Critical business networks

What’s so good about IP centrex?

It’s been one of the most hyped up services in recent years but the arguments for IP centrex are far from clear-cut

written by Stewart Baines, Futurity Media

Not a day goes by without another IP centrex announcement. You would be forgiven for thinking that the PBX had finally given up the ghost and carriers have at last found a way to migrate users to IP telephony. But how much truth is there behind the PR façade? Is IP centrex anything more than a revamped version of its old PSTN namesake?

		Table 1

Proclamations of IP centrex service launches have come from an array of alternative carriers (Table 1) and there is a wide range of speculation, most of it positive, from analyst firms about how big the IP centrex market can be. Ovum estimates that IP centrex and hosted VoIP revenues will amount to around US$6.6 bn worldwide by 2008, up from US$255 m this year. Frost & Sullivan puts a US$7.9 bn valuation on the market by 2009.

Financial estimates, when rounded up to such a level, may be of little appeal to carriers preparing a business plan. BT’s projections, however, may prove more insightful. Current centrex services for the UK incumbent account for around eight per cent of seven million business lines. It expects, conservatively, that IP centrex will account for at least 15 per cent of all business lines by the turn of the decade, double the size of the current centrex market.

And will it be at the expense of old-school centrex services? Surely selling to a converted audience will be harvesting low-hanging fruit. But in the US, the old centrex market is in decline. Eastern Management Group calculates that centrex lines have dropped from 18.5 million in the 2001 to 13.4 million in 2004, without a substantial number of them turning to IP centrex, instead opting to return to CPE-based systems. Yet IP centrex services have been available in the US from 2001/2002.

Helping VoIP gain a foothold

So why should IP centrex be such a success? For one, VoIP has needed a leg up, at least in the enterprise arena. While the number of PBX ports shipped this year — roughly half — are IP-enabled, it would be a gross exaggeration to suggest that they are being used as IP rather than the PBX’s native digital protocols.

“I’m really convinced VoIP will take off when users do the next round of technology refresh,” says John Blake, head of hosted IP telephony at BT Global Services. “There’s enough credible evidence that the voice quality is good enough. So when the big boys edge there way into VoIP, a little at a time through their organisations, then the medium-sized companies will follow. And the small business will want the same toys the big boys have. But I don’t presume it will necessarily be easy to churn them from old technology to VoIP and IP centrex. It will need to be cheaper and more powerful than what they already have.”

Why IP centrex makes sense is that it is relatively risk free — assuming the service provider can provide adequate service — and it’s cheaper. Firstly, there’s no capital expenditure and even operational expenditure (despite paying per-seat licence rentals) can be cheaper than a CPE equivalent service when PBX maintenance contracts are taken into account. The key advantage it has over traditional moves, adds and changes on a PBX is that a US$100 an hour maintenance engineer is not required to set up a new extension. IP centrex services typically offer a simple, web-based interface for ordering, fault management, administration and doing those all-important (and expensive) moves, adds and changes.

Who’s interested?

There is much debate about which customer-type will be most interested in outsourcing outsource VoIP and those that won’t. “Its still hard to accurately predict which segment is going to be most attracted to IP centrex,” says Peter Hall, a research director at analyst firm Ovum. “The early market will probably hit larger companies who are, after all, more familiar, with outsourcing.”

Large companies can be split into two different scenarios: distributed organisations with a number of branch locations where the cost of running PBXs at small locations and maintaining staff to support them would be considerably more than paying a carrier do it all centrally; and large office sites where staff frequently move around desks or from home, and are often in city centre locations where communications rooms take up expensive floorspace. In both cases, the larger the organisation, the more experience it has with negotiation outsourcing contracts and reading between the lines of an SLA.

At the opposite end of the spectrum, there are small- and medium-sized businesses. Bearing in mind their comparative lack of telecoms experience (to buy and maintain a PBX), uncertain futures (will they grow or contract) and being extremely cost conscious (opex preferable to capex), you would may think that many would want to outsource? Market watchers AMI Partners estimates that IP telephony spending among SMEs worldwide will amount to US$1.2 bn by the end of this year. However, a mighty US$1.1 bn of that is accounted for in IP PBX and IP phones. Just over US$100 m is accounted for in VoIP services. So much for the theory.

