The game of technology-tag has been an interesting one. Analogue cameras have lasted far longer than many industry pundits predicted. It’s not as if digital cameras haven’t been around a very long time – patent records show they were invented by Australia’s own Zone Technology back in the 1980s. But the momentum of research and development in the first half of the noughties gave analog cameras a second lease on life thanks to major advances that include but aren’t limited to the Pixim chip, as well as successive generations of superb cameras from leading manufacturers. The delay hasn’t just been about IP camera performance. Taking the market segment as a whole it’s quite clear that much of the problem from an integration perspective lies with the networks themselves. It’s not just bandwidth. Fact is the router-to-router network techniques that continueto support LANs and WANs have created a flawed substrate across which it’sextremely difficult to drive real time video streams without performance andreliability issues. It’s probably no more that slightly unfair to claim that poor network design and the limitations of current networking technology have hamstrung WAN-based video surveillance systems. And despite manufacturers waxing lyrical about the performance of IP cameras/encoders that work as edge devices on existing networks, even the best and most capable of these solutions is all too often grafted to a comms solution that’s about as conducive to real time movement as gumboots at a beach sprint. Most case studies undertaken by Security Electronics & Networks Magazine over the past couple of years show that end users who are serious about multiple input CCTV solutions that offer acceptable image quality and frame rate get around feeble data networks in one or another of 3 ways. They might rely on dedicated carrier grade networks that they beg or borrow. They sometimes install dedicated point-to-point links that they port to routers/switches in an IT rack at the control room to deliver what is then hopefully described as ‘networked video’. And they most often do all the serious surveillancework at their remote sites using local DVR/NVR/RAID combos and download image streams one input at a time across private WANs (with reduced frame rate and resolution) when something happens. But whatever else they do, these end users never send multiple, simultaneous, live, 4CIF video feeds across wide area data networks to centralized locations for monitoring and storage. Why is this a problem, you ask? Well, it’s not really a problem at all. There’s no particular reason video surveillance systems shouldn’t run on dedicated data networks instead of on dedicated analogue networks. No reason except it entirely misses the point. I don’t want to criticize any of these ‘networked’ solutions too much – they all represent a best-effort attempt by installers and integrators to give end users the sort of performance they demand for the sort of dollars they’re prepared to spend. Having said that, all these solutions represent an attempt to burrow under one salient truth – most networks are complete and utter crap. Security integrators have zero chance of offering customers solutions that support multiple live video feeds until IT departments wake up to the fact that skinny 512Kbps asynchronous private networks at 150 bucks a month are a seriously bad deal. The answer is 2 or 10Mb WAN links employing multi protocol label switching (MPLS), a method of building networks that allows improved packet exchange and exploitation of network latency to increase transmission speeds and provide assured signal integrity. There’s no real chance widespread MPLS WANs will ever be driven by CCTV applications though – the only realistic catalyst from a dollar-saving perspective is VoIP. Given the trudging take-up of VoIP in the Australian market, all those wholesalers plugging away with analogue cameras and standalone DVRs can rest assured they’ll be in business for many years to come. John Adams, editor, Security Electronics & Networks Magazine
