The IP opportunity
Feb 1, 2005 10:26 AM, By Tom Corbett
IP connectivity is the next big feature of the newer audio and video products. Trade shows are abuzz with words like "IP" and "network capable." Some devices have internally assignable, static IP addresses and can be truly networked. The main benefit is that audio and video can be moved from one location to another using unshielded twisted pair (UTP) wire.
UTP connectivity brings a new dimension to audio and video trades, since any building's UTP wiring infrastructure can be used for connecting rooms without costly, time-consuming wiring upgrades. The greatest advantage of UTP is that audio, video and control signals can be handled just like computer data using computer technology. Harnessing the power of the computer provides extensive signal-processing capability and uses inexpensive hardware.
As promising as this new hardware and software technology is, the real challenge is implementation. These new digital devices are more and more commonplace and offer greater and greater power, but implementation requires coordination. Without a properly designed dedicated network, video will be latent, which translates into a lower frame rate, jerky video display, a smaller image on the screen, or lower resolution. A dedicated network will provide full motion video and clean audio.
A Project Gone Awry
Pro Media, a San Francisco bay area AV contractor, installed the audio and video systems in a new building on a major campus that shall remain nameless. Besides local presentation, the campus needed interdepartmental sharing and even shared research with off-campus colleagues.
During construction meetings, locations were established and rack footprints were discussed. Conduits were coordinated, power requirements were outlined, and drops were installed. Then the building was built. At the last minute before the system went online, the network administration representative realized that the burden on the building's computer network was a greater unknown than previously realized. Access to the network was denied. The faculty was frustrated. Researchers were frustrated. Noone was happy.
The lessons of this project are many. The main lesson is that the IT administrator has the prime responsibility of keeping the network running for the building occupants. Audio and video accessibility are secondary to computing and maybe even IP telephony.
Before installing a network over UTP like this one, all demands on the facility's main computer network should be examined.
Computing burden: A reasonable computing environment's average daytime computing traffic might be close to 30mpbs, but an Ethernet network is considered saturated at around 40% of the capacity of the 100mbps switch gear. The computing traffic is instantaneous and constantly fluctuating, and the information technologist is required to keep the network running at transparent speeds. The only way to verify traffic capability is to connect a network traffic generator and test the results.
Email burden: The amount of email traffic has risen substantially in the last few years. Email can be internal or external to the building. Email is indistinguishable from computing traffic: Documents, spreadsheets, databases, and CAD projects might be the stuff of business, but web searches, emails, and memos are the communications that represent research as well as products. Part of the email bandwidth requirement is within a building, but it also represents a part of its interconnectivity with the outside world. Cisco cites the 20/80 rule: 20% of the network traffic is internal and 80% goes outside.
IP telephony network burden: The ability to connect calls on the network is particularly attractive to large businesses with several office sites. A great deal of coordination is required to set up such capabilities. Interestingly, much of multi-site connectivity is carried on T1 phone lines. This means higher call density and lower cost because calls are not charged on a unit basis. This is a bandwidth requirement for some networks. Unlike data, however, if a voice call were to be bogged down because of bandwidth restrictions, the resultant call latency would be totally unacceptable.
Building management network burden: Building management systems have been migrating to computer networks for some time. This has not been a burden to networks because these pulses of information connected with control are small packets that can easily fit into a larger scheme of information. Building management can include lighting controls, heating and air conditioning, and security systems. Interestingly, even security cameras are now being connected to networks and passing video as well as control for pan/tilt camera heads, door locks, and card readers.
And Now to AV
Thus far, the higher bandwidth requirements of video and its higher clock rate needs have limited the entry of AV to the network. However, analog-to-digital input converter and the digital-to-analog output converter clock rates have improved and made higher audio fidelity and full motion frame rates possible. AV's demands on networks are different from those of regular business applications. Factors to be considered include:
Audio Burden: Analog audio's clock rate is around 20khz. In order to digitize audio, its clock rate must be at least 44.1 or 48khz. These older clock rates have been largely replaced by 96 or 128kbps and even 192kbps to provide a better reconstructed audio signal at the eventual loudspeaker. Higher clock rates increase the data requirement on the network. One of the major audio transport technologies, Cobranet, will put 1.5mbps worth of traffic on a network to ship eight channels of audio.
Video Burden: The bandwidth requirement for video is much higher than for audio. Compressed video will operate at 3mbps, while uncompressed video has a bandwidth (serial bit rate) of 8 to 18mbps, with 9mbps being a good baseline number. Sending one channel of video to be viewed on the network requires enough bandwidth to recommend a subnet. Advanced video signals require greater capacity. HD video signal operates at 36mbps.
