Fixed or wireless communications?

Optelecom-NKF's Coen Hooghiemstra considers the play-offs and pay-offs involved when deciding whether to go for fixed or wireless communications solutions

Cost of ownership is a major driver of acquisition and deployment decisions - and never more so than, as now, during a period of economic difficulty. Although the use of wireless communications has increased significantly across a wide range of business sectors in recent years, driven by increases in capability as well as perceptions of the ease of deployment and cost, 2229 Optelecom-NKF's vice president of engineering Coen Hooghiemstra cautions against making uninformed assumptions. There are, he says, a whole series of technical and economic factors at play which make a choice between wireless and fixed networks less than straightforward.

"There are always going to be locations, particularly in cities, where installation or replacement of a fixed network is prohibitively expensive; Chicago and New York are examples of where wireless has been implemented widely as a result.

"In simple terms, where it's expensive to dig you're going to see a lot more wireless," he says. "But you also have to look at the applications which need to be provided with data services. In the ITS sector, there are numerous technologies which can be used for detection and monitoring. The ubiquitous loop remains as valid as it ever did for some applications but images are very useful for flow control and absolutely essential for incident management.

"That creates a need for bandwidth and gives rise to deployment decisions. With fibre optic, bandwidth is no issue. With wireless, it is - we see many situations where wireless applications are limited to less than 100Mb. Even where physical infrastructure already exists there are constraints. With copper wire for example distance can be an issue, although you can use signal amplifiers." Bandwidth requirements will only increase in the future, he continues: "More efficient compression techniques mean that you need less for the same quality. Codec standards such as H.264 help but the introduction of High-Definition [HD] cameras only accelerates the trend; HD cameras need more bandwidth than their standard-definition predecessors, for instance, and the numbers of cameras in use go on increasing.

"A readily acknowledged solution is to find ways of not transporting all images at full frame rates. Intelligence is being pushed to the front end, with in-camera processing becoming more common. Storage is becoming cheaper so there is a tendency to store more, which creates a bandwidth burden, but you can limit frame rates and resolutions or only record significant events. Certainly in Europe, most highway cameras don't store all the time. That makes traffic management different to some other CCTV applications, where everything is retained."

Reliability

Although wireless technology's reliability continues to improve, it is still dependent on several different factors.

Hooghiemstra: "The technology used, distances involved and frequencies available all have an effect. Antenna system and power are further considerations, as are jamming and interference. Malicious acts are more common in the security field but they can't be ruled out in ITS. The long and the short of it is that there is always a risk of disturbance between two wireless antennas.

"Modern technology helps but latency remains an issue - there will be a need to resend data packages when interference occurs. Fixed networks don't suffer from such problems.

"These issues are solvable but the solutions can be costly. You can, for example, have intelligence reside in all of the cameras on your network and reduce transmission distances but the fact remains that our telecommunications frequency bands are becoming increasingly crowded and in some countries professional bands are reserved for military and emergency service use. Another solution, often favoured by police forces for mobile sites, is to connect 10 to 20 cameras at a location using cable and then wirelessly transmit the data generated."

SLAs and the current market

In many IT projects, availability is now incorporated into specifications in the form of a Service Level Agreement (SLA). Hooghiemstra sees SLAs becoming increasingly common.

On the one hand, he says, they offer opportunities for suppliers to improve their profitability by providing for users' maintenance needs. Availability, where it is a specified item, also dictates that certain strategies be implemented, such as multiple routing and other means of signal redundancy.

On the other, the rise of the SLA increases the pressures on suppliers: "Money is very much an issue in the current market. Customers are looking for ever-increasing levels of quality and functionality for a given price. You see this often; suppliers have to meet a specified availability rate then make a very sharp offer. How sharp depends on how close you are to the project and how much of the customer specification you can already address. It's certainly increasing competition, with more technology and additional features being added all the time." Again, this has both positive and negative effects. It helps to advance the state of the technological art. But it squeezes both offered solutions and their suppliers, who have to contend with reduced margins and at the same time anticipate where and how to innovate. Whether it is a sustainable situation in the longer term remains to be seen.

Differentiation is key and one trend is to apply video analytics - Automated Incident Detection (AID) by any other name.

"That affects the cost equation," Hooghiemstra continues. "With the numbers of cameras increasing all the time, the ability to reduce the numbers of operators has a major effect on operating costs. The installation might not be cheaper but the running costs are significantly reduced. The greater use of AID is just one factor that's supporting a current trend: the amalgamation of smaller Traffic Management Centres [TMCs] into bigger, regional ones.

"That also reflects a continuing need for dedicated networks. Certainly in the Netherlands where TMCs have been combined they have retained dedicated communications links. Wireless just doesn't offer the necessary reliability. GSM (cellular) networks tend to become overloaded during emergencies or times of major disaster - just when you don't want them to - as everyone is trying all at once to contact family and friends. PSTN links offer reliability but suffer from being prohibitively expensive to lease." Where using PSTNs (public networks) does make more sense is at the strategic, country-to-country level.

"There, it makes a lot more sense to use guaranteed connections provided by the telcos, rather than dedicated networks, because the dynamics are different. The bandwidth needs are only moderate because in the main only data is going to be passed, not large numbers of high-quality images."

Looking forward

Any discussion of the future data needs of cooperative infrastructures (vehicle-to-infrastructure and vehicle-to-vehicle communications) has to consider the broadband requirements. According to Hooghiemstra, however, there is a rather more pressing need: to bottom out the business cases for deployment and operation.

"There is certainly a lot of work going on at the technical level to develop the infrastructure and application elements but where will the impetus for deployment come from?" he asks. "The drive will certainly come from the individual user but we're nowhere near that situation yet." Similarly, there is potential in future-generation cellular (4G and follow-on) technology as a possible means of communication but only in strictly defined circumstances.

"If you need an extremely fast install, cameras up and ready inside a few hours or a couple of days, wireless and cellular might be useful first steps. But they're no solution in the longer term.

"It's difficult to give absolute figures for future bandwidth needs and availability. The only safe prediction is that the need will increase and the hunt will continue. That's a universal truth inside and outside ITS.

"We're already chasing a scarce commodity, though. There are potential solutions. The military in many countries is currently laying claim to a lot of frequency space which it doesn't use. Perhaps some of that can be freed up for more general civilian use. Whichever way, it will become more expensive to use dedicated frequency bands - one need only look at the huge sums raised in recent years by national governments which have held public auctions.

"Increasing intelligence at the front end and ongoing developments in compression techniques will help - or you could just use cable. In the end it's not necessarily more expensive, especially at the roadside where road operators already own and have access to the real estate."