February 2001

Network Re-Soul 

BY KEVIN MAYER

A few months ago, in an article entitled "Soul Of A New Network," we described various architectures for packet-based telephony. Among other points, we described the tradeoffs between so-called open approaches and platforms built "from scratch." In general, the open approaches -- that is, those relying on commercial components, such as those available from Natural MicroSystems (NMS) and Dialogic -- emphasized the benefits of fast time-to-market and development flexibility. And the so-called custom platforms emphasized the importance of carrier-class density and capacity, while indicating that the scale of the solutions amply justified proprietary components.

Since that article appeared, we've received feedback that the distinctions between these different approaches may have been overstated. Also, in the interim, vendor announcements suggest commercial approaches are persistently chipping away at capacity limitations.

For example, according to Heidi Bersin of Clarent, "Whether the components are sourced from open systems vendors like NMS/Dialogic or manufactured based on TI/Samsung chipsets on customized boards doesn't really make much difference in the end. It's a question of architecture, and the real differentiation is on scale, density, and whether system features are configurable on the gateway."

As for the critical point -- architecture -- Bersin adds that most carriers will probably have a mix of components in their networks, depending on their migration needs, and the feature sets they need to execute compelling business cases. Thus, evolving next-generation networks may consist of integrated access devices (DSL/cable/WLL/T1/E1/DS3) on the premises and a combination of feature gateways and pure packet-based switches in the network. The feature gateways are platforms that accomplish voice compression, IVR, TDM, and advanced features, and the packet switches are tandem devices that use IP as the trunk select rather than pure TDM.

Bersin notes that while many are focused on high-density tandem replacement, the tandem is only about 11 percent of the switch market. "The big focus and source of the greatest return on investment will be on the access side of the equation -- integrated access devices that live on the premises, but are controlled and provisioned in the network by carriers," she said.

As for new commercial developments, we'd emphasize interim solutions intended to minimize the limitations of bus architectures while maintaining the benefits of mass-market silicon and software. The consensus is that solutions such as Infiniband are well into the future. Instead, look at architectures such as those suggested by StarGen and PLX, as well as extensions and supplements to CompactPCI and H.110.

[ Return To The Top ]
[ Return To The February 2001 Table Of Contents ]

BY ALI KAFEL

Access technologies are constantly changing and expanding. Recently, the only way you could remotely access the network was by dialing through the traditional PSTN. The rise of the Internet has brought us the birth of DSL, cable Internet, and even wireless Internet. The end user's appetite for more bandwidth and new services motivates carriers and network infrastructure vendors to find new and better ways to provide access to the Net.

Although less than 2 percent of Internet users, only 6 million people worldwide, will access the Internet via wireless technology this year, that number will explode to 484 million by 2005 (according to Ovum research). This dramatic growth represents the culmination of three major trends: First, the Internet is transforming the way we communicate, encouraging new applications that demand more from both the network and end-user devices. Second, we have a proliferation of powerful and relatively inexpensive notebook computers, PDAs, smart phones, and other handheld computing devices. Third, the worldwide deployment of cellular infrastructure continues apace.

In light of these trends, carriers need new infrastructure that accommodates all methods of access. Users will demand relatively seamless and unified services regardless of their access type. One approach to this challenging problem is implementing a multi-access intelligent system at the edge of a unified core transport infrastructure. The mixture of voice/data traffic over a wireless access technology, like any other access, needs to be groomed and switched based on intelligence defined in customer profiles as well as service profiles, and matched to the specific wireless parameters.

Integration of these various profiles on one multiservice platform allows a unified approach for offering services and improves both the economics and the simplicity of providing multiple services. The network architecture would include elements such as hardware switching, signaling, call control, and application servers to support access methods such as traditional PSTN, DSL, cable, and fixed wireless. For mobile subscribers, location management functions and additional elements such as home location register (HLR) and visitor location register (VLR) need to be added.

Network architects and planners building the new public network should support multiple access methods for their subscribers, especially ones designed to support wireless Internet. This means a typical VoIP gateway that simply converts TDM traffic to packet technology is not the answer. Carrier class multiservice switching platforms designed to support broadband and narrowband access for fixed and wireless networks is a better alternative. A solution offering with building blocks for all access types, including wireless, utilizing a phased approach, is the ideal path to the new public network.

Ali Kafel is vice president of marketing for Telica, a developer of intelligent multiservice broadband switching platforms for CLECs, ILECs, and next-gen service providers. 

[ Return To The Top ]
[ Return To The February 2001 Table Of Contents ]

BY DENNIS GATENS

There is no doubt that the convergence of voice and data services on a single, next-gen packet-based network is imminent. Crucial technical barriers have been overcome, and the business drivers are clear.

VoIP technology is widely accepted, particularly with the emergence of DSP-based processing engines supporting carrier class voice features. On the technical front, other favorable signs include the maturity of key technologies such as call control, the deployment of enabling transport technologies such as DSL and cable, and industry efforts towards proving interoperability.

The business drivers are compelling: a capital cost advantage for packet-based versus circuit-switched communications, open architectures enabling the timely deployment of new and enhanced services, and new revenue opportunities.

However, for next-gen solutions to succeed, equipment manufacturers must offer scalable solutions from PCM-only to full-featured carrier class gateway solutions that meet their demands for high-density, low power, and low-cost per channel. Required features include: G.711 PCM; G.168-2000 echo cancellation (up to 128 msec with multiple reflections); full tone detection and generation; voice activity detection/comfort noise generation; network management support; and packetization (IP and ATM). It will also be necessary to support the following features: fax relay; low bit rate CODECs (including G.723.1A, G.726, G.728, G.729AB, G.729E, GSM-ER, GSM-EFR); and RTPUDP/IP header compression.

In addition to this feature set, which includes several complex, computation-intensive algorithms, platform providers will need to consider new standards, the demand for new and enhanced services, and requirements for global interoperability. As service providers target densities of 100,000 channels per bay, carrier class gateway platforms will require a new generation of programmable DSPs that will make it possible to process many channels of full-featured voice and fax on the same DSP while significantly reducing the cost and power per port.

Service providers are looking to the next generation of platforms to scale to very high densities supporting their converged voice and data service offering strategies and business models. But these platforms also must support carrier class voice quality. The objectives of density and quality can be achieved with next-gen DSP solutions that provide high channel density and the advantage of programmability to support a dynamic evolution to a packet-based infrastructure.

Dennis Gatens is product management director for Telogy Networks, a Texas Instruments Company.

[ Return To The Top ]
[ Return To The February 2001 Table Of Contents ]