Two years ago, softswitch advocates emphasized softswitch architecture would leverage In- ternet technologies, along with a broad community of independent software developers using popular application development tools, to create a rapid innovation spree akin to the World Wide Web. By sharing web technologies, voice and communications in general could become an integral component of any commerce, supply chain, customer support or other business process already popular on the web. They also admitted writing new calling features for legacy telephone switches could take years.
So, heralding the applications promise took a back seat to painstaking labor and debates over signaling protocols and 'hooks' needed to allow applications to talk to underlying network equipment and arduous attempts to replicate thousands of existing Class switch features before getting onto the business of creating truly new stuff.
The result: Frustration has set in among some pioneer vendors.
"The energy spent on replicating Class 5 features may have been in vain," says David Butler, director of softswitch product management for Convergent Networks Inc. and founder of Technology Control Services, which Convergent acquired in 2000. "A lot of work has been done on Class 5 features, and there's also a lot of political and regulatory barriers to migrating from them."
Consequently, Convergent has thrown up its hands and turned to ILEC competitors, along with an emerging set of application service provider partners. "We've taken the approach today that focuses on empowering other service providers to compete with the ILECs instead," Butler says. "That may be the interexchange carriers who are very interested in taking business applications forward and taking top customers away from the incumbents."
Other vendors, such as 3Com Corp. subsidiary CommWorks Corp., Alcatel Alsthom, Hughes Software Systems, Nortel Networks Ltd. and Siemens AG, are less vocal about their disappointment. They say they accept an evolutionary pace.
"The Internet brought user control, with each person's access as personalized as it gets, while horizontal telephone network features are as impersonal as it gets," says Houman Modarres, director of softswitch products for CommWorks. "There's a space between those extremes where microservices augment those macro telephone services to increase customer loyalty."
As a practical matter, CommWorks and others say they must offer multi-protocol mediation to accommodate the ILECs' need to protect their huge investments in web-unfriendly technology. Chief architect of Siemens SURPASS IP packet-circuit consolidation solutions Van Phung says it will be up to the incumbents to make the integration of packet into the world's vast, legacy public network happen. He notes competitive providers carry only 10 percent of voice traffic.
From Convergent Networks' perspective, there are few en- cumbered candidates among that 10 percent who are champing at the bit to follow through on the original promise of packet telephony, which is service innovation.
"The real threat to the ILEC lies with ASPs, who understand [Internet-based] SIP [session initiation protocol] distributed call control and only need the hooks in the softswitch to participate deeply in service creation," Butler says. "ASPs would directly intervene in the process of turning [AOL Time Warner Inc.'s] AOL subscribers quickly into telephony customers. The IXC has the opportunity to let the ASP, working with the softswitch, become their advanced intelligent network."
He adds, "We're seeing hosted PBX players getting worn down by the lack of uptake by the ILECs and beginning to look at the ASP outsourcer. You don't see [ILECs offering] a PBX service designed specifically for law firms, but an ASP is going to do unique things for a law firm or hospital or other vertical industry, because they're deeply entrenched in the ways their customers do business."
IBM Corp., a veteran of AIN (advanced intelligent networking, a way to bring enhanced services to Class switches by using peripheral devices) and ASP infrastructure and operations, has unified what had been separate Websphere ASP and Telsphere telco product units, Bulter notes.
Another veteran, Hewlett-Packard Co. "has developed models and specifications for automated discovery of an application, rather than the AIN's batch loading of applications in lock step in every node before executing," Butler says. Now, with SIP, "you can proxy an application from one machine to another, with service discovery working underneath to automate the flow from service creation to implementation."
Big Stick
Butler is not alone in marking the potential for a few competitors to force the incumbents to pick up the pace.
|
Voice web platform provider Telera Inc. brought Butler's formula to fruition in a partnership that marries IXC Qwest Communications International Inc.'s packet and telephony networks with enterprise voice applications hosted by an ASP. In this case the ASP also is Qwest in the form of its Cyber.Solutions unit. The offering combines voice recognition and call center management applications, enabling customers like Time Inc.'s Synapse magazine subscription unit to save $500,000 a year by offloading network and call center infrastructure while retaining control of features and call routing.
