Beyond the DVR

Oct. 27, 2008
Storage and retrieval applications are blossoming in the security market.

The security industry continues to be challenged by ever increasing requirements for surveillance applications. Comprehensive security and surveillance strategies are including digital video surveillance as a first option in ever increasing numbers. These strategies are designed to provide users with better threat detection and prevention, a low cost of ownership, streamlined security operations and decreased liability.

Along with this increasing need for surveillance systems is the increased need to effectively manage and use the massive amount of recorded images that must now be stored. Security directors who would manage from a few to hundreds of cameras just a few years ago, must now provide security surveillance services that include several hundred to even thousands of cameras — all of which produce recorded images that must be stored and available on demand.

Traditional Video Surveillance

The classic CCTV implementation that is still in use in many locations has not changed dramatically since the invention of the video recorder. These systems have been used successfully for many years. The typical installation included analog CCTV cameras which had separate wires pulled to them for power and for transmitting images over proprietary wired coaxial connections to VHS video recorders. Add a monitor and the security guard can watch live video or watch playback on the VHS.

Larger implementations of these systems would have security guards watching an ever increasing number of monitors — each of which would present either a continuous image of the area under surveillance or it would “cycle” between several camera fields of view in a predetermined sequence. Practical limitations for the number of monitors effectively screened by a security officer began to drive innovation towards enhancing these systems. Multiplexers were added to these systems to allow the recording of several streams of video onto the same tape, yet separated into discrete viewable streams on playback. Tapes only had a few hours of recordable surface, so multiplexers were used to drop frames in the stream to create the notion of the time-lapse VCR to permit longer recording time coverage on the same tape — although it would do so by reducing the number of actual recorded images.

Many environments still use these systems — even though by today's standards, analog camera and tape systems have a number of limitations. These implementations need manual tape storage and re-use procedures to ensure that retention of the images was available. The process was tedious and prone to human error leading to misplaced or lost information. Video tapes have a limited life span and would need to be replaced often to ensure the quality of the images remained within acceptable tolerances. Keeping the recorders cleaned and serviced meant taking them out of service. Finally, the coaxial cables had limited effective distances which meant that these systems were deployed locally to the tape recorders and could not be transmitted between remote facilities.

The real bad news for users of these systems is that many of the manufacturers of tape systems have opted to discontinue making them as recently as 2005, which makes it imperative to consider alternatives as replacements, service and parts will soon become significantly more difficult to find.

The First Real Innovation: the DVR

Rapid development in video compression algorithms such as JPEG, MJPEG, MPEG and others coupled with lower data storage costs on digital media prompted the creation of the Digital Video Recorder, or DVR. DVR technology comes directly from the computer world using the same disk technology found in servers. Conceptually the DVR is quite similar to the TiVo system you may have connected to your television set at home, and it has several advantages over traditional tape systems.

DVR systems are more reliable since there are no tapes to jam. They offer better video quality as they do not have tape wear problems. They have a low risk of degaussing or signal loss. They can store significantly more images — from weeks to years worth of data. They also eliminated the need for multiplexers as they provided that capability built into the DVR itself. The most important enhancement is that they are effectively automated so that you do not need a human resource to rotate tapes or remember to push the recording button.

The primary reasons for DVR implementations were to address the above issues; however they still have limitations. They still were primarily connected to analog cameras and were therefore still implemented local to the cameras themselves, making remote viewing problematic. They still were limited in the number of camera ports available, as most of them allowed up to 16 channels or video ports. They also tended to be based on a proprietary analog-to-digital conversion capture cards which were not interchangeable with another manufacturer's DVR. Components and accessories were also proprietary. This also meant that the DVR manufacturer was your only real option for service and replacement — limiting the security director's options for consolidating service contracts. Backup and recovery of recorded images is typically not available — if you lose the device, you usually lose the images recorded on the disk.

The advent of the DVR provided advantages over tape-based systems, but more was needed to respond to security industry requirements.

The Impact of IP Based Systems: the NVR

The convergence of IT systems with physical security applications has been widely described. Suffice to say that the incredible growth of IP-based solutions is in parallel with the explosive growth of the Internet and browser technologies. Companies started to see the opportunities presented by using IT networks for more than business applications. Building management systems, fire control systems, HVAC, alarm and access controls, elevator controls, lighting controls, telecommunications, PA Systems, e-commerce applications, and many other previously siloed systems have joined this technological revolution. Physical security is no exception. The move to IP-based solutions on open architectures is under way. IP solutions open a wide variety of options and advantages to the security industry.

