TMCnet - World's Largest Communications and Technology Community




Special Focus.gif (4903 bytes)
February 2000


The State Of Switching


Unless you’ve been living in a cave for the last few years, you know at least a bit about the efforts underway to transform the circuit-switched PSTN (public switched telephone network) of the last hundred years into a packet-based network capable of delivering a wealth of enhanced services. Our coverage often tends to the actual services that will become available as vendors deploy all sorts of new switching elements into the fabric of the next-generation network. In this Special Focus, I’ve undertaken the task of asking several influential vendors in the networking and switching community for their views regarding the state of switching today and the potential for the future of telecommunications.

One interesting trend I found in most of the responses is the belief that the so-called next-generation network has already arrived and is starting to carry more minutes and offer more services to the lucky few who are jumping aboard early on. For argument’s sake, I asked these forward thinkers the following question regarding today’s PSTN: “If it ain’t broke, why fix it?” The responses all gravitated to the ease and efficiency of deploying enhanced services, and the potential for generating revenue from such services. All the vendors agreed that no carrier is ready to swap out billions of dollars of legacy equipment overnight, but that in time, the PSTN as we know it will evolve into something altogether different. Carriers may not replace everything right away, but when the time comes to buy new equipment, they’ll be looking at next-gen elements to replace their legacy iron. Some respondents defended the best points of today’s PSTN (ubiquity, reliability, quality of service…) while still acknowledging that as technology allows us to replicate those features on tomorrow’s networks, the PSTN will slowly fade into the background.

I’d like to thank the participants for their frank and insightful answers. (A complete list of Web sites for all the participating vendors appears in a sidebar at the end of this article.) I urge readers to not only visit these Web sites, but to contact me with any questions or comments about this feature, perhaps offering their own views on the state of the next-generation network, and what steps need to be taken to ensure that the evolution stays on track. Feel free to send e-mail to ggalitzine@tmcnet.com

State Of Switching Featured Participants

Cisco Systems
As evidenced by SBC’s recent $6 billion investment, service providers are investing heavily in a new digital broadband infrastructure to capture new revenue and profits from data and integrated data/voice services on packet networks. While the PSTN has served well as the multiservice network of the past, it is no longer capable of generating new revenues and profits. As data volume predominates, driven by high growth, it only makes sense to find a way to merge voice onto data networks.

Service providers with a large installed base of data on today’s voice network (e.g., private lines) must move this data onto their packet networks to enable value-added services — their new profit engine. Providers who choose to retain their costly and inflexible “Old World” equipment will also be at a significant cost and time-to-market disadvantage relative to providers with New World packet networks.

To be successful, packet networks must accommodate multiple transport types with a common set of services and policies. MPLS will be a key enabler, as it is transport independent and enables value added service delivery though VPNs, QoS, and traffic engineering. Open protocols such as MGCP and JAIN will be critical to voice service delivery.

[return to the listing]

Mariposa Technology
During the AIN era (still underway), the traditional PSTN switch has been broken into its constituent parts, including intelligence, switching fabric, intelligent processors, signaling interfaces, and traffic interfaces. What we are seeing now is that the traditional components are being replaced by next-generation components: Servers, packet switching fabrics, signaling mediation servers, and edge devices that include multiplexing and media gateway functionality. In many cases, the edge has been extended to the customer premise, in the form of a next-gen IAD. In effect, the AIN guys accomplished the spirit of their mission, but in a way totally foreign to their original expectations.

The next-generation network (NGN) is ATM-based. The Internet is already mostly built on an ATM backbone, and ATM is still the most effective way to integrate services in the bandwidth-restricted local network. But ATM is simply a very efficient transport technology.

The NEXT next-generation network (NNGN) will use IP as the convergence layer. However, it will be an IP that bears only passing resemblance to yesterday’s IP. IP is gradually developing the toolkit to allow it to provide service quality and manageability that is required to deliver reliable, high-quality, converged services. It still has a ways to go, but it will get there. When it does, it will provide new service-creation capabilities that will fuel a century of innovation.
The next-generation network is being built in parallel with the last generation network (LGN). It is gradually eroding sales of new LGN equipment... capping the LGN... and accounting for the largest proportion of network growth. It is very much a work in progress (see above).

