Expert Tips on Networking Video
While the functions of video surveillance cameras and the benefits of networking are well known to the great majority of SDM readers, connections of such devices onto networks and the Internet are not as well understood.
Here are three common questions dealers have about network video system set up. SDMâ€™s contributing technology writer, Dave Engebretson, who has written a book on the subject, provides painstakingly detailed answers that technicians can use in the field to save hours of labor.
Packet loss is when individual TCP/IP data packets get â€œlostâ€ during their travel from one point to another on a network. Packets may be corrupted during transmission, particularly over the Internet, and the network switches along the path will discard those packets that are defective.
Packet loss can be tested using the ping -t test from the Command line in Windows.
After typing this command and pressing enter, the testing computer will initiate the ping test, but instead of only sending four packets, the ping -t test sends packets continually until it is stopped by simultaneously pressing the control and c key.
This test should be run for a minimum of 30 seconds. A longer test, say a minute or two, will provide a statistically more accurate test. If this test shows any amount of packet loss, there may be some problem with cabling, connectors, network congestion, or electromagnetic interference that is causing LAN communication problems.
Letâ€™s look at a packet loss test for an Internet connection. In this case we are testing from Chicago to a computer in Germany: (see Figure 1)
Here we see actual packet loss, as the â€œrequest timed outâ€ line indicates a lost packet. At the bottom we see compiled statistics showing the percentage of packet loss, and the latency of the communications between the two computers. Note that a 12 percent packet loss is very high, but this report reflects a very short testing time. The same test run for two minutes resulted in an 8 percent packet loss figure.
Itâ€™s important to remember that packet loss problems over the Internet may be temporary or sporadic as ISPs dynamically manage their huge amount of Internet data traffic.
There are a number of handy tools to use when troubleshooting a particular Internet connection. One website, www.dslreports.com, provides a wide variety of Internet connections tests. After establishing an account (which can be free for limited services or upgraded for a charge), users can run a Line Quality test to view the quality of their Internet connection. Running the Line Quality test usually takes about 10 minutes. Letâ€™s look at two sections of the report from a recent test: (see Figure 2)
This section of the report details a test run by dslreports.com that provides information regarding each hop or Internet router that their test packets are passing through from the West Coast (U.S.) to the testing location in Chicago.
Available bandwidth will directly affect the amount of data packets that can be transmitted from a network video device, and the quantity of packets that can be received by a remote viewing computer over the network and/or Internet. Bandwidth is typically characterized in bits per second, such as kbps (kilobits) and Mbps (megabits per second).
For network video servers or cameras that will be viewed over the same LAN, the available bandwidth is a function of the cabling, connectors, and network devices such as hubs and switches. While the available bandwidth of a LAN is usually more than adequate to transmit multiple video data streams, security integrators need to work closely with IT personnel so that the video streams donâ€™t overwhelm the network and cause slowdowns or problems with other network traffic such as enterprise data.
Letâ€™s look at one set of test results from the authorâ€™s DSL ISP (see image).
This report is typical of DSL connections. While the download bandwidth is 1.26 Mbps, the upload bandwidth is only 147 kbps, or roughly one-tenth of the download bandwidth. DSL services are typically configured to provide much greater downloading capability than uploading, as most Internet usage consists of reaching a Web page with a few mouse clicks and downloading large amounts of data such as images and audio.
Here is the same test on the authorâ€™s cable ISP:
Here we see the superiority of cable modem Internet connections over DSL, at least with the authorâ€™s particular ISPs at the time of testing. The cable modem provides 2.50 Mbps downloading, and 359 kbps upstream connection bandwidth.
Sharing of the Internet connection by others while the video stream(s) are being transmitted and received will also affect the available bandwidth. This may result in high-quality video being remotely received during off-hours, while the remote video signals slow down or degrade during business hours when enterprise users are connecting to the Internet.
When connecting IP-enabled video devices for over-the-Internet viewing, both the downloading and the uploading available bandwidths must be considered. At the transmission end, the uploading bandwidth is what matters, as the amount of bandwidth will directly affect the amount of video that can be sent up to the Internet. On the receiving end, the downloading measurement indicates how many fps can be viewed at the remote location.
Available bandwidth also can be impacted by features of the LAN at either end. For example, if the remote viewer is using a Wi-Fi wireless computer to access the network camera or DVR, the amount and quality of the video may be reduced based on the quality of the Wi-Fi connection. At both the transmission and receiving end, proxy servers or firewalls may reduce the amount of video that can be sent up to the Internet.
These tests can be run either from the technicianâ€™s laptop if connected to the target network or from a network-connected PC at the clientâ€™s location.
