Home
About PCB-Search
Advertising
Contact Us
Free Magazines
PCB Blog
PCB News
PCB Resources

PCB and Electronics News

Special Market Focus: Collaborative Design
By Ann Steffora Mutschler -- Electronic News, 1/27/2005

In the consumer electronics space, where products are embraced at breakneck speed, making it to market first can mean the difference between leading the pack, slugging it out later on pricing or -- worse -- being ignored completely.

Design organizations are under pressure to get superior designs to market faster than the competition, to take full advantage of shrinking windows of opportunity for maximum ROI. To meet the strict demands of today’s market, semiconductor companies have two choices, buy it or build it. If the expertise lies within the company’