Mediation devices are being built that allow translation of voice and data among the PSTN, ATM, and IP networks. More importantly, mediation of signaling and network applications must be accomplished. This is still in very early stages. It will be a long time before the packet voice network has the service richness of the PSTN (CLASS and custom calling features, LNP, etc.). It will almost certainly have its own, unique features first (unified messaging, click to talk, etc.).

As for the capabilities and qualities that must be exhibited by “next-generation” switches in order to take the state of switching to the next level?

  • Scalability to terabits... if not the individual devices (in the short term), at least networks.
  • Reliability equivalent to current PSTN (99.999%).
  • Service Flexibility — user-programmable services that don’t have to be deployed ubiquitously to be useful.
  • Manageability — the carrier (and increasingly the user) must be able to measure and manage all aspects, from the physical to the application layer.
  • Reverse compatibility with the PSTN.
  • Interoperability with other vendor equipment using standard protocols.

All this is fun, but it is going to take decades, not just years. Old technologies don’t just vanish... they become those annoying “legacy” things that you just have to deal with. The ideas are exciting and the hype sells magazines and conferences, but people buy stuff that works reliably and cost effectively. For now, the cutting edge of reliable and cost effective — at least for converged network services — is ATM.

[return to the listing]

Lucent Technologies
The “switching network of the future” will be the ultimate convergence solution connecting a variety of networks — wireless, optical, voice, and data. From the enterprise to the core of the public network, next-generation switches combined with high-speed optical interfaces will deliver unlimited scalability to support the exploding demand for bandwidth.

These multiservice network switches will link any device or protocol to the network with the unprecedented reliability and quality of service featured in today’s voice network. They will be highly flexible with the ability to handle voice, video, and data traffic over wireless and wireline networks.

Proprietary carrier class voice switching solutions that work on single platforms are fast disappearing. Next-generation switches will be better because they will be able to handle voice, video, and data services seamlessly on a variety of server platforms, all interoperable with each other.

Next-generation multiservice switches (both hardware and software solutions) will be able to handle hundreds, perhaps thousands of applications and services — from basic voice calling to exciting next-generation services such as interactive conference calls and click-to-dial Web services. These intelligent voice/data applications will be controlled by innovative policy management software that enables administrators to provision quality of service to meet changing business needs.

Lucent’s vision for next-generation switches is to build universal platforms that support circuit and packet switching technology, using open protocols or Applications Program Interfaces (APIs) such as TSAPI, and developing software that is flexible enough for the switching technology to meet the demands of users who want bandwidth immediately.

[return to the listing]

The “next-generation network” is already here. It has been brought about by the influx of new technologies into a deregulated telecommunications environment. Three trends are emerging: A packetized local loop extending all the way to very small businesses and homes, open softswitch architectures which divide call processing intelligence from the underlying transport and switching, and packet voice trunking, which is reducing the overall cost of switching and transmission. The PSTN network is evolving rapidly from a closed system to an open network architecture with hooks for new service offerings provided by open software platforms.

Packetized access in the form of DSL, cable modems, wireless local loop, and powerline technologies has democratized the choice of service and service providers. These technologies in a deregulated environment are the on-ramp to the new network. Originally targeted for packet data, these access networks are rapidly being transformed into multiservice networks with support for voice and other services.

The edge of the network has also evolved. Call control, which has typically been embedded in the network hardware, is being ported to open platforms. There is substantial progress being made to standardize common call control for both packet and circuit-switched networks. This will unite the new network with the legacy network and create an environment for standard service delivery as well as new service creation with a common language for handing services off between networks. CORBA is being adopted widely as an open object-based protocol for developing the service creation and service assurance applications, which will bring scale and reliability to these networks.

Packet voice trunking enables service providers to freely mix voice and data networks on a common high-speed backbone with dynamic bandwidth allocation and increasing quality of service.

These three trends are being manifested in the network today. The network is already different and in its “next generation.” The coming years will continue to see evolution along these three basic lines towards a single network for all services.

