By Steve Jennis, Senior Vice President, Corporate Development,
PrismTech In recent years the 3G base station (BTS) community has pursued cost reduction with a passion. The combination
of tougher competition, especially from Asia, and increasingly price-sensitive operators (following the .com bust) created a need for drastic action that
lead to downsizing, consolidation, increased outsourcing and strategic
attempts to lower overall industry costs. One of these
strategic initiatives is a broad attempt to lower bought-in
costs through interface standardization with the intent of: - increasing network equipment provider (NEP) purchase choice; - lowering costly customization and integration requirements; - facilitating the use of more COTS components; and, - raising production volumes of commodity sub-systems. This drive led to the establishment
of several standards bodies earlier in the decade, including the Open Base Station Architecture Initiative (OBSAI) and the Common Public Radio Interface (CPRI), to define industry standard interface specifications for hardware
modules. These efforts are still on-going and producing valuable results related to hardware costs, which representapproximately 70 percent of aBTS's cost today. The fastest growing cost component of any BTS, however, is software. The software content of a BTS today represents approximately 30 percent of the total cost (source: Unstrung Insider). This figure jumps to 50 percent if the RF sub-system
is excluded. Furthermore, this percentage is expected to rise steadily as new functionality requirements (e.g. dynamic multi-mode, multi-band operation and air interface upgrades) are best addressed through software re-configuration. But software-based functional enhancement
driving-up software cost content is only half the story. Software-based re-configurability also offers the
prospect of reducing time-to-market (and thus increasing competitiveness) and
lowering OPEX costs by reducing or even eliminating hardware upgrades). So, functional, time-to-market
and field maintenance requirements are all driving greater software content. Consequently the ability of a NEP to efficiently manage the software lifecycle is rapidly becoming a critical business requirement. With software management so
critical to future competitiveness and cost control, where is the
software equivalent of OBSAI or CPRI, an industry body defining BTS software standards to facilitate reuse, outsourcing and component-based software system assembly? Is the lack of a software equivalent
of OBSAI or CPRI because: · NEP vendors still have a hardware mentality to cost control? · Are they conveniently ignoring the growing software cost content because software engineering is outside of their skill base · Is it because they see software development as a 'craft'
industry with little scope for standardization or reuse? · Is it because the hardware vendors
have no interest in supporting hardware-independent software
development (and thus only produce software tools
for their own processor families)? · Is it because software re-configurability is assumed to come at a performance cost that
is unacceptable? Maybe it's a bit of each? But, whatever the rational, a head-in-the-sand attitude to software cost management will soon be a major strategic weakness. Fortunately there are initiatives
that the BTS community can leverage. The Object Management Group has for several
years hosted a domain task force developing standards for software-based
communication systems. This group has developed platform-independent models (PIM) for building software radios that
can be mapped to various platform-specific architectures (PSM).
With these standards-based models, COTS software
vendors can produce implementations and tools that not only provide highly-productive software development, but also support the concept of component-based assembly and There may be good reasons for BTS vendors to move to new hardware regularly (as semiconductors get steadily more powerful, efficient, and cheaper), but there's no Moore's Law for
software. It's more expensive
tomorrow than today, At a recent International Base Stations Conference, Alan Gatherer, CTO of
TI's Communications Infrastructure business presented the following
predictions: Is this the future of Base stations? COTS embedded middleware providers, such as PrismTech, are active in addressing the issues
of software proliferation, complexity and reliability faced by
OEMs when bringing-to-market multi-standard,
reconfigurable, low-cost base stations. Unless better-quality software environments
are deployed (see bullet #3 above), time-to-market, maintainability
and cost control will be increasing-- · a clean separation between software
applications (waveforms) and hardware platforms (processors); to
ease both software portability (re-use) and hardware upgrades, thus facilitating cost reduction; and, · software component frameworks to facilitate the outsourcing of application
development and the assembly of software applications from components
from disparate sources, again facilitating cost reduction in an
increasingly distributed supply chain. In summary, market demands for increasing functionality, shorter time-to-market and cost reduction in a multi-Standard
world provide the 'business drivers' for Standards-based embedded
middleware and high productivity application design and deployment
tools. Just as RTOSs became 'buy, not build' COTS products in the
late 1990s, COTS embedded middleware is now on the market to support
the development of reconfigurable base stations. If you would like more information
on software tools and embedded COTS software platforms for base stations, please contact PrismTech at www.prismtech.com Author's Biography Steve Jennis is responsible for devising, implementing
and coordinating PrismTech's strategic and tactical business plans
throughout its worldwide operations. He has more than 20 years
experience in sales, marketing and general management positions
in high technology organization. While at PrismTech, he has established
PrismTech's North American operations and developed the worldwide
sales and marketing functions. Prior to joining PrismTech in 1994,
Mr. Jennis was the general manager of Texas Instrument's Computer
Products Division and his 16 years experience at TI included more
than seven years in international marketing and strategic planning
positions. He is Physical Sciences graduate from Loughborough University.
Software
Cost Control in Base Stations ? ? software reuse
both within and between product lines. This approach is often referred
to as product line software engineering. Without these standards-based software architectures, COTS middleware and tools are not developed, the maintenance and evolution of custom software is expensive, and reuse is a pipedream. especially if you have poor tools and no strategy for reuse.
, not declining--, problems. For its part, PrismTech is investing heavily in standard-based communications software products
(frameworks, middleware, development tools, etc.) to deliver: