residential problem, would be to provide additional incentives to these guys...

Mike: Right. Absolutely.

Jon:... To do this voluntary participation, right? Is there anything like that happening?

Mike: Well, there are some. Mostly they come in the flavor of standard rebates. The good news is they help offset the cost of a significant amount of putting a system like ours in. They range anywhere from paying 25 to 100 percent of the cost of putting something like this in.

Jon: Oh, really? Okay, and that comes from?

Mike: Fom PG&E, San Diego Gas and Electric, it comes from the utilities. Some of that's government funded, some of it's not, but it's usually the utilities. Almost the same thing as putting up solar panels in your house.

Jon: Okay, so this is very cool. So where are you in terms of the life cycle of your company, and where are you headed?

Mike: We're young. We're a three-year-old company. We like to call ourselves the fastest growing EMS company in America. I'm pretty sure that's true. [laughs]

Jon: EMS?

Mike: Yeah, Energy Management Systems. We've won all of the big deals, that we know about anyway, that have come down the last year or so. We're growing really, really rapidly, but we're a young technology company. We're not a controls company.

Jon: It seems like such an obvious idea, but then perhaps it's obvious to folks like me who come from the computer and networking world, and are accustomed to thinking about these kinds of systems in these terms. And maybe the reason that this hasn't happened to date is, like you said, that the people who are in the HVAC world, their culture isn't to think in terms of network effects.

Mike: It's interesting. I believe the new wave of applications for technology are applications like this, where you go deep in a specific application, deep in a specific vertical, and you really customize the technology. We're using ZigBee, we're using 802.11, we're using Linux, we're using Java, but we're specializing it, so we're bringing all the power of the latest and greatest technology out there and going deep into a specific vertical to solve a huge problem.

Jon: Right. So part of it is network thinking, and part of it -- which is a relatively new phenomenon still -- is the broad availability of a range of componentry that makes it easy and economical to cobble these kinds of solutions together.

Mike: Absolutely. If you look at the Honeywells and the Johnson Controls of the past with customized solutions, before now they would never go into the small footprint. Those guys can't touch a Subway or a Starbucks, you know, they just can't. Whereas, when you're using really low-cost, high-volume computer technology, not only can you get to a lower price where you can address those markets, but you've got far better technology at the end of the day anyway.

Jon: Yes. Well, this is excellent. I think this is one of the greatest examples I have heard of in a long time of a business that is going to, in Tim O'Reilly's phrase, "Do well by doing good." There are not very many hopeful examples to point to in the world of energy, and it's easy to get kind of depressed about the state of affairs, but it's nice to see some really smart and encouraging action.

Mike: Well, good! I appreciate it.

Jon: So thanks very much, and have a good flight, and hopefully we'll talk again.