[return to the listing]

Avici Systems
The switching market is evolving in a number of directions, with a single central theme: As silicon becomes more capable, the depth of function and speed of operation continue to skyrocket. In both the LAN and the WAN this is evident in the rise of multi-layer capabilities, which allow switches to provide QoS at wire rate to diverse network flows. The next trends are going to be increasing interoperation of the IP service layer and the underlying optical layer.

The next generation network is here. A number of carriers are either rolling out or trialling new networks built with IP running directly on top of an optical core, without the intervening layer of SONET equipment.

The biggest help to a smooth migration is the deployment of DWDM (dense wave division multiplexing) on existing fibers. This allows carriers to run parallel networks — legacy and next-gen — concurrently. As the next-gen network picks up an increasing percentage of the traffic, the legacy traffic can be migrated onto the new network.

Next-generation networks will be increasingly auto-provisioned. As the intelligence in switches increases, the switch hardware will be able to detect network bottlenecks and increase service capacities in real time. The end result is that switches will need to be scalable, reliable, and have the ability to dynamically add or redirect bandwidth as network load changes.

[return to the listing]

Nortel Networks
Convergence in a data-centric telecommunications world has resulted in service providers’ Points of Presence (PoP) starting to look more and more like large campus backbones based on Ethernet switched architectures. Layer 2 gigabit Ethernet switches are providing low-cost connectivity, scalability, and simplified connection management (over point-to-point architectures). This is enabling service providers to meet the challenges of burgeoning bandwidth demand, and of provisioning new services profitably.

As demand skyrockets for Internet telephony applications and the delivery of managed services, Ethernet switches in the next-generation service provider network will require 10 gigabit Ethernet interfaces that address capacity requirements and ensure compatibility with carriers’ and service providers’ optical infrastructure. These 10-gigabit Ethernet WAN-compatible interfaces will forego the creation of additional delays and infrastructure costs where data and optical networks meet, thus enabling seamless integration of local-, metropolitan- and wide-area networks (LAN, MAN, WAN) based on a common technology. Among market segments to realize the greatest benefit from this type of 10-gigabit Ethernet switched PoP is the application service provider (ASP) industry. ASPs will provide outsourced enterprise eBusiness solutions that require a tighter bond between the private and public worlds of telecommunications.

[return to the listing]

Lara Technology
Throughout the Internet’s existence, its traffic load has steadily increased at an average monthly rate of 10 percent, and it’s currently projected to double every three to six months. Much of this Internet data explosion is now being fueled by the combined effects of millions of new users, far richer content (e.g., multimedia), and the migration of corporate traffic onto the Internet. Overall traffic (both voice and data) is predicted to continue to increase rapidly throughout the next few years. As escalating Internet traffic demands push beyond the limits of today’s switching architectures, an entirely new generation of highly scalable switches will hold the key to successful evolution of the network infrastructure to meet tomorrow’s requirements. Currently, the convergence of two major trends is dramatically changing the basic nature of the switching environment. First, the emergence of Internet Protocol (IP) as the de facto choice for carrying the whole range of network traffic, from data to voice, is laying the foundation for a seamless internetworking environment.

We feel the communication industry is undergoing phenomenal changes and in our opinion, a big communications tornado is forming on the horizon. It is bigger than the world’s telephone network and will offer services and applications beyond what we can imagine today.

As a result of this tornado we expect a widespread deployment of a broadband, packet-based public network. We believe it is currently undergoing the early adoption phase and will be ready to take off in the tornado mode within the next few years.

This next-generation switching architecture must be based on an entirely new approach to delivering services that is specifically designed to accomplish the following:

  • Deliver robust switching functionality at a cost that is an order of magnitude lower than traditional, proprietary Class-5 switches.
  • Distribute switching functionality to the edge of the network to protect existing investments by supporting all current analog and digital network standards, interfaces, media, and service elements.
  • Reduce the number of network elements by combining a range of telephony, application, and service-delivery functions to enable new service creation through programmability and the flexibility of an open application programming interface (API).
  • Provide a high degree of scalability, enabling network operators to expand their subscriber base rapidly and cost-effectively.
  • Promote extensibility through open architecture design and, thus, take advantage of future technological advances to redefine true, carrier-class design for maximum fault tolerance and zero downtime.
  • Reduce operating costs by employing advanced remote maintenance and diagnostics capabilities.
  • Increase revenues by shortening time to market, reducing up-front costs, and providing remote management capabilities.

