VoIP telephony is a wonderful example of the chasm that appears between home experience of the cloud and the reality of company life. Consumers see VoIP as Skype: after all, it's voice, right, and the conversation travels over IP, doesn't it?
This doesn’t square with the reality of the astronomical cost of enterprise telephony and the nightmare of implementing a VoIP system. Why is it so hard to make VoIP really work?
It's partly because the same term has different meanings. "VoIP" is not a term protected by standards, or by a brand. Within VoIP there are more than ten distinct protocols, routing methods and support services. Simple endpoint clients presented by a major league telephony player like Mitel or Alcatel, are actually not using any of the same network structures to support the waving kids you can see on your laptop. Technically they are probably using H.323 standard IP traffic; public internet VoIP traffic uses SIP and a portfolio of other protocols.
This is why VoIP networks are often being deployed in what is euphemistically referred to as an "unconverged" state - that is, each desk has two wires leading to it. One for the computer, from the company LAN; and another for the phone, which in many deployments, has no physical connection to the main LAN, whatsoever. Sometimes this is because the phone budget and the LAN budget are owned by two antagonistic empires, but it can also be done because network tweaks could break the reliability level that people are used to from their phones.
This does not look like Skype to an end-user, nor indeed does it look like the free MSN Messenger (which is easy to accommodate in a security plan). Microsoft Lync is appearing on people’s radar not just because it is being offered free (assuming you have the rest of the software in place to support it), but because it doesn’t rely on VoIP to transport internal communications.
This basic architecture decision ie that the nasty bits of SIP-based VoIP should be hidden in a cloud suite at a telecommunications firm – is the basic premise of managed PBX. Everything that used to be done by a grey tin box in a cupboard is now up in a hosting centre somewhere. All your phone devices – be they traditional phones, videoconferencing devices or software on your work PC – make a simple point-to-point link to that hosting centre.
Make no mistake about the intensity of competition here. Traditionally, companies would spend more on their internal phone system during a replacement project than they would have done on their LAN. Phone system prices were kept high, thanks to the slow pace of innovation, proprietary technologies and the absolute need for telephony available to all, on every desk, come fire, flood, asteroid or ice age.
The business sector which lived off those kinds of deals has not taken well to the more price-competitive world of IT muscling in on their revenue stream. This is a large part of the justification for both separate VoIP LANs (when proposed by traditional telephony salesmen) and all that stuff about "convergence" (when sold by LAN hardware architects). Voice is just data, and the bandwidth should fit into a tiny corner of a network – except, very often, it simply doesn’t.
VoIP is treated with horror by telcos: it’s a direct threat to their core business. I’ve done VoIP-managed PBX rollouts in companies of fewer than ten people where the savings are enough, per month, to hire another person - that’s a lot of money.
It should come as no surprise, then, given that telcos are still in charge of the giant public data networks that VoIP traffic is frequently seen as no more as the 21st century equivalent of the miner’s canary eg, a sign of excessively draconian traffic-shaping by a telco pretending to be a full-service ISP, while in reality being nothing of the sort. Rarefied pundit articles have been alluding to a two-tier web for some time, as a distant prospect, but the truth of the matter is that when you start to look at VoIP as revenue threats to old dinosaurs, then accidental cutting out of particular traffic types follows.
Not that the signs are as blatantly evident. Telephony has only two error messages – three, if you count total silence. VoIP is no different: Where hardened old TCP/IP hacks have traceroute and Syslog, VoIP very often simply fails to complete or dries up altogether: The standard “polite email to the support team” approach changes when telephony enters the picture. It’s not a big problem when opening a Word document brings up the hourglass, but if a phone call drops while talking to a client, emotions rule and the hunt for the guilty party takes on a new intensity.
This is why you will be hearing so much about Managed PBX services. Businesses struggle to build the kind of connection portfolio demanded by VoIP, yet the prospect of thousands of pounds a month cannot be ignored. The siren call of all those keen advertisers, promising to take all the headaches away, is dreadfully appealing, but there are some basic questions to ask.
- Are you entirely dependent on phones for your business? Given that repair times for Internet links can lead to a day or two of downtime per quarter, could you survive that kind of interval?
- Is permanent 24/7 call-ability a prerequisite, with staff reachable at home or wherever they might be? This used to be important but the slow shift to email and SMS has relieved the pressure. If it is important, though, it can drive adoption of hosted VoIP.
- Secure communication via voice is taken for granted, but VoIP is composed of network packets and can be sniffed like every other type of network traffic. If security is a concern then traditional telephony is definitely still the right choice.
- Multi-party telephone conferencing with lots of the parties inside the company LAN can drag an internet link to its knees in a managed PBX environment.
- If your provider started to under-perform, could you tolerate the inevitable bumps in continuity of service which will come about as part of moving from one provider to another?
Despite these caveats, managed telephony is a real cloud application; it’s set up so you can choose a supplier at the moment you come to use it, rather than at the moment you sign a binding and disadvantageous contract. However, this is true mostly for outbound traffic, not inbound. If you are depending on a reliable inbound stream of revenue-generating phone calls, VoIP has some way to go before it is worthy of your undying trust.
