 | Issue: | North America 2009 |
| Article no.: | 9 | |
| Topic: | The ICT relay race | |
| Author: | Cliff Townsend | |
| Title: | Vice President, Business Development, | |
| Organisation: | Galazar Networks | |
| PDF size: | 200KB |
About author
Cliff Townsend is the Vice President of Business Development at Galazar Networks. Previously, Mr Townsend established and oversaw Galazar’s sales, marketing and operations teams. Mr Townsend has held a variety of senior engineering and business roles at telecommunication technology companies such as Nortel Networks, Skystone Systems, Cisco Systems, and Innovance Networks. Cliff Townsend holds a BE in Engineering Physics from the University of Saskatchewan and an MBA from Queen’s School of Business.
Article abstract
To compete effectively, network operators must constantly upgrade their networks to handle rising volumes, steadily mounting bandwidth requirements and increasingly demanding applications and content. Operators rely on network equipment providers (NEPs) not only for faster, better, more cost-effective equipment, but also for the knowhow to get it running, and stay running, smoothly. To provide better functionality at lower cost, NEPs increasingly depend upon network semiconductor providers (NSPs) to design these functions in ever more complex, sophisticated and lower-cost chips.
Full Article
The race to deliver new information and communication technologies has not been an individual sport for decades. The AT&T divestiture in the early 80s began to change process the communications industry from a vertically integrated model to a value chain with many highly specialized members. That process continues to evolve the way information and communication technologies (ICTs) are delivered and is causing a dramatic imbalance between value creation and profit margins. Before divestiture, a single entity planned the network, defined the products necessary to build it, did the R&D and manufactured the products needed to turn the plan into reality. In a monopolistic model like this, the entire organization knows precisely what product characteristics are needed (quality, features, price) and when. It is also easy to maintain a profitable business. Fast-forward to the present. Network operators provide the services and plan the network evolution. They rely on network equipment providers (NEPs) to supply platforms with the right characteristics when they need them. NEPs draw on contract manufacturers (CMs) for platform assembly and on component and module suppliers for platform features. Component and module manufactures in turn draw on their own CMs. This model is much more complex and open to competitive forces. Competition is good – right? It drives innovation and keeps costs down. The problem is that post-bubble competition continues to influence the profit margin of the NEPs. They in turn are passing the pain on to their suppliers; contract manufacturers, component and module suppliers. The entire value chain is under severe stress and members are competing horizontally and vertically for the margin they need to survive. The margin necessary for survival varies significantly for different members. Organizations that bulked up during the boom have infrastructure costs that are out of line with a stable market, let alone today’s economic environment. In order to survive they must extract margin from other members in the value chain that they inhabit. A good example is the thin margin that the suppliers of optical modules receive compared with the margin some NEPs receive for reselling the same modules. Evolve or perish. The value chain for communication technologies continues to evolve and so must its members. We have seen consolidation and restructuring aimed at adjusting the financial characteristics of larger organizations for better alignment with the industry. Time will tell if this evolutionary branch can support sustainable organizations in the long-term, and it is certainly creating short-term uncertainty. Another evolutionary branch is spawning smaller, more focused organizations that create value within their own segment of the supply chain. This model is less encumbered by overbuilt infrastructure and is, therefore, able to maximize the relative investment in innovation. A specific example is the value chain that includes suppliers of integrated circuits for complex networking functions – network semiconductor providers (NSPs). NSPs have grown to fill the market need for effective, highly integrated solutions to complex communications networking functions. Some were created by the divestiture of the semiconductor divisions at many of the large NEPs (e.g. Lucent – Agere). Others have grown organically with venture financing. The challenge for NSPs is to get the product characteristics right (features, price, quality) and to hit the market window. Market timing is a growing challenge. Predictions this year at OFC/NFOEC 2009 (Optical Fiber Communication Conference and Exposition/National Fiber Optic Engineers Conference) were for annual network traffic growth ranging from >50 per cent for wireline to >100 per cent for wireless. At that rate of growth it’s difficult – some say impossible – to avoid a limited supply situation. NEPs face a difficult decision when developing new products. Should they shorten the development cycle or increase the designed capacity? How many NEP CFOs feel comfortable investing in a development program today that yields a 10x performance improvement to full field deployment in less than four years? Quality is usually the first victim of aggressive product developments. After all, you can set premium prices if you are well ahead of your competition; you can hold the market with your product