The advantages of using IP-based technology are a significant leap forward over previous systems. First and foremost is the simple ability to connect the IP-enabled camera systems to your IT network, eliminating the coax cables. It also means that the ability to transmit those images across geographic areas is now only limited by your network's limitations. Local storage is no longer required. Getting power to an analog camera has always been a major obstacle and cost. The IEEE 802.3af standard for Power over Ethernet (PoE) was designed to address this problem, leading to significant cost savings in deployment of IP cameras. PoE means that networking devices get power from a PoE-enabled switch over the same kind of Category 5 cable that transmits data and video. It also means that cameras can get centralized backup power from the computer room backup power systems, so in the event of a power failure, they will continue to operate. Legacy analog cameras can be connected via a video server (which converts the analog to digital format) to the IP network to allow for migration plans and capital expenditures to be planned within budgets. IP digital cameras do not need to be converted.

One of the drivers in this industry is the sheer volume of video that is being captured. Security directors need to be sure that they record the information that is critical, and not capture information that is of little value. A camera watching an empty hallway is only valued when activity is presented. Intelligent video is one of the next big trends. Network cameras can have built-in motion detection and alarm management at the camera itself. The camera decides when to send video, at what frame rate and resolution, and when to alert a specific operator for monitoring or response. Intelligent algorithms for license plate recognition, people counting, facial recognition, etc., are being integrated into network cameras.

The next significant advantage is that you can base your video storage applications anywhere on the network. This generated the new term Network Video Recorder, or NVR, which has two types of implementations. The first is a “network aware” DVR developed on PC-based architecture, and the second simply uses a standard server which uses the same storage systems and disks as in any other servers capable of storing video imaging. This allows for the centralization of video storage and management in a Security Operations Center (SOC). You can also choose to have distributed SOC controls as well. This also allows for ease of remote monitoring across the network using standard PCs and browser technology for controlling pan/tilt/zoom (PTZ) cameras and video surveillance. Gone are the 50-pound, standalone monitors.

Capacity and scalability have taken a significant leap ahead on NVR systems. Previous DVR systems could typically support 16 cameras, some higher-end systems could support up to 64 cameras. NVR systems can support camera counts from 50 up to 1,000 each. As your needs grow, you can simply add more capacity. Maximum resolution is also effectively unlimited. As digital cameras continue to provide increasing resolution, NVR technology can receive and decode without needing to be modified. NVRs can operate on almost any network topology, including wireless systems. You can install and configure effectively an unlimited number of servers for this purpose. Since the NVR is a server, it can be mirrored, providing you with backup and recovery capability of the stored images, thus preventing loss of the information in the event a server disk goes bad. Also, if the failure is in the server itself and not on the disk, the disk can be pulled and put into another server without losing the data. The typical cost to replace a failed DVR is the same as the unit cost, as they usually need to be replaced at a price that can be in the thousands. The same failure in a disk drive on an NVR server is only a couple of hundred dollars.

Computer industry leaders are making great strides in storage technology. The future of the NVR will be significantly impacted by these developments. Companies like CISCO and IBM, to name a few, are taking traditional IT storage ideas used in the data network world and applying them to traditional video. The following links can provide more information:

* http://www.cisco.com/en/US/products/ps6936/products_white_paper0900aecd804a3e89.shtml

* http://www-935.ibm.com/services/us/index.wss/summary/imc/a1027383?cntxt=a1005270

Spotlight on the Steelbox Solution

The number of companies in this rapidly evolving area can provide a bewildering array of choices — many of which are simply creative uses of existing IT products and services available in the market today. There are also innovative companies like Steelbox Networks that have come up with options of their own in this rapidly evolving area.

To illustrate this rapidly developing market we asked Richard "Chip" Howes, Steelbox president and CEO, and Andrew Flood, lead sales engineer, to provide details about their solutions.

“The Steelbox solution uses only two basic module types to build video surveillance and storage networks of any size — a Digital Matrix Storage Switch, or DMSS, controller module, and a Storage System module. The DMSS modules provide control logic to capture up to hundreds of concurrent video streams for live as well as online storage for presentation to any number of surveillance stations, command and control centers and video walls.”

