When cable operators first approached Voice over Broadband (VoBB) services, there was much about the technology that was yet unknown. Even once the case was made for the technology in general, cable operators still faced many challenges in actual deployment of VoIP networks. Although many of these challenges remain, the indisputable fact is that VoIP is here and it is here to stay. With the abundance of VoIP networks, whether or not VoIP will take the Darwinian step of replacing the traditional Public Switched Telephone Network (PSTN) depends on the ability of innovators to bridge the islands of VoIP, and provide interoperability and interconnectivity solutions, ultimately turning all VoIP networks into one global network.
Voice over Broadband services provided by cable operators are a compelling offering for a number of reasons. On the business side, a triple-play package of Internet access, content (TV channels), and voice is extremely attractive to consumers. On the technical side, since the cable network spans that last mile directly into the home, Quality of Service (QoS) can be controlled to assure positive user experience. By contrast, non-cable Internet Telephony Service Providers (ITSPs) like Vonage or Skype will have an uphill battle on both these points. It is little wonder that the cable VoIP offerings are the fastest growing segment in the consumer space and have by far overtaken other (paying) VoIP services in numbers of users, minutes of use, and revenue generated.
Cable companies are now looking beyond the birth pains of their initial deployments to the next generation of opportunities and challenges especially to VoIP Peering between Multiple System Operators (MSOs). Peering allows the cable networks to keep a VoIP call IP to IP, end-to-end. Currently, a VoBB user calling another user of the same Cable voice service will never go off-net. But a call placed to a user of a different VoBB provider, even if both end users have identical IP phones, will leave the first cable operators IP network, drop off to the legacy PSTN, and continue on to the terminating MSOs IP network and on to the end users IP phone. This detour in the communications route presents MSOs with a number of challenges and difficulties.
Firstly, multiple conversions between VoIP and PSTN protocols degrade voice quality. Secondly, an unnecessary cost component of PSTN per minute charges is added. This can be quite substantial in international dialing, and ultimately causes service providers to either transfer costs to users, or reduce revenue. The third, perhaps most compelling argument for peering is the retention of advanced features allowing IP communications to realize its true potential. When a call leaves the IP network, it drops to the lowest common denominator with the PSTN, namely toll-quality voice. On the other end of the spectrum, IP end-to-end communications allow for video, CD quality voice, voice/data integration, and features and applications yet to be developed.
While VoIP peering makes sense for other ITSPs, organizations, universities, and large enterprises as well, it is natural for the Cable industry to be the first mover in this space. MSOs are typically regional and dont compete for each others user base, so cooperation between them is a non-threatening, win-win situation.
The challenges involved in VoIP peering fall into three categories: Interoperability, Routing, and Security. VoIP network deployments using different softswitches or SIP proxies (not to mention different protocols) rarely work seamlessly with each other. It is usually necessary to tweak the call flow in one way or another. Session Border Controllers (SBCs) can often solve these problems, but the dynamic nature of the problem set with new versions of software or additional features and applications causing trouble makes maintaining, patching, and upgrading an SBC an operational headache.
Routing calls between MSOs requires each operator to determine if the call is destined for a VoBB user of another MSO or for the regular PSTN. Since many countries support Local Number Portability (LNP) there is no way a-priori to know if a particular phone number will end up on an IP network without a directory service comprising the phone numbers of all members. That database needs to be queried on a call by call basis. The protocol of choice for this lookup is Electronic Number Mapping (ENUM).
Security challenges are the most formidable and complex, and certainly not yet completely understood. By opening up their networks to each other, MSOs are breaching their walled garden and are thus susceptible to new vulnerabilities. These include things like Denial of Service (DoS) attacks, worms and viruses embedded in the signaling or media, and malformed protocol packets (accidental or malicious) capable of crashing VoIP systems. There are also higher level problems such as Identity theft (which is especially tricky in VoIP as the CallerID is usually open and unprotected) and Spam for Internet Telephony (SPIT).
SPIT differs from regular telemarketing on the one hand and from e-mail SPAM on the other. For spammers, since there is essentially zero cost per call, there is no economic constraint on the volume of calls. MSOs know that as soon as the VoIP space becomes a safe haven for spammers, users will relinquish these services, and they must act fast to prevent SPIT. However, methods of detecting SPIT are problematic since there is no content to examine and filter, the caller identity is generally untrusted, and the ringing phone alone is already intrusive and annoying.
Many of the above issues are best addressed by having a federated peering authority instead of the ITSPs attempting to peer directly with each other. The bookkeeping involved in managing the directory services, for example, is much simpler in the federated architecture. There is also greater flexibility for solving security issues with a central authority involved. Policy can be enforced and abuses monitored more efficiently. SPIT is another area that can be combated more effectively within a federation. It is easier to detect patterns that could identify a caller as abusive if calls flow through a central point than if traffic is solely peer to peer. For example, if a user sends out a large number of calls to users of multiple MSOs, each network will receive only a small fraction of those calls, and may therefore not notice any irregularity. Mining the call patterns at the crossroads between all federation members, however, allows for greater statistical accuracy in identifying abuse. The federation
also has information about the type of service its different members are providing, like call centers, and that profile may be important to other members deciding whether to accept or filter calls. The information regarding the profile, identity, and security level of the call needs to be embedded within the call flow so that it is available to a network or end user on a call by call basis. For example, a user may decide that specific types of calls at certain hours should divert directly to voice mail instead of disturbing him/her with a phone ring.
The first generation of VoIP was about proving the technological concepts, and the second generation focused on aggregating groups of users and deploying VoIP networks. The future of the industry lies in bridging these islands and connecting up the networks so that VoIP can reach its full potential and eventually replace the legacy phone system with a feature rich and robust communication platform. IT
Baruch Sterman, Ph.D., is founder and chief executive officer of Kayote Networks. For more information, please visit the company online at www.kayote.com (news - alerts).
If you are interested in purchasing reprints of this article (in either print or PDF format), please visit Reprint Management Services online at www.reprintbuyer.com or contact a representative via e-mail at [email protected] or by phone at 800-290-5460.