AV on the Network: An audio or video signal broadcast on a network allows for that signal to play over the network, permitting anyone to view the stream as it arrives at their computer terminal. Video on demand, however, means that a user requests a file be played and the file streams to that computer. Another user might request the same file a fraction of a second later and the file starts again for that computer. The network traffic incrementally doubles for this use alone. Because the local computer buffers the signal, though, network inconsistencies are not necessarily noticed.
A network's bandwidth can be enhanced by constructing subnets. Adding complexity to any system, however, means that complications are introduced, which can add to potential failures. If the network is not comprised of all new equipment, legacy equipment can be an operational unknown. Because networks have a lot of equipment, the chance of having different levels of technology in the switchgear, the hubs and routers, or any other device is greater and increases as time goes on.
10/100/1000 network with CAT5e or CAT6 cable: Claiming to have a signal rate of “gigaspeed,” this network's real rate is determined by connectivity and wire condition. Improper terminations will drop the speed of the cable dramatically. If the cable is stepped on or crushed during installation, it will not pass data at the same rate. If the cable is pulled with too much force, it may reduce video frame rate. Installed UTP cable is tested with a time domain reflectometer (TDR) to assess its installed condition. This device will provide a printed readout, which identifies not only the speed capability of the individually tested cable, but also where in the measured distance a loss occurs.
10/100/1000 network switchers: Routing switchers in the network have an operational speed capability. Hubs will not work for video because they have a dedicated bandwidth and broadcast signal to all connected users, whereas switchers use micro segmentation to provide a signal where it is needed.
One approach to handling audio and video netword demands is to configure subnetworks. Users in one area can be separated into a subnet where the higher local traffic does not affect the basic building backbone. If the viewers are widespread, the next step is to assign quality of service levels to the signals. Queues can have multiple levels of priority, with higher priority going to time-sensitive data. A higher level of service priority might be assigned to a video signal, where loss of data would result in a decreased frame rate for the viewer. Note that MPEG video playback is not rate adaptive. An application such as video conferencing, however, is rate adaptive because the codec can adjust the frame rate as needed.
Probably the best solution is to provide a separate wire connection. Dedicated CAT wire will be free of any other burden and will operate without interruption. If, however, the identified user’s computer must be connected to the regular network for normal computing and also connected to a specialized network, a server will be needed to select the appropriate signal connection. Most often, heavy video and audio traffic devices have a CAT connector only for connecting to another similar device. Although this RJ 45 connector appears to be a standard Ethernet connection, it is actually a proprietary connection between like devices. (Symnet is an example of this type of dedicated use of CAT wire between audio devices sharing digital audio.) If a device has an IP address, it is looking for an IP Ethernet network connection for interconnectivity with non-dedicated processor environments.
Optical fiber has much more bandwidth to offer. Optical carrier 3 (OC3) is an interbuilding connection over optical cable at 155mbps. A major campus link or a link to a service provider might have OC12 with a data rate of 622mbps. Copper limitation is DS-3 with 45mbps.
Optical fiber, however, needs transmitters at the origin of the fiber and a receiver at the destination. The cost of transceivers is high and may be more than the cost of the audio mixer at the origin. Fiber is an end-to-end connection and does not go as many places as UTP networking.
Comparatively, the fiber network may not be connected to the desktop because of the cost of the encoders and decoders. Fiber may be the best connection for the bandwidth requirements of an auditorium display, but this is not the everyday computational load. Without fiber, the desktop environment may well depend upon CAT cable. The answer may be at least two networks.
Any new facility should consider a dedicated network for audio, video, and associated control. Although control represents a very small network activity, it is generally linked to the video or audio signal and would need the same connectivity.
Other Questions to Consider
Don't forget these details in your planning:
- Who maintains the AV switches as opposed to the data switches and hubs?
- Who sets the color standard for all the CAT wire and RJ 45 receptacles?
- Who gets what locations in the MDF and IDF rooms?
- Can the systems cross patch?
- Will the two (or more) networks racks stand next each other in the MDF or IDF room?
- Can users cross-connect receptacle functions?
- Who will control the tel/data room so that it doesn’t become an uncontrolled hangout?
When designing a network, the ongoing construction meetings often bypass the larger issues while a general contractor stands waiting for an answer to the request for information. The issues discussed in this article are for early programming discussions while the new building is being designed and the occupants are looking for solutions to their individual and group needs. Address these needs. Don’t build a monument to unfulfilled desires and frustrated dreams.