Everybody's an expert
While hardware to support VoIP is certainly widely available, VoIP is about far more than just equipment.
Unless you've got extensive inhouse expertise in the area--unlikely in all but the largest companies--you're going to be relying heavily on the good advice of your integrator and vendor's implementation team.
Unfortunately, that's not always a safe bet. One by-product of the early hype over IP voice has been an explosion in the number of companies purporting to offer VoIP expertise.
Recognising the technology's natural appeal to customers, those companies have jumped on the bandwagon assuming their data skills are adequate.
"VoIP certainly seems to be becoming a mature opportunity in the Australian market," says Gary Spooner, managing director of Centile Australia/New Zealand, which opened up shop here late last year to spruik its VoIP software to local interests.
Centile's Hosted iPBX solution is aimed at service providers who would like to offer virtual PABXes to their customers, but can also be run inhouse within businesses eager to get into VoIP.
It requires Sun Microsystems Solaris-based Netra servers and Cisco Systems network switches to run, so the company is counting on its customers' already having the necessary infrastructure in place. In a cluster of five Netra t1s, Centile claims the system supports up to 100 virtual PABXes and 10,000 extensions.
Sun servers aren't cheap however, and it will take a good business case for all but the largest companies to justify such an expenditure. And that's only the beginning: without adequate bandwidth planning and commissioning, quality of service issues can easily persist despite having a technically correct software implementation.
Major systems integrators probably have access to the right kinds of people to avoid such errors, but customers need to retain a healthy scepticism to make sure that smaller companies aren't just talking the talk. It's all too common for suppliers to underestimate the complexity of VoIP implementations and assume that the technology will take care of itself.
For a data expert to assume that VoIP is just another form of data is a woefully common but potentially damaging mistake, since telephony issues such as signalling and feature management don't go away just because you're using VoIP.
"Data people tend to see VoIP as just 64Kbps data, but in particular voice signalling is not well understood," says Panaseer's Hampel. "This is reflected in naïve implementations that don't actually deliver results. Sometimes companies get into a situation where they've just thrown that VoIP data in with their general data, and neglected issues about how to focus dedicated capacity on that delivery."
Hampel estimates that effectively deploying VoIP as a replacement for conventional trunk lines requires a significant investment in bandwidth-four times the actual requirement, in his estimation-to ensure there's enough available bandwidth to support real-world traffic volumes.
"Often there are substantial hidden costs in terms of data upgrades that people don't fully take into account," he says. "Our company is doing a fair amount of business developing and implementing voice systems for people who have previously invested in the equipment to do data and VoIP believing they'd be able to do it all. They now realise they're not going to be able to get the result they expected."
Although he recommends VoIP in niche applications like call centres, commissioning so much excess capacity can make the network necessary to support VoIP prohibitively expensive.
For companies that can't justify the cost of VoIP's other benefits, Hampel says, a better solution may be conventional time division multiplexing (TDM) solutions using 8Kbps voice channels over less expensive leased lines.
Poorly architected VoIP implementations can also introduce problems should features not work like they did on the old networks. Incorrect implementation of VoIP call handling procedures, for example, may mean a call cannot be transferred when it's delivered to a handset.
One early VoIP adopter found that while call forwarding worked fine, calls would spin into an endless loop if the number to which the call was sent had also been set up to forward.
Users, of course, hate this--and they're the ones you want to inconvenience as little as possible. Without proper guidance, your VoIP project could easily become the bane of the company.
Make sure your VoIP implementation discussions include an assessment of the pedigree of potential service providers, who should be able to competently address issues including integration with existing systems; quality of service parameters; future scalability; power, server and network redundancy; VoIP administration and management strategies; and security procedures and software.
This is a very interesting site well done.