October 2000

BY RAMESH RAO

IP technology is opening the way to cheaper, better enhanced services. In the wake of global deregulation, increased competition is driving Internet service providers, local exchange carriers, and cable service operators to introduce brand-differentiating services, including toll bypass, wireless, and video, while at the same time trying to find new ways to both standardize and reduce expenses. These new services will be offered via "converged platforms," that are capable of handling voice, data, and video streams over the same connections. In conjunction with these new services and options, voice-over-IP (VoIP) technology has provided -- and will continue to provide -- telecommunication operators new approaches to network implementation and service deployment.

AN OLD MODEL, A NEW WAY
While the world moves to IP-based communication, most terminal equipment -- plain telephones and fax machines -- is not packet-based. These devices, and the local exchange switches or PBX installations to which they are connected, are not likely to be replaced by intelligent packet-based devices in the short term. Therefore, extra hardware is needed to access the packet network, which will replace traditional circuit-switched networks.

The challenge for OEMs is to develop systems that can deliver all of these services not only cost effectively, but at the same level of availability that public switched telephone network (PSTN) operating companies have conditioned consumers to expect. On the hardware side of the system, carrier-grade, full-scale media gateways will be required to manage thousands of media channels at a single site. They could also potentially manage thousands more at several remote sites. On the software side, a level of management infrastructure based on today's traditional telephone model networks will be necessary.

DEFINING THE ARCHITECTURE
In accomplishing these ends, it will not be practical to simply re-use the tools employed in the PSTN, because those systems are based on proprietary architectures and are meant to support circuit-switched protocols. What then is needed is a commercial off-the-shelf set of components designed in order to support packet-switched protocols and network elements in conjunction with a robust, proven UNIX operating system, an "open systems approach," if you will. With commonly available hardware and software, service providers will no longer have to choose proprietary technology and the associated costs to acquire and maintain specialized networks. They will be able to focus instead on service features and functions, while relying on a vigorously competing community of OEMs to provide standardized building blocks. Success in this arena depends on three factors:

• High-performance computing for service delivery based on commercial off-the-shelf RISC processors and board-level technologies;
• A high-availability system-level framework for the open systems platform; and
• High-performance computing for switching and voice and data processing based on commercial off-the-shelf DSP board-level technologies.

Until now, commercial off-the-shelf equipment could not provide the level of availability established by traditional PSTN systems at an acceptable price point. Inexpensive PC-based system components have appeared, but they lack the crucial characteristics to support high availability, a trait users have learned to expect from a telecommunications company. Alternatively, high-availability system elements offered have been expensive because each product is based on a different proprietary design.

STANDARDIZING THE BUILDING BLOCKS
Fortunately, new standards-based systems are now emerging based on well-engineered, scalable processing subsystems that can speed development time and substantially improve time-to-market. Standards-based hardware, an open framework, and the requisite firmware and higher-level application software to support cost-effective, high-availability solutions for IP-based telecommunication are finally in place. Vendors have a choice of several components: CompactPCI as the hardware platform; ECTF H.110 for telecom buses; cellular and G.7xx standards for codecs; and V.90 and H.3xx for access platforms. For OEMs, this means that open standards solutions are, at last, capable of meeting their needs.

The gap between the PBX and IP-based communications can now be filled with a new high-availability system-level framework that is built around off-the-shelf commercial components and then coupled with high-performance computing solutions for voice processing and switching. Now telecommunication providers will be able to deliver advanced voice, data, and video features and functions, while relying on a vigorously competing community of OEMs to provide, and adhere to, standardized building blocks.

Ramesh Rao is group product marketing manager at Sun Microsystems. Sun Microsystems, Inc., has long been synonymous with leading edge technology. Now, after 18 years of telling the world "The Network is the Computer," Sun is at the forefront in the emerging network-driven economy. 

[ Return To The October 2000 Table Of Contents ]

BY JON KENTON

What started as an Internet revolution has become a communication evolution. A far cry from a "simple" 24/7 dial tone over our handsets, telephone networks comprised of switched-networks have become overloaded by new data-transmission demands. The incredible pace of change in the telecom industry, comprised of new technologies that have been shoe-horned onto the network's fabric -- WAP, VoIP and a host of others -- is forcing the telecom networks to adapt a feature-rich, digitally based, open-system approach to high-availability voice and data convergence. In fact, telecommunication networks must adopt new standards and methodology concerning the implementation of hardware and software in order to handle the loads created by next-generation systems.

HIGH AVAILABILITY IS NOT AN OPTION, IT IS A REQUIREMENT
Having piggybacked data onto the voice network for so long, it didn't take a big leap to look at how to add voice onto data networks as well. This has lead to many opportunities in which new companies have set up operations as voice providers when they build interconnections into the conventional phone networks when they own only a packet-switched data infrastructure. Service revenues from VoIP have been small by comparison to existing services but are forecast to exceed the $5 billion mark by 2003. VoIP gateways, which are responsible for the interconnect of the old and new networks, have a CAGR approaching 40 percent according to the MultiMedia Telecommunications Association (MMTA).