Q: How can existing CCTV systems be connected to LANs and the Internet? A: With the large quantity of CCTV systems in existence, the ability to connect previously installed CCTV cameras and systems to networks and the Internet is a great opportunity for security companies to upgrade their clientsâ€™ systems. Security equipment manufacturers are responding to this marketing opportunity by providing a variety of ways to network existing CCTV equipment.
One way to network a current clientâ€™s CCTV system is the â€œforklift upgrade,â€ where the entire old system is replaced with new IP-addressable cameras. While attractive from the salespersonâ€™s point of view, many clients wonâ€™t buy this approach due to cost.
If a client has a substantial investment in their existing CCTV system, then an incremental conversion to IP-addressed equipment will best suit their needs.
There are a number of ways to approach this incremental process. For smaller systems, say 16 cameras or fewer, the simplest and best approach is most likely the replacement of their existing recording equipment with a network-enabled DVR. This is a common method, as the DVR is a â€œdrop-inâ€ replacement for an existing VHS-based recorder.
If the client is content with his or her existing system, but would like to connect one or more cameras to their network and/or Internet, dual-function cameras are available from vendors such as Panasonic, Sanyo, and others. These cameras provide both a standard analog output and an Ethernet connection, while retaining the traditional form factor of a CCTV camera. These devices allow the dealer to replace only specific cameras, such as at an entry door or loading dock. Once installed, the analog output of the dual-function cameras can be connected to the existing recorders and monitors, so the local system retains its previous functionality. The new cameraâ€™s network connection will provide video transmission to authorized users on the LAN and/or Internet, simultaneously with the analog video. Because these dual-function cameras can be purchased in typical CCTV forms, including pan-tilt-zoom domes, they can be used anywhere a standard CCTV camera can be placed, and utilize a variety of lens types.
A less expensive option than replacing cameras is to use video servers, which take the video output of a camera, monitor or other CCTV device and produce a network video stream. These devices are very attractive, as they lower the labor cost of converting an analog video signal into a network stream. Technicians do not have to pull down an old camera, replace it, adjust the lens, etc., when using video servers. Installers can connect a network video server such as the Napco VIP-Gateway to the cable coming from the existing camera, and use the video out connection on the VIP-Gateway to feed the video signal to the existing local recording device and viewing monitor.
Letâ€™s look at two examples of latency testing. Using the ping command, the latency of communication from one computer to another on the same LAN is tested:
Looking at the â€œtimeâ€ column, we see that communications between the two computers are accomplished in less than 1 millisecond. If this test indicated that communications were completed in more than 5 milliseconds over a LAN, there may be a problem with the cabling, connectors, or network devices, which is slowing network communications.
Here the time for packets to travel is between 164 and 162 milliseconds between Chicago and Helsinki. Because the data is traveling through the Internet and across the Atlantic, it takes more time for the packets to reach their destination.
The amount of latency in a particular point-to-point network connection will affect network video transmission in two ways. The first is that the amount of latency â€“ added to the time it takes for the video signals to be compressed and decompressed â€“ will cause a viewing time lag between what is viewed on the monitoring computer and what is actually happening right now at the camera/video end. When viewing cameras or DVR video over the Internet, this time lag could amount to a few seconds. Typically, this is not a major concern, but this aspect of viewing network video should be pointed out to the end user.
If excessive latency is present between two network points, problems can arise when attempting network video transmissions using certain compression codecs. IP cameras and network-enabled video devices using JPEG or a variant will transmit in TCP, or Transport Control Protocol. TCP is a â€œconnectedâ€ network communication, and the transmitting device is expecting to receive acknowledgement or ACK receipt packets indicating that what is being sent is being received properly, and within an acceptable time frame. If there is excessive latency of more than 250 milliseconds in a network connection, JPEG-transmitting cameras and video encoders may provide erratic viewing or none at all, because the receipts are not received promptly enough.
MPEG-2 and -4 video devices will send their packets in UDP or User Datagram Protocol. UDP is a â€œconnectionlessâ€ communication, with the transmitting device not asking for or expecting receipts. Excessive latency will have little effect on MPEG transmissions, although the lag time between the scene at the connected camera now and whatâ€™s being remotely viewed may increase.
Latency is an important issue, and is easy to test. Using a laptop computer, a technician can address it onto the local network, plug it into the switch or hub port to which the camera or encoder will be connected, and use the ping command to ensure that LAN communication latency is acceptable. If planning to provide Internet viewing to a remote computer, the technician can use the same ping test to check the latency between the cameraâ€™s proposed LAN connection and the remote viewer, provided that the technician has the public or Internet IP address of the remote viewing computer.