THE IRONY OF INTEGRATION
According to the Security Industry Association (SIA): Systems integration means whatever the customer says it means. It’s a bit ironic for SIA’s Systems Integration Industry Group to have arrived at that definition, considering that systems integration was the buzz word of the millennium.
Instead of being hung up on trying to further define it, the industry today has turned to harnessing its full potential. However, while true plug-and-play systems integration is what integrators and their clients really need, competition among manufacturers may be what’s holding such development back.
Among current advancements, what is considered cutting edge in integration varies from manufacturer to manufacturer. To some, it means using a common protocol or computer language to interconnect systems from different companies in a plug-and-play fashion, similar to the way in which a printer usually can be configured with minimal difficulty to work with many different brands of computers.
To others, cutting-edge integration means developing quick and easy ways for other companies’ devices to interface or interconnect with your company’s system through software development kits (SDKs) or application program interfaces (APIs) without revealing too much about your proprietary system.
Many manufacturers provide dealers with SDKs which allow dealers and integrators to connect to specific manufacturers’ products. In one example of such an “open alliance,” Imprivata, which manufactures an IT/physical security product, currently works with Tyco, Lenel, and S2 Security and is working to add more vendor partners.
“I think [open alliances are] a very forward-thinking way to approach this,” says David Heinen, product marketing manager for enterprise systems for Bosch Security Systems Inc., Fairport, N.Y.
But that door hasn’t opened far enough, he maintains. “They are providing a very slim connectivity to allow you to have some functionality. You could monitor what is going on, but not make any change unless you actually booted up their software,” Heinen asserts. “You may be able to open a door, but you wouldn’t have a single unified database. You’d have their database and a separate database for our product.”
BEING TAKEN FOR A SPIN?Security manufacturing companies may need to connect to other companies’ systems for several reasons. It may be that they cannot supply an entire complement of products themselves, or because their customers are engineering systems that include a mix of different manufacturers’ products, or simply because legacy systems may have equipment that is no longer available.
“We’d love to connect to our competitors, and sometimes we’d like them to connect to us, but when we have a project being sold, we’d like as much of the project to have our equipment in it as possible,” Heinen admits. “Most companies, regardless of how they want to present themselves as open â€” they are proprietary.
“If you want management from a single screen, you really have to use all our products,” he concedes. “We may be able to connect to other things, but you’re going to have a separate screen. You may be able to manage it, but you can’t control it.
“Our competitors have that very same issue,” he declares. “They can connect some way through an API or SDK, but they’re not going to have full and complete control, because each manufacturer wants to keep some cards back. We’re not to the point where the IT industry is, where they have open connectivity. We’re still a little proprietary whether we care to admit it or not.
“People say we’re ‘open standards,’ when in reality they’re open to their own products and have a small connectivity to other people’s products, but not at the level you would expect,” Heinen comments.
Bob McCarthy, vice president of technology for Dedicated Micros Inc., Chantilly, Va., points out that it is in his company’s interest to integrate with as many systems as possible, but he is not sure that is true across the industry.
“The biggest companies may really not have much of a financial interest in a standard, because they do have access control and DVR and central station monitoring, and all those products, and it’s interoperable and their proprietary solution, so they’re not looking for competition,” he declares.
“A lot of people claim to be all for open integration and openness, and the proof is in how many other systems they’ve integrated with to make life easier for their customers,” he continues. “A lot say they’re open, and they’ve listed the same two or three partners for the last four years because they really aren’t interested.
“A lot have house-brand DVRs, and so they don’t want to integrate,” McCarthy maintains. “Others say ‘we’re open access,’ but then they charge $15,000, which is a lot of money, for the SDK. So the proof is, how many systems are they integrated with?
“It’s in our best interest because we are a video surveillance company,” he emphasizes. “That’s really a strength of ours. If our parent company bought an access control system, and that’s the only one we were allowed to integrate with, that would be a distinct disadvantage for us.”
Heinen thinks the rise of middleware companies may be aided by this reticence on the parts of major security companies to interconnect their systems.
LOWEST COMMON DENOMINATORDeveloping systems that integrate different companies’ components has been difficult, McCarthy thinks. “I know there have been some companies that have attempted to create multi-vendor integrated solutions where you can choose your DVR and access control, but I don’t think any have made it because there are no standards in place,” he laments.