Infinite dial tone and now infinite data tone are what we expect. Unavailability or outages in these networks, effecting our businesses and personal lives simply will not be tolerated.

CompactPCI = Open Systems
Today's service providers and operators demand a flexibility that requires a fresh approach. Distributed, open-standard, IP digital networks that provide new feature-rich, affordable, easy-to-access, and value-added services require underlying technologies that provide unprecedented connectivity, reliability, availability, scalability, and serviceability.

For equipment providers, time-to-market is now a major facet in staying competitive and open standards hold the key to being able to respond quickly to market demands. Features and functions of the overall solution are the basis for decisions made by operators, and while hardware design is still a factor, the bus or processor that powers a switch or a gateway matters much less than the speed at which they can bring services online and start billing customers.

CompactPCI is now the open standard of choice for new infrastructure designs. Resources previously deployed on hardware designs are now being focused on software value-add. As an open standard, CompactPCI brings many benefits:

• Standard processor, chassis, and communications hardware are readily available "off-the-shelf." This availability, along with industry-standard operating environments and software, help to dramatically cut development times;
• Without the need to train on proprietary architectures, staffing becomes easier due to an increased knowledge pool;
• Vendor competition is promoted by open standards, ultimately speeding innovation and reducing prices.
• Competition also ensures upgrade paths to the latest chipsets and technologies; and
• Single sourcing issues are made easier, as products may be purchased from a range of suppliers.

The attributes of CompactPCI from a technology perspective reflect the needs of the telecom community. The PCI bus was well established and the associated wealth of slot cards prevalent with PCs was attractive. However, there were limitations that made it difficult to safely include them within designs requiring high levels of availability and reliability. PCI's physical shortfalls were addressed by combining IEEE 1101 mechanical standards with the PCI electrical specification creating CompactPCI, with the robustness required for carrier-grade infrastructure deployment. CompactPCI had high availability in mind from the get-go, as evidenced by a key feature -- the hot swap specification -- that defines three models: Basic, full, and high availability.

Hot Swap and High Availability
The ability to add and remove components from a live system, reducing system downtime is fundamental for high availability. Shutting down and powering off a system prior to the addition, removal, or replacement of system boards and components causes unacceptable levels of unavailability. Systems with hot-swap capabilities range in their ability to reduce this downtime from hours of system unavailability per year to less than five minutes per year. Predictably, as the availability of a system increases, so too, does the complexity of the hardware and software of that system.

The three levels of hot-swap capabilities defined by PICMG build upon each other, increasing system availability at each step.

Basic hot swap establishes the fundamental capability to add and remove boards from an active system. Staged power pins and control circuitry ensure that the physical insertion and removal of a board will not cause bus glitches or damage the component itself. At the basic level, operator intervention is required to control the software ensuring it is in the correct state for insertion or removal.

Full hot swap extends the basic model with the addition of hardware and software enabling the system to be notified automatically of a board's removal or insertion. The operating system controls the configuration without operator direction, reducing opportunities for error, as well as the time required to perform hot-swap activity. A microswitch is added to the ejector handle of the CompactPCI board. Activation of the microswitch upon removal generates an interrupt (ENUM). The operating system identifies the board that is about to be removed and gracefully un-configures it. The operator is notified that it is safe to complete the removal of the board when an LED is illuminated on the face of the board. The microswitch activates the reverse process when a new board is inserted, ultimately the LED turns off to indicate the completion of configuration and that the board has been accepted into the system.

High availability systems significantly increase the levels of control that the operating environment has over the total chassis. APIs and middleware (user level software outside the operating system) are added, allowing a greater granularity of management and an interface for the user applications themselves. In a high-availability system, the state of chassis components can be further controlled via the addition of the hot swap controller (HSC). An HSC providing bus and slot control, as well as chassis alarm and status registers, is itself hot-swappable. Individual slots can literally be turned on and off through software control providing significant benefits in fault isolation.

Software and Hardware Working Together
Hardware architectures such as CompactPCI can create the foundation for high availability, but to build a truly highly available solution requires further hardware considerations such as redundancy schemes and a broader, holistic approach. Three major building blocks working in cooperation are needed to build such a solution -- the hardware, the operating environment, and ultimately the application itself.

High availability is a real requirement and we all want it, now. The speed at which end users are devouring network bandwidth and subscribing to new services continues to increase, as does our dependency on them. CompactPCI offers developers a step up, providing the platform for applications needing high-availability attributes, as well as meeting time-to-market goals. 

John Kenton is the telecom marketing manager for Motorola Computer Group. Motorola is a global leader in providing integrated communications solutions and embedded electronic solutions. 

[ Return To The October 2000 Table Of Contents ]