“I think a lot of small businesses will be quite reluctant to look at IP centrex,” admits Steve Kennedy, director of technology futures at Thus, a UK alternative operator. “A 25-person company with one small PBX and a primary rate ISDN line may find it extremely expensive to rip out something that’s already working and move to IP. They’ll need to completely upgrade their LAN for instance to support IP telephony to the desktop.”

He continues: “If a company is to move to IP centrex it won’t be because they’re getting more features. What they’ll want and get is the same features they have with their PBX but a lot cheaper. If you don’t get cost savings, what’s the point?”

Thus has not announced an IP centrex service, despite being one of the more IP-centric operators in the UK, while rivals BT and Cable & Wireless have been positively gushing about IP centrex’s potential. Thus is expected to announce a hosted IP telephony platform rather than edging customers down the centrex route. “I think when we launch we’ll be looking at a hybrid approach that supports the continued use of some existing infrastructure. The whole point of VoIP is that it’s supposed to be cheaper,” says Kennedy.

Barriers to adoption

The most obvious cause for the slow uptake of IP centrex is the continued presence of the PBX. With a ten-year plus working life, many have been upgraded, renovated or simply patched up. Extra NICs, more ports, interfaces to voice mail servers or media gateways, and new operating software, means than most PBXs are not the same technology of a decade ago. Which is good, because ten years ago deregulation was on ly just kicking in and ISDN was considered the next big thing. Unless an alternative is cheaper or offers considerable more functionality (presuming its gizmos that can actually help business productivity), then many end-users will decide to simply stay put with something that works.

Another barrier to IP centrex take-up is the telecom managers themselves. Few will want to see themselves out of a job. “I think that’s a real barrier,” explains Carine Ziol of French IP centrex platform developer NetCentrex, which includes France Telecom as one of its IP centrex customers. “They will tell their bosses that IP centrex does not make sense so I think carriers would get more success selling to smaller businesses without telecom managers.”

Pricing has certainly been an issue in most nascent services. Old centrex services were seen by many of its users as more expensive than owning and managing their own internal telephony infrastructure. IP centrex, will of course be considerably cheaper than that, but how much? As all call control, adds, moves, changes and even line provisioning can be largely automated via web portals, there’s little overhead required in servicing a customer. So should IP centrex be pitched as a small premium to PSTN/ISDN line rental? Or should the fact that a user is saving money on PBX/handset ownership and management mean that IP centrex should be sold at a significant premium to simple line rental but slightly cheaper than the DIY option.

Fear of lock-in

The Enterprise VPN Users Association (EVUA) should be well versed in VoIP by now, partly because it’s one of the two main drivers of migration from frame and ATM to MPLS/IP VPN. Surely they would want to outsource their voice with their data? “I don’t think many of our members are too interested in IP centrex, despite them being more open to outsourcing than most,” explains Phil Barton of the EVUA. “They’re just too worried about lock-in. I think that’s why the original centrex services weren’t too successful”. Barton believes that once you’ve removed a capital cost from the balance sheet, most finance directors would be adverse to re-introducing it. Put simply, if you dump your PBX for an outsourced service, don’t ever ask for another one if it all goes wrong with the outsourcers.

And this is a considerable lock-in. A lack of centrex providers in the past meant that many companies were tied into one provider’s offering. And that can mean expensive adds and changes, and no guarantee of getting the most competitive call tariffs. An overly expensive usage bill could obviate any savings on equipment not required.

“If the customer is unhappy with what they’re getting from the service provider, they’re not in a great bargaining position,” argues Barton. “If the operator owns the wiring and the handsets it can be very tricky for business users. If you do go into a centrex contract you better make sure you at least own your own wiring or the assets are transferred to you at the end of the contract. And avoid long contracts.”

The are many sceptics who think the total addressable market for IP centrex will be the sum of all the existing centrex extensions. Is that enough? For some, maybe. But that’s hardly the leg-up VoIP needs. Looks like it’s back to the drawing board.