With many CLECs out of the way, "ILECs have cash flow and a monopoly and don't need to change quickly, so the dinosaurs have inherited the Earth, and vendors have to operate in that context," says Telera CEO and co-founder Prem Uppaluru.
On the other hand, "The IXCs face price erosion in consumer services and a mature enterprise services side requiring new services," Uppaluru adds. "Qwest is well positioned to attack the small and medium business market. It is a network service provider that decided to become an ASP, and we've proven the application infrastructure technology and business models are there to allow incremental migration."
Other IXCs, like Level 3 Communications Inc., "are proving that a large IP network operator can elegantly integrate your services over one pipe to reduce your costs and increase productivity," Phung says. "That puts pressure on the ILEC, who makes 50 percent of his revenue from the 17 percent of lines that deliver business Centrex. The ILECs can grow revenue only through geographic expansion or new service creation, and both are supported by softswitch deployment."
The executive vice president for Probe Re- search Inc. Hilary Mine agrees.
Enhanced services features like Centrex and more standard add-ons like call waiting and caller ID generated more than $11 billion in revenue and $7 billion in profit in 2000 for U.S. ILECs, according to Probe's "Super-Competition" study. Mine says incumbent service providers want to guard these high-margin services closely but are being challenged by a variety of competitors, including IXCs and AOL.
The fact that some 70 percent of traffic within an enterprise is intra-company creates an opportunity for VoIP as well, says Mine, who points to SBC Communications Inc. that recently lost tens of thousands of dollars from a Centrex customer deciding to move to an IP solution. Similarly, military, government and educational institutions are big, price-sensitive users of Centrex and are willing to try new technologies.
Mine concludes that IP-PBXs could take a big bite out of incumbent Centrex revenue, if they don't offer a comparably priced alternative.
To be sure, the incumbents are migrating away from mainframe telephony switches to server-based softswitches.
A study the International Softswitch Consortium commissioned, "Current & Planned Uses of Softswitches & Next Generation Networks in the U.S.," found 67 percent of ILECs and 43 percent of IXCs interviewed already are carrying live customer traffic on a softswitch (for more on ILECs and VoIP see xchange February). However, the same survey, released in January, found only 19 percent of carriers created new revenue streams through enhanced services, behind Internet offload at 29 percent and Class 4 replacement at 26 percent.
Convergent's Butler suggests ILECs are ignoring the power of their own reach. "The telcos are in a position to expose the most users to an application," he says. "The carriers know when I'm on my phone, when I'm on my cell phone and when I'm on my Internet-connected device. What more does an application layer need?"
CommWorks' Modarres agrees with Convergent on the need for a common applications layer that can be decoupled from network type, but he puts the point differently. "Right now, my wireline phone, cellular phone, paging and Internet services don't know that I'm the same guy with the same wallet. They don't offer me the same services."
He also agrees devoting time to replicating what the PSTN already can do is "a sucker's bet. Parity shouldn't be underestimated, but that energy is best spent on new services that haven't existed before," particularly where they can be integrated with and augment business processes for high-margin enterprise customers.
|
While Convergent essentially has turned its ship to tack around what it sees as legacy Luddites, CommWorks advocates a path that accommodates legacy and envisions services to augment rather than to compete directly with or replace PSTN services.
Butler says an applications layer that gets to the single wallet and opens multi-service innovation can become independent of "the plumbing" only through highly distributed call and feature set-up controls offered by SIP -- not the Media Gateway Access Control Protocol (Megaco) favored by the ILECs.
"Click on Outlook to generate a call has not been deployed widely because it doesn't fit into the centralized AIN implementation model," he explains. SIP enables a network to "discover" an application from anywhere across a highly distributed set of proxy call control points.
Modarres and others laud SIP and argue 3G wireless components, such as the call-state control proxy, already are effectively the same as SIP, which "has good traction and seems positioned to become the de facto, especially in the end devices, because of its small footprint," says Ajay Gupta, vice president of Hughes Software Systems operations, Americas.
"We've concluded it will be SIP in 3G user terminals, and the application server will become a SIP development server and will also dominate 3G."
Gupta continues, that more work is needed in Parlay and other application logic exchange protocols to enable feature interaction with the legacy phone networks. "The Class 5 has to communicate with that distributed app."
Convergent's Butler says that may come eventually, but in the near term "the primary line can sit over in the corner, and on the second line, we can give customers all the new services."