Integrators often cite bandwidth as a key concern for deploying IP cameras. Let's examine what the potential issues are and where they may be applicable:
- You do not have enough bandwidth for IP video (see our bandwidth tutorial if you want a review of the basics)
Why might you not have enough?
- Existing local network is old or antiquated
- Your local network is fairly new but already has too much traffic
- You are trying to send IP video over a wide area network (DSL, Cable Modem, T1, E1, DS3, etc.)
What might block you from solving this problem?
- It is too expensive to upgrade your internal network
- It is not possible or extremely expensive to upgrade the wide area network
- It is costly/complex to implement traffic engineering
We have 3 cases to examine:
Upgrade the Internal Network
You have dozens of IP cameras in a building/facility but the existing network does not have enough bandwidth to support IP cameras. The simplest solution is to simply deploy a secondary network to support the IP cameras. This is almost invariably cheaper than deploying analog cameras unless the cameras are very close (under 100 meters) to the server/headend. Slightly more complicated but perhaps cheaper is upgrading the existing network (say from 100Mb/s to 1000Mb/s).
Upgrade the Wide Area Network
You have IP cameras in a building or campus and want to stream those cameras to a central location for storage. This I think is just
not feasible. Technically it can be done but financially it is rarely justifiable. In most countries, the cost of 10MB/s WAN bandwidth is very expensive - in the range of $3,000 to $30,000 USD per month. At that price point, it is simply cheaper to store video from IP cameras at each local site.
This is only a problem for IP video to the extent that if it was solved it would benefit IP over analog. However, it is certainly not a shortcoming of IP.
Implement Traffic Engineering
You have a converged network where you are sending corporate data, voice over IP phone calls and IP video surveillance. To ensure that IP video does not ruin phone calls or disrupt important data transfers, you want to implement advanced network capabilities to ensure all services work without interfering with each other. This can certainly be expensive both in deploying equipment with these capabilities and dedicated trained engineers to configure/optimize them (see
Cisco's webinar on this). If you are deploying a large IP video system with hundreds of cameras across long distances, this is probably still cheaper than building out an analog CCTV system. If the number of cameras are smaller and the distances are shorter, you may solve this by deploying a secondary network simply dedicated to IP video surveillance.
Discussion
This should be an interesting discussion. I just wanted to open with some general observations. There are lots of other aspects to consider. Please expand, disagree and debate this.