    [return to the listing]

Convergent Networks
The movement to the “new public network” is the single most important development in communications since the creation of the telephone itself more than 100 years ago. In the past, communications services have been delivered through devices that operate over a series of independent networks. In the future, communications services will be delivered through an entirely new set of devices that operate over a single integrated network. The simple elegance of the new public network benefits service providers and end users alike.

But this future will only be realized, and tolerated, if the new public network provides communications services with the same level of quality that consumers expect from the PSTN today over technology that provides carriers with a significant return on their investment. As a result, next-generation switching equipment must meet particularly high standards for reliability, scalability, and flexibility. Next-generation switches need to achieve the same “five 9s” of availability and high QoS of traditional switches. In addition, next-generation solutions must scale to unprecedented heights and achieve new lows in per-port cost to justify the needed investment. Finally, next-generation equipment must be useful for multiple applications and must interoperate with the existing PSTN. Flexibility will allow service providers to further justify their investment, and interoperability will provide a smooth migration path from old to new.

[return to the listing]

Performance Technologies, Inc.
Just a few years ago, advances in switching were mainly concentrated in the enterprise LAN market. With the passing of the 1996 Telecommunications Act, the wide area and metropolitan area communications markets are now experiencing the same level of dynamism. “Convergence” is the WAN buzzword with new start-ups being minted monthly. In the LAN space, many switch vendors are addressing needs for load balancing, QoS, and total fault tolerance. Many companies are swapping old hubs for layer 2 and 3 switches to improve performance and simplify overall IT services management. Going from 10 Mbps to 100 Mbps is a “no-brainer” and adding gigabit Ethernet won’t strain the systems administrator’s support and training budget. Another trend is the move toward server farms and creation of SANs (Storage Area Networks), the driving force being the need to gain better control over the server environment in terms of data reliability while improving the overall management.

In the past, bandwidth was extremely expensive in the WAN, less so in the LAN. But with the emergence of DWDM (Dense Wave Division Multiplexing) and terabit switching for the WAN, we are about to enter a new era of bandwidth economics where demand for raw bandwidth may no longer outstrip supply. Trusted industry analysts forecast a $4.5 billion investment in long-haul DWDM networks by 2002 as many network operators face fiber exhaustion in specific locations. Further, fundamental breakthroughs in optical physics will radically change the WAN landscape, with implications for switching technology, architecture, and economics. Lastly, IP telephony will become an enormous growth market as service providers rush to develop new services for packetized voice.

For reasons of pure economics, carriers are not about to rip out decades of installed equipment. But they are prepared to build overlay and bypass infrastructures while maintaining the existing equipment as they transition to a new service paradigm. Many telecommunications vendors are building gateways and conversion products to solve the vexing problem service providers experience operating and maintaining multiple networks while trying to provide new types of services. The new breed of service provider does not have the baggage of the traditional carrier, but must deal with the existing network and is therefore not immune from interoperability requirements. As the public network evolves from today’s circuit-switched PSTN to the IP-centric packet network of tomorrow, there will be ample opportunity for vendors to build “bridge” products of all types to ease this transition.

We are starting to reach the limits of speed and performance with silicon. Next-generation switches must be all-optical. To get to the next level, photonics must replace electronics. This will provide the ultimate jump in reliability and corresponding drop in cost for building and operating the next-gen nets. One of the most significant benefits of an all-optical switch will be increasing the velocity of provisioning an OC-48 or OC-192 circuit, shrinking from today’s typical 80 to 90 days down to but a few days over the next year. The economic ramifications of this are enormous, not to mention the enhanced customer satisfaction and increased loyalty for the competitive service provider.