This necessitates custom integration work for each individual vendor, and when that vendor rolls out a new product or a customer switches to another, the SDK may change significantly, he notes.
“So it’s really hard for those all-in-one companies to stay caught up to the latest products,” McCarthy maintains. “It’s a lot of work just for us to work with all our integration partners and keep those up-to-date.
“While we are very open with who we integrate with, we do try to stick to people that either have a decent market share or an exciting new technology,” he emphasizes.
One possibility for close integration of security systems that the Security Industry Association (SIA) mentions in its definition of open systems integration (OSI) is use of an extensible markup language, such as XML, to integrate different security applications.
“I know it’s been done successfully in other industries,” points out David Weinbach, product manager for access control solutions, Lenel Systems Inc., Rochester N.Y., a UTC fire and security company. “There’s a tradeoff whenever you try to standardize configuration because the unique benefits from each vendor tend to be unsupported in the XML.
“By nature, it has to be the lowest common denominator data that everyone can share,” he points out. “So you might get the basic functionality from each device, but the unique feature that caused you to pick it in the first place might not be available with XML. But it’s still possible once the device has been turned on with plug-and-play to go into some vendor-specific software to customize its behavior.”
That customization is a necessity, McCarthy thinks. “This industry is constant integration, constant movement. The open standard they define needs to give room for new features and custom features that certain vendors create, because otherwise the standard won’t be used,” he says.
This would make any standard irrelevant, he fears. “You usually can add something on most of these standards, such as a custom command to be defined by a vendor, and the vendor has to provide very specific instructions on how it works and follow something very similar to the test protocol,” McCarthy asserts.
He suggests agreement on a simple point as a strong first step, such as standardizing use of Greenwich Mean Time, on a time stamp. “It’s hard to define something and get agreement from all the big and small companies,” McCarthy concedes. “If you’re forcing [manufacturers] to totally change the way they do things, there has to be some benefit for them. We’re for, not against it, but I see it difficult getting everyone to agree on the spec.”
A BABEL OF LANGUAGES?“We’re part of the SIA working groups,” points out Peter Boriskin, director of product management for access control in Tyco Fire and Security’s access control and video systems business unit, Boca Raton, Fla. “Some of the things [SIA is] looking at are SPML (service provisioning markup language). SPML is an extensible set of classifications based on XML, but is particularly for provision of services to a user.
“It does come from the IT side of the world, but the methodology is one we could use in physical security just as easily as they could use in logical security,” Boriskin maintains. “SAML (security assertion markup language) also is an extensible language based on XML that can grant users permissions and access to assets. Again, it comes from the IT world, where those assets are devices, databases and logical definitions, but there is no reason those couldn’t be about physical devices as well.
“Since [SAML] is an event management system, it really doesn’t necessarily pigeonhole us anymore into the domain of physical security management,” he asserts. “There’s nothing unique or specific about it for physical security. We happen to have deployed it as a security management system, but nothing is precluding us from managing any other type of event.”
This means that logical security could be provided as easily as physical security.
“A lot is to be determined, but it boils down to either choosing and expanding it to physical security or going back to traditional XML and then creating a unique object classification that is specific to physical security â€” so when you talk about a door and I talk about a door, we know what that means and what constitutes a door,” Boriskin describes.
“Right now, there is no standardization to the level SIA is talking about,” he continues. “It’s a bold first step and I think it’s great, especially when you want to get to a place where you have role-based access control irrespective of whether physical or logical. So based on [an end-user’s] role in the company, they should have access to certain assets, whether a database here or a car park there.”
His company also is considering using XML as another means to integrate products in the future. “It’s just pretty amazing what you can do with XML,” he concedes. “Our preference is to use OPC, but we have used SDKs in order to connect to other people’s systems on different projects.”
The actual language used is less important than the underlying infrastructure, stresses Robert Siegel, general manager of video and software solutions for GE Security, Bradenton, Fla.
“.NET under Windows is basically a set of application interfaces, and the language becomes less important,” Siegel maintains. “What is the underlying infrastructure that allows these applications to talk to each other? XML is just a language for moving different types of data back and forth.
“Linux-based platforms also allow for some ease of use and interoperability,” he notes. “Let’s go back to IT networks, where you will have Apache servers talking to Microsoft exchange servers talking to Windows on a desktop PC.
