On duty: 24-hour video guards

Call it the changing of the video guard. Analog cameras, coaxial cable and tape have served as the standard components of video surveillance for decades. Closed-circuit TV (CCTV) soldiers on as a widely used solution for monitoring buildings, highways and other areas for security and public safety. But although those systems are serviceable, they have limitations. CCTV’s nature hampers the sharing of imagery. Videocassettes are unwieldy to search, they require frequent replacement, and the CCTV network is yet another piece of infrastructure to manage.Digital video surveillance, however, promises to break through the limitations of analog CCTV. Digital systems enable remote viewing through standard IP data networks, so multiple parties can observe the same imagery. Digital video is easier to search and doesn’t degrade the way analog tapes do. Furthermore, the technology allows video surveillance to mesh with an organization’s information technology infrastructure and opens opportunities for analysis unheard of in analog systems. For example, you can set up a digital system to generate alerts when it captures suspicious behavior.Police departments, transportation authorities and other government agencies have already embraced digital video surveillance. Kirk Talbott, chief information officer at the North County Transit District in San Diego, cites flexibility as digital video’s main benefit. The transit organization, which operates a commuter rail line, has begun deploying digital cameras along its right of way. “You get the ability to leverage the existing IP infrastructure,” Talbott said, adding that IP networks can deliver video to any location where it is needed. Despite the obvious advantages of digital video, making the switch is often difficult for large organizations that have a sizable investment in analog equipment. To keep costs down, many agencies introduce digital gear incrementally or operate hybrid systems that incorporate analog cameras into a digital backbone. Another difficulty in switching from analog is the dizzying number of digital video options. Agencies must evaluate a wide array of IP cameras with differing capabilities and consider numerous solutions that incorporate encoders, switches, and various storage and recording options.  “The array of technology and devices that are out there under the rubric of digital video surveillance is enormous,” said Dirk van der Vaart, vice president of the National Security Strategic Business Area at American Systems, an integrator that specializes in digital video surveillance. Given the variety of solutions available, van der Vaart advises buyers to make a business case for video surveillance before focusing on a specific solution. He said an agency’s selection of one system instead of another should be based on the answer to this fundamental question: “What is it they want to protect and why?”Sam Docknevich, director of IBM’s Digital Video Surveillance business, said a threat assessment can help an organization define what it is up against, and a risk analysis can shed light on how much it should spend on a security solution.Threat assessment is especially important when selecting the camera component of a digital video surveillance system. “When you look at the cost of a system, cameras are typically the highest percentage of any installation,” Docknevich said. Camera costs and capabilities vary widely. IP cameras can cost a couple of hundred dollars to $4,000, said Jim Clark, vice president and general manager of Verint Video Systems.A digital camera with 640 x 480 resolution is considered entry level, said Kirby Russell, director of product marketing at Strix Systems. A camera with that resolution is comparable to an analog one rated at 400 to 500 TV lines.Megapixel video cameras offer higher resolution levels. Russell said 1.3 megapixel cameras, with 1,280 x 1,024 resolution, are at the lower end of the high-resolution choices.  So what type of camera should a security-conscious organization buy? “There’s no stock answer,” van der Vaart said. A camera used to detect motion and trigger an alarm doesn’t have to produce high-quality images, he added. But if an agency wants to use video imagery to identify and prosecute an intruder, it will have to invest in a high-resolution camera. Another consideration is whether to buy a camera with a fixed gaze or one that can pan, tilt and zoom. The latter is more expensive, and customers tend to use it sparingly.The Georgia Department of Transportation (GDOT) uses 1,300 fixed, black-and-white cameras to gauge the flow of traffic and about 300 color cameras that pan, tilt and zoom to zero in on trouble spots, such as car accidents. Most of the cameras are analog. However, Hugh Colton, Intelligent Transportation System projects manager at GDOT, said the department is gradually converting to digital technology.“We keep improving our specs,” he said. GDOT’s cameras use charge-coupled devices (CCDs) as image sensors, but the department might switch to cameras that have complementary metal-oxide semiconductor (CMOS) sensors. CMOS sensors cost less than CCDs, and their image quality, which used to be inferior to that of CCDs, has improved.