[return to the listing]

Merlot Communications
Much attention has been paid to the “Softswitch Revolution” and the development of new-generation networks. However, what is really exciting about this transition is what it enables — a new breed of service provider. Prior to the emergence of new-generation networks, carriers were required to provide both transport and services. Thanks to the advances in softswitch networking, service providers can provide just that: Services.

Carriers are migrating to packet-based (cell or frame) networks for two reasons: Cost and regulatory pressure. This process will not be immediate and for some time there will be a mix of PSTN, ATM, and IP networks. However, this is merely the first step. This migration is causing the commoditization of network transport, opening the door for new-generation service providers to focus on providing customized enhanced applications and services and merely reselling transport from the lowest cost provider. This dramatically lowers the cost of entry since capital expenditure is limited to intelligent premises-based softswitches and a centrally located feature repository. Together, these elements comprise what is envisioned as a ServicePage Architecture.

In this architecture, service providers can define features to suit specific industries or even specific customers and store them on the feature repository (or server) as XML-based Web pages known as ServicePages. When a customer requests a particular feature, the premises-based equipment accesses the appropriate feature definition on the feature repository and executes it. This model is transport agnostic; it will work whether the network is IP, ATM, or even the PSTN. These new service providers that leverage this ServicePage Architecture are now able to differentiate themselves on the basis of the applications and services they provision rather than just their transport methods and costs.

[return to the listing]

The rise of the Internet along with such business, regulatory, and technology factors as IP telephony, local competition, and “next generation” data-centric service providers, has started a revolution in telecommunications markets, services, and infrastructures. The next generation of services is enabled by the convergence of packet and circuit networks and services, which will continue to gain momentum into the next century. SALIX envisions a highly competitive service provider environment that will demand rapid — and cost-effective — deployment of innovative services based on evolving customer needs and requirements.

Driving this evolution is the concept of Class Independent Switching. With Class Independent Switches, traditional services can be merged with communications services based on the more flexible Internet protocol, enabling the delivery of services never before possible — all while maintaining quality and reliability.

Class Independent Switches make the deployment of cost-effective, unified voice, data, and video services possible. Converged network infrastructures will replace service-specific networks and provide integrated services on demand. Unified service packages, tailored for specific markets, will enable service providers to rapidly address new markets with custom applications.

As we move forward, scalability and flexibility of the Class Independent approach will provide the investment protection and robust technology required to deliver new services well into the future.

[return to the listing]

Spring Tide Networks
Switching means different things to different people: Frame relay, ATM, or perhaps Ethernet if we’re talking about the LAN environment. ATM was a great innovation, because it allowed us to perform switching in hardware, rather than software, for the first time. Over the last few years, we’ve seen a move to perform IP forwarding in hardware as well. For a time, cell switching was the low cost way to move data, but with the new packet hardware, we’ve seen the switching costs on the packet side come down dramatically. From our perspective, we’re all about applying service functions uniquely to individual IP user flows, based on a user-level policy.

Things are changing very rapidly. In the transmission network, we see things moving away from legacy SONET ring architectures to partially meshed topologies than leverage DWDM. Soon we expect to see wavelength routers, which can find a best path for an individual wavelength across a complex network topology. At the edge of the optical network, there will be mediation platforms that map traffic from various edge service platforms. Of most direct interest are the emergence of the new IP services layer, which needs to exist between the core switch/routers and the edge aggregation gear (RAS boxes, DSLAMs, edge routers, FR/ATM switches, etc.). It will be from this strategic location that value-added IP services get delivered.

This means having the right interfaces and software features, so that you can deploy a next-generation network element in an existing network architecture. It also means having the right set of instrumentation so that the equipment can be managed and monitored using the existing set of OSS tools and operator skills. For us, this means performing multiple functions: User authentication and authorization, service selection, virtual routing, encryption, compression, address management, firewall, and QoS services.

[return to the listing]

Technology Marketing Corporation

2 Trap Falls Road Suite 106, Shelton, CT 06484 USA
Ph: +1-203-852-6800, 800-243-6002

General comments: tmc@tmcnet.com.
Comments about this site: webmaster@tmcnet.com.


© 2020 Technology Marketing Corporation. All rights reserved | Privacy Policy