Optical/IP Networks

IP Centrex: Build or Buy?

Competitive hosted VOIP service providers are questioning the suitability of commercial IP Centrex platforms for carriers, following Level 3 Communications Inc.'s (Nasdaq: LVLT) decision to exit the hosted VOIP market (see Level 3 Provides VOIP Lesson).

Two of the upstarts hoping to capitalize on Level 3's decision say they're using their own in-house platforms, and not systems sourced from vendors.

Jason Talley, CEO at Nuvio Corp., says his company is in talks with a number of current 3(Tone) channel partners looking for a new hosted VOIP wholesaler (see Nuvio Treads in Level 3's Tracks).

"We developed our own services platform, and that pleases the channel partners. It makes them confident that they won't suffer the outage issue that Level 3 had," says Talley. (Level 3 hasn't admitted to any outage problems -- see Level 3 Provides VOIP Lesson).

"We checked out the commercial IP Centrex suppliers, but there weren't any advantages to switching off our own system. Their license costs are high, and every time you want to change something you need to call out someone from the supplier to code a new feature. The usual suspects didn't have anything going for them."

Those "usual suspects" include BroadSoft Inc., Sylantro Systems Corp., and VocalData Inc., which is now part of the Tekelec Inc. (Nasdaq: TKLC) empire (see Tekelec Connects With VocalData).

Talley's view is largely echoed by John Wind, senior director of marketing at Volo Communications Inc., who says his company is also fielding calls from current 3(Tone) channel partners.

"Our position has always been that we prefer to develop our own applications. We know about other platforms and the features and functions they have, but the key for us is to have the source code and to own and manage the interfaces with our other systems, rather than have a box parked in our network that we can't fully manage on our own. It's all about the level of management that's required -- we don't want to have to call an external agent in each time we want to change something."

He adds that "any service provider, large or small, using external technology will always face scaling problems. That's the nature of the beast."

And Wind is concerned that Level 3's decision to drop 3(Tone) will reflect badly on the Centrex market in general. "The hosted platform business does work, and it's safe and stable."

Not surprisingly, the commercial system vendors are eager to counter the claims of inflexibility. On the question of having to call in external support to make changes and updates, Broadsoft's head of marketing Scott Wharton says that's "patently untrue."

"Our system is designed to be changed and managed not only by our customers, but also through a Web interface by the end users of the service. Our system is highly flexible to a very granular level -- just ask any of our customers, such as BellSouth or MCI. This sounds like posturing to me. I'd be very surprised if these [in-house] systems could match anything near our capabilities."

And Wharton says Level 3's experience shows how important it is to have a system that allows service providers to be able to make the changes themselves. "Not everyone's needs fit the cookie-cutter. You need to address that need for flexibility."

Over at Sylantro, senior VP of marketing David Illing is very upbeat about the state of the IP Centrex market and demand for his company's products (see Sylantro Ready for Its Public). He is also quick to point to Level 3's admission that it is dropping 3(Tone) because it adopted the wrong strategy, and not because the technology (Sylantro's platform, in this case) didn't work properly.

"It has always been the case that service providers need to decide whether to build or buy, but if a service provider builds a system it needs to look at the features it can get out of its own technology in a time-to-market situation. The whole testing process, making a platform reliable -- all that work has to be done before services can be launched," and with a commercial system all that work has already been done, says Illing.

"We are testing millions of calls every night on our system. That's the sort of process that needs to be done. But it's up to each operator how much time and money they want to invest in a system. A commercial platform is good for some and not for others," he concedes.

On the question of supporting systems such as Sylantro's platform, Illing says his company trains its customers' technical staff, but that "people have to learn about the interfaces and how they work, and that can take time, maybe longer than if they had built the system themselves."

And it seems there is no shortage of service providers willing to take the support regime and license costs on board. Broadsoft, Sylantro, and Tekelec are all adding new customers, with Illing claiming that Sylantro added 11 new customers in the fourth quarter of 2004 alone. (See Sylantro Lands New Customers, McLeodUSA Selects Broadsoft , iiNet Offers VOIP With Broadsoft, XO Picks BroadSoft's VOIP Apps, Broadsoft Names New Customers, One Connect IP Picks Tekelec , and Telekenex Uses VocalData for VOIP.)

— Ray Le Maistre, International News Editor, Light Reading

iometro 12/5/2012 | 3:26:49 AM
re: IP Centrex: Build or Buy? I wonder if L3's real reasons are based on the fact that use of open source IP-PBX alternatives like Asterix and SIPfoundry (PingTel) are becoming much more widespread? Thus realizing they can't compete with free have decided to play to their strength which is wholesale termination.

For the perspective of customers and large end-users; why spend boatloads of money to be locked into a system where your are charged for every incremental increase when you can buy a few linux boxes, hire a few code-geeks and mass produce a dirt-cheap infrastructre?

Hosted Centrix I believe will end up following the market dynamics of the email market (i.e. hosted vs. on-site).
dljvjbsl 12/5/2012 | 3:26:44 AM
re: IP Centrex: Build or Buy?
why spend boatloads of money to be locked into a system where your are charged for every incremental increase when you can buy a few linux boxes, hire a few code-geeks and mass produce a dirt-cheap infrastructre?

This is going to be very good news to PBX suppliers. When this advice is followed there are going to be a great many companies which discover that their communication systems don't work and can't be fixed. PBX suppliers will be able to charge what they like.

These stories about code geeks creating customized communicaion systems have been around for at least 25 years. CTI is reinvented every five years or so and given a new name. It always ends in tears. Code geeks know about writing code they do not know about the desing of mission critical communication systems. There is a big difference.

paolo.franzoi 12/5/2012 | 3:26:42 AM
re: IP Centrex: Build or Buy?

Let me use my wife's cousin as an example. He is a mortgage broker with approximately 30 employees. These employees are in 3 locations. Needless to say he has no IT staff, just him. So, he takes many simplifications to get his employees computers in place as well as spares, service, etc. In that environment, Centrex services work. He can network his 3 locations, get better voice services, and employ no staff to buy or manage the switch.

alchemy 12/5/2012 | 3:26:42 AM
re: IP Centrex: Build or Buy? I've never particularly understood the value proposition of IP Centrex. Legacy Centrex mostly died decades ago when companies did the cost benefit analysis of owning their own PBX compared to letting AT&T charge them $75/line/month. I've seen IP Centrex/Hosted PBX get traction in Japan but that's due to regulatory arbitrage. In the US, I just don't see where a vendor can charge enough for the service when they're competing against a relatively cheap IP PBX and bulk minutes (either TDM or IP-based) deal. The PBX vendors have always given their gear away to get the upgrade business later so the initial CAPEX is very low.

I guess if you price it cheaply enough, you might be able to sign up some smaller businesses that have a geographically dispersed workforce. I just don't see where you could ever sell this kind of service to the Fortune 1000.

Am I missing something here?
dljvjbsl 12/5/2012 | 3:26:41 AM
re: IP Centrex: Build or Buy?
guess if you price it cheaply enough, you might be able to sign up some smaller businesses

The bulk of teh market is in teh sub-100 line size. Very large systems genrally require a great deal of customization and hence low margins. PBX vendors concentrate on the medium and low line size segments whre they obtain higher margins.

I suppose with IP Centrx the theory is a economy of scale. With current technology it is easy enough to develop central call controllers that can servicex many thousands of lines. This call controller can then be shared by multiple small businesses that are dispersed over a large area (in contrast to TDM Centrex). In any event it is the sets that will drive the acceptance of IP technology not the call controllers.
Sign In