“All those devices can talk to each other through a variety of SDKs and APIs and sometimes .NET or XML,” Siegel points out. “There’s not one right answer â€” it’s about acting, thinking and behaving like a network.”
COME OUT AND PLUG & PLAYTrue plug-and-play integration is still in the future, Lenel’s Weinbach thinks. “To my knowledge, there’s no plug-and-play integrated solution, but there are plug-and-play parts,” he states. “There’s limitations on plug-and-play, because not everybody wants to play.”
But Mike Garcia, vice president of marketing, MDI Security Systems Inc., San Antonio, Texas, thinks his company has achieved plug-and-play integration and can make changing components easy.
“With some systems, when you do something as simple as change a camera out on a DVR, you have to call the integrator to change the camera DVR settings and head-end setting,” he maintains. “In a unified system, you change that camera, and the system recognizes it and adapts the camera in, and you go. There’s no major engineering expense for software upgrades.”
Jim Lowder, MDI’s vice president of engineering and chief technical officer, explains further. “It’s the same concept Microsoft employed with plug-and-play technology,” he relates. “We’re going to create a standard method to interface with us. If you want to be plug-and-play, you follow these specifications. It’s also very, very simple to work with other companies.” Integrating with other companies’ products should only take a few hours, Lowder thinks.
Another benefit is ease of training. “With this type of platform, your operators only have to know one system,” he asserts.
“There’s definitely drudgery, like setting addresses of devices, that will be simplified in the future,” he predicts. About the issuing of IP addresses, he adds, “It should happen behind the scenes â€” you shouldn’t have to worry about it. You still need to configure the device once it is on the LAN, but you don’t have to enter any address to get them on the LAN,” he points out. “So it’s removing the drudgery part but keeping the flexibility that you want.”
Some network video recorders, such as models from Panasonic Security Systems, have that feature built in, points out Steve Surfaro, director strategic technical liaison for Panasonic Systems Solutions Co., Secaucus, N.J.
“This is a very integrated product that actually has a built-in application to recognize a camera that is on the network and to automatically set up the camera,” Surfaro points out. “So it allows direct access to that camera, recognizes it, sets the IP address of the recorder and allows you to start recording immediately. So it’s making IP-addressable cameras plug-and-play.”
Following the lead of the IP industry, it’s probably inevitable that the security industry will become more open over time, Heinen predicts. “A company like ours needs to sell products in order to afford to do the R&D so we can keep going forward, and the best way to do that, we have found, is to try to get as many people to buy as many of our products as possible.
“I think that’s a reality,” Heinen insists. “Anyone who says it’s different is trying to spin you. It’s a marketing effort more than a reality, just in my humble opinion.”
Sidebar: Head to the WebEnthusiasm about Web-based solutions is expressed by Bob McCarthy, vice president of technology for Dedicated Micros Inc., Chantilly, Va. “One thing we’ve found is you need more than one type of integration capability out there, because we go to work with various partners at customers’ requests,” he relates. “We find one wrote their program in Visual Basic and another in Java and another in .NET and another has a Web-based solution, and they all want to integrate with us.
“So we’ve developed Web-based integration tools, Java Applets and Active X controls, as well as Windows and Java SDK, because it’s not our choice what the other partner is using,” he points out. “We need to be prepared to offer up the tools they need to work with their product.”
John Smith, senior marking manager, Honeywell Access Systems, Louisville, Ky., thinks cutting-edge integration would include Web access. “Where we’re headed is integrating everything together, and putting it on the Web, and making it available, so you have access from anywhere with real-time control,” Smith describes.
“You can control your cameras, arm and disarm your system, open and close doors,” he lists. “Tying it all together and driving that single interface is ideal for the customer. Putting it up on the Web is the big step for the next generation of product. Nobody is doing that now. In my opinion, we’re probably the closest we’ve seen to it.
“What we’ve seen with the competition are components of access control, video and intrusion, but not a complete system that is integrated fully across the board,” Smith maintains. “We see everything is moving IP-based, everything is moving Web-based, and all our products are moving in that direction, and this is going to continue.”
He thinks individual components such as access control readers have more open architecture so they can work with many different systems, but actual system control is still proprietary with most companies.