“It is my contention that…GDOT will be moving to CMOS in a year or two to take advantage of the cost and quality benefits,” Colton said. The IP camera is the front end of a video surveillance system. The camera’s live video stream travels across an organization’s IP network, where it can be viewed in real time and recorded on a storage device for later playback. Storage options include network video recorders (NVRs), which store video on a hard drive. Such devices fill the niche occupied by videocassette recorders in analog systems. However, specialized appliances aren’t required. Customers can instead choose to store video on a PC hard drive or a network storage device.Docknevich said customers buying IP-based digital surveillance systems favor enterprise storage over proprietary devices such as NVRs. That way the stored video has the same backup and disaster recovery protection afforded the other data on the network. “There’s more stability and more security,” Docknevich said. Treating video like any other enterprise application also helps lower the cost of surveillance solutions, he added.Another infrastructure component, typically found in large installations, is a switch that distributes live or recorded video streams to multiple monitors. Those monitors might be PCs used as surveillance consoles or more expensive video wall displays.The video’s journey from camera to display often takes place across the same IP network that handles an organization’s other data traffic. “Most organizations that I work with are not putting in a separate network just for video,” Docknevich said.The use-what-you-have philosophy, however, places additional demands on networks and storage devices. The Regional Command and Control Communications (3Cs) project in Southern California illustrates the effect. The project distributes video from helicopter-mounted cameras to participating agencies via a 155 megabits/sec IP-over-microwave network. Participants include San Diego’s Police, Sheriff’s and Fire-Rescue departments, and the local office of the California Department of Forestry and Fire Protection.The video streams consume two-thirds of the network’s bandwidth, said Sara Diaz, 3Cs project manager and a contract employee at the San Diego Police Department. The video streams must share that bandwidth with other applications, including videoconferencing and voice over IP.In the next few months, however, the project will adopt Multiprotocol Label Switching to address the bandwidth-sharing issue. MPLS boosts an IP network’s ability to manage traffic. “We’re going to MPLS soon, and we will put each one [multicast video and videoconferencing] in its own channel,” Diaz said. That approach will improve performance, she added.Industry executives say most organizations find ways to manage the bandwidth demands that digital video creates. Tony Lapolito, vice president of marketing at VidSys, said IT shops reluctant to add video to their networks are reminiscent of organizations that balked at adopting voice over IP. For that technology, forward-thinking customers employed techniques such as virtual local-area networks to separate voice traffic from mission-critical data traffic, Lapolito said. Organizations can use similar best practices to deliver video surveillance over IP, he added. In addition to engineering techniques that help maintain network performance, improvements in data compression are making digital video less of a network pariah. Two years ago, IT managers kept video off LANs because they were concerned about it degrading network performance. But today’s digital video is highly compressed and easier to manage, van der Vaart said.The improvements in compression come from H.264, also known as MPEG-4 Part 10. It has become the algorithm of choice in video compression, said Sean Lessman, senior director of advanced technologies at Tandberg. The algorithm, which became a standard in 2003, lets users obtain about the same picture quality for half the bandwidth, Lessman said. Although compression eases the demands on storage, they remain considerable. “There’s no more intensive storage requirement than video,” Clark said. The amount of storage an organization needs depends on variables such as image quality, frame rate and the number of cameras in the system. Even entry-level systems can require 1T of local storage, Clark said, and a large facility with 60 cameras might require 40T.Building a purely digital video surveillance system is challenging enough, but most organizations don’t have that option. Instead, they use their existing analog cameras on an IP-based network, which means they must install encoders to convert the cameras’ analog signals into a digital stream that an IP network can carry.“Hybrids prevail,” said Rich Mavrogeanes, founder and chief technology officer at VBrick Systems, which makes appliances that handle video-encoding chores. In a hybrid system, an analog camera attaches to an encoder via a coaxial cable. The encoder compresses and digitizes the video stream before transferring it across the IP network.Organizations can deploy encoders as appliances at prices that generally run into the thousands of dollars. For instance, a VBrick network appliance starts at $4,500. For a few hundred dollars, agencies can buy a video-encoder card to install on a desktop or server PC.  GDOT plans to eventually have an all-digital network as part of its intelligent transportation system. In the meantime, the department might replace aging video switches with encoders that will digitize signals from analog cameras. Nearly all of the video switches are obsolete, Colton said, and the manufacturer is no longer in business. The 3Cs project uses VBrick encoders to digitize video streams from helicopter cameras. The cameras send video via a downlink transmitter to a receiver on a mountaintop, which also serves as an entry point to the project’s communications network. Diaz said local agencies bought video equipment for their air units that was beyond the project’s scope. But via the network, the project can use the camera equipment already in place.  Many organizations are in a similar evolutionary state with regard to video surveillance. “We rarely find someone who is ripping out the entire infrastructure,” said Mariann McDonagh, vice president of global marketing at Verint. “It is important to see this market as an evolution and not a revolution.”