Ever since the personal computer revolution of the early 1980s, industries of all kinds have been looking for the “killer app” — that one application or use of technology that makes it all worthwhile, that gives something real value in a marketplace. A killer app legitimizes a technology and makes people sit up and take notice. 

When we talk about the convergence between the traditional security industry and IT, what do we see for our killer app? Many new uses for the merging of analog and digital have come into being in all areas of physical security. But is there one that stands out among the rest? 

While it may be arguable, I would venture to say that networked video might just be that killer app. It has become the rallying cry for our industry to hop on the network. It appears to provide solutions that have been seemingly impossible or unduly expensive until now. 

Distributing video across the network may be what has led you to this point. It has become the application that brings more and more security integrators into this field. Let’s look at some of the things you need to know to be successful in the convergence wave. 

Video’s Impact on the Network
One of the big attractions to digital networked video is the ability to send it across an enterprise (the business LAN/WAN). Distributed video systems are already beginning to replace DVR-based systems for many large or midsized installations. A distributed video system is one that places video encoders or IP cameras in remote locations, connected to any nearby network drop. A DVR-based system requires cameras to be home-run to a central box, then sent onto the network. 

This distributed solution has many benefits, not the least of which is labor and saving time, but it does add a great load to the network. Video takes up a large amount of network bandwidth. The file size of each video stream can be up to or exceed 4-5Mbps (megabits per second). For one or two cameras, it’s not a big deal, but what if you have hundreds? This extra utilization is what generally causes friction between the integrator and a client’s IT staff. 

There are several methods of sending data across the network: unicast, multicast, and broadcast. Two of these are appropriate for video; one is the enemy. 

Broadcasting Presents Problems
Let’s get this out of the way right up front: Broadcasting video data on a network is bad news. Broadcast data is data that is sent to every receiver on a network. For a network switch, that means that the data is sent to every port. For small pieces of data, like a computer announcing its address to the network, that’s not a big deal. But if you were to try and do that with a 5Mbps video stream it would be another story. 

Most routers on a network will stop broadcast traffic, as will switches that have been configured with VLANs (virtual LANs). I guarantee, however, that if your video causes a broadcast storm on a switch, you’ll be hearing from the IT guy. 

The most common way video makes it across the network is via unicast transmissions. Unicast is basically what it sounds like, a single (uni) transmission. Unicast data is sent between one sender and one receiver. It generally requires no special network configuration. If you have a DVR and one remote client, or a single IP camera, chances are you’re viewing a unicast transmission. 

Unicast is sent via TCP (transmission control protocol). This protocol sets up a direct link between two devices, and uses several methods of guaranteeing delivery of the data. Guaranteed delivery of data is a good thing when you’re dealing with most network documents, like E-mail. In fact it seems like a good idea for video, eliminating dropped frames, and sometimes, when used in moderation, it is. 

Unicast video is usually used if there are few receiving devices. An IP camera or encoder, for instance, will generally use unicast to send its stream to an NVR (network video recorder), ensuring delivery of every frame. If that NVR or a DVR can only support one remote client, you will probably also see unicast transmissions between those devices as well. 

What happens, however, if you have multiple receivers looking at the same stream of video? With unicast, the transmitting device would have to generate multiple video streams, a new stream for each viewing station. This would quickly cause overwhelming amounts of network traffic, stress for the IT staff and headaches for you. This is where multicast comes in. 

Multicasting video is one of the relatively newer technologies on the network. It is one of the technologies that really make distributed network video possible.

If unicast transmissions can be considered one-to-one, then multicast transmissions would be one-to-many. Multicast video transmission eliminates the need for multiple streams traveling around the entire network if a video source needs to be used in multiple ways. 

Multicast video takes advantage of network switching and routing hardware on the network that can speak its language. This is where the first complication comes in. The network infrastructure needs to speak multicast or several bad things happen, not the least of which is broadcast storms, or data simply being blocked completely. 

You need to be aware of a few terms when you are speaking to the IT staff about putting multicast video onto their network. The first is IGMP (Internet group management protocol). It is the foundation of multicast on the network.

IGMP Facilitates Multicasting
Multicast data is sent around the network through a series of groups. Individual receivers (viewing stations or PC clients) join these groups to let the network know that they want a certain stream. IGMP is the protocol that governs the joining and leaving of these groups. 

When a viewing device wants access to a stream, it sends a message to the switch it’s connected to that says just that, “I want to join the stream at address 224.1.32.10” (the IP address is just an example, but there is a special range of IP addresses set aside just for multicast data). That “join” message is sent up the network all the way to the switch that is hosting the requested stream. 

Each switch that received the message forwards the video stream back down the same path that the “join” came through. This way, the video stream only goes where it was requested; it doesn’t need to flood the network unnecessarily. When the receiver is done with the stream, it sends a “leave” message back the same way, and the stream is dropped. 

Routing Protocols Mind Subnets
A few more important items to remember: Routing protocols keep track of all the different subnets on the network, so data can be sent to the right place. They are what make a large network possible. Multicast data also needs to be routed between subnets, but requires some special handling. 

One of the most common multicast routing protocols you will see is PIM (protocol independent multicast). It is compatible with most switches and routing hardware out there. As the name indicates, it will work no matter what standard routing protocol is being used on the network. You will run into this one more often than not. 

PIM keeps track of the layout of the network, as well as the locations of multicast groups. It makes sure that the multicast data is sent across the network via the shortest paths possible, again, to keep the multicast data from flooding everywhere.

Introducing the ‘Killer App’:Video

Video on the network has shown itself to be the killer app of the convergence wave. It has pushed the security industry
onto the network faster than any other technology use. Taking the time to understand how video impacts the network will help you communicate better with the IT personnel you run into, and hopefully ease some of the headaches of a converged installation.

If you enjoyed this article and want to receive more valuable industry content like this, click here to sign up for our FREE digital newsletters!

Related Content