footprint (NEP or NSP) as you address the quality shortfalls. Unfortunately, network cost must ultimately increase to cover product upgrades and dependable service is often unacceptably delayed. Features are also commonly traded-off to gain time. You can often get a product into dependable service with an acceptably reduced feature set and a solid plan for upgrade. The ultimate platform for this approach would offer field-provisionable hardware and software for the complex networking features. Components exist today in the form of FPGAs (field-programmable gate array semiconductors) that can partially satisfy this approach within the limits imposed by technology and cost. Unfortunately, you cannot download FPGA resources or microprocessor memory when new network features exceed their hard limits. Again, network cost must increase to cover unanticipated upgrades and intended features are often unacceptably delayed. Price is the elephant under the table. Consumers have become accustomed to getting more for less when it comes to communication technology, so they expect network operators to increase bandwidth without increasing service fees. NEPs, therefore, are pressured to supply equipment that meets increasing bandwidth demands at a declining price – by some estimates you need four times the bandwidth at less than twice the price. The question is – who is making the investments required to achieve these gains? For one, the NSPs are. Moore’s law predicts a semiconductor technology platform that can satisfy the growing demands for network bandwidth, give or take a bit. NSPs are exploiting that semiconductor platform to deliver increasingly complex networking functions. The good ones are evolving their development techniques to reduce their costs and improving their support infrastructure to improve their NEP customers’ development cycle. The result is an overall reduction in development costs and development time for new communication products. How does it work? Refer to the standard development process; the exploration phase identifies the appropriate product, the planning phase assures a deterministic downstream process, the development phase produces the intended product, the validation phase assures that product characteristics are aligned with the intent and the production phase delivers it to the market. To be successful NSPs need to embrace many of the characteristics typically associated with NEPs within their development process. First, the NSP has to understand the market need. It is not good enough to listen to any one member of the value chain, and the listener is just as important as the speaker. That leads to the first important characteristic for an effective NSP; the team must be comprised of people with direct experience and contacts within the downstream value chain. That gives the NSP the ability to appreciate the subtleties of network needs and provides them with a virtual consultancy of contacts within the industry. That often resolves the question of which products to develop. Apart from good execution during the planning and development phases, the next significant requirement is a network approach to product validation. For this, the NEP must validate their ICs (integrated circuits) within a representative hardware and software platform using realistic network scenarios. Without this platform, only a limited set of configurations can be exercised and only in a static environment. With the flexibility enabled by this platform, an automated validation process can exercise various and dynamic network scenarios around the clock. That solves the requirement for delivering a quality product. Extending the NSPs definition of the validated product to include the IC abstraction software allows NEPs to quickly integrate pre-validated functions into their network equipment. This makes it possible for NEPs to integrate new equipment into networks quickly by minimizing their development and validation time associated with the functions provided by the NSPs’ ICs and software. That solves the requirement for delivering new networking products at the appropriate time. If an NSP can deliver the right product with good quality at the right time, what is left? The elephant – price. Most NEPs have developed their own ICs in the past and their CFOs know that they cost 0.2 cents per pico-hectare. That is the cost of goods that NEPs traditionally used for calculating their equipment cost. Along comes an NSP pricing a similar IC at two cents per pico-hectare and it looks like an order of magnitude increase in the NEP’s cost of goods. The fixed cost associated with the IC development that was previously buried in the equipment’s development cost is now exposed in the COGs. There is the imbalance between value creation and profit margins. IC development costs have migrated from NEPs to NSPs along with the responsibility for network level features, quality and market timing. Margins, however, are harder to let go. Many NSPs have established effective processes to innovate and deliver the functions required to meet the network growth demands. All they need in order to continue their contribution to the success of the value chain is their fair share of margin to fund their investments. The message is simple; the excesses of the telecommunications boom are long over and nobody is approving investments without a solid business case. For a healthy value chain, each member must remain viable. Otherwise, the components, modules and networking equipment necessary to sustain the predicted growth will not emerge. Like a relay race, each member must make a clean transfer, adding value as the information and communication technology moves up the value chain toward the consumer.