“The Steelbox DMSS pricing is based on the input bandwidth required for “x” number of streams. For example, a 16-camera system based on Axis MPEG4 cameras would require about 50 Mbps of bandwidth. (This assumes a stream consisting of 4CIF, 30FPS imagery.) The system scales to support 1.5 Gbps of input bandwidth (i.e. – 512 3 Mbps camera streams), 1.5 Gbps of output bandwidth (i.e. – 128 stations at 12 Mbps per station), and 1.5 Gbps of throughput to storage arrays”.

“The DMSS supports redundant power supplies to mitigate power failures and provides up to 14 GigE network ports or ATM connectivity. The Steelbox Raid solutions provide redundant power supplies and supports Raid 0 and Raid 5 configurations for fault tolerance. In addition, the smart load balancing that the DMSS provides allows for Raid units to be taken offline for servicing without disrupting stream recording.”

“The product supports both local and distributed deployment schemes. The deployment flexibility offers the ability to manage the solution and provide critical video to remote locations in a bandwidth-conscious methodology. It also uses a smart load balancing technology that allows Raid arrays to be taken offline, repaired, or replaced without loss of stream recording. It can also be integrated in a variety of network infrastructures to provide the flexibility required for existing installations”.

The scalability of the DMSS and Raid Storage system is impressive. The DMSS was tested by a third party, the Tolly Group, in March, 2007. This single unit was tested as supporting 512 cameras and 80 TB's of available storage.

http://www.tolly.com/DocDetail.aspx?DocNumber=207202

“The system interface, Multi-View, is designed around simplicity for the operator. From a design and integration perspective, the DMSS offers several ways to configure the system, via command-line interface (either remote Telnet session or direct console connectivity) or via Windows graphical interface. The DMSS supports a variety of network infrastructures and supporting configurations. Support for existing analog camera devices can be added using IP encoders. Steelbox provides a comprehensive SDK and API that allows third-party systems to integrate with and control DMSS functionality. The DMSS controls the amount of bandwidth required by offering the option of frame rate reduction and transmitting video across the WAN on event or on demand.”

“The DMSS supports both distributed and centralized models easily. The distributed network model requires additional design and implementation knowledge to support transmission of video across the WAN.”

“Using the DMSS to provide distribution only, system designers can tailor systems around a centralized storage model using video distribution to provide all frames back to a secured environment. At this point, storage via disk drive remains the best option for long term availability of video data.”

Pitfalls for NVR Installations

The biggest bear trap to deal with in designing any NVR installation is dealing with the bandwidth available to you in your network. You really need to work with your IT staff early in your planning to ensure that they have the “pipes” ready for your implementation. Look for their assistance in designing your security solution.

There are many techniques to minimize this problem. Some of them have been mentioned earlier — such as smart cameras which only transmit what you really want, technologies that compress transmissions, distributed networks and NVRs to keep the bulk of the activity in certain network segments, etc. But the bottom line is that video streaming uses a lot of bandwidth. Plan accordingly; remember your IT staff are your best friends here.

The Next Round of Innovations

There are many areas of opportunity ahead of us. One is the adoption of mass storage devices for video data. The IT industry is coming out with new approaches to handle the increasing volumes of data we all deal with.

What does Steelbox say about next areas of innovation?

“It is necessary for the industry to adopt a standards-based format for storing, indexing, and retrieving meta-data associated with video analytics. This will allow best-of-breed applications to develop, allowing for stored and real-time video searches similar to the way we “Google” the Internet for word matches. Without this, I think that video analysis will advance at a very slow pace and take years to reach a point where it is generally deployable”.

Another area of development just over the horizon is actually the human interface with all of this technology. For a glimpse of where we are heading, take a look at the development of Microsoft Surface technology: http://www.microsoft.com/surface/

Imagine using this technology to control PTZ cameras, to identify and zoom in on a suspect, to trigger alarms and lock doors, to retrieve and transmit video images to anyone (such as the authorities responding to your 911 call), to the handset carried by your responding security guards, etc., simply by the movements of your hands over a control application on a “surface enabled” console.

Eduard L. Telders is the director of enterprise information security at T-Mobile. Since 1981 he has served in physical security, information security, corporate contingency planning and safety programs in the banking, insurance and financial industries. He is active in a number of security trade groups and associations such as ASIS, ISACA, InfraGard, the AGORA, CyberGuard Advisory Council, and others, for both physical and information security, and he is a contributing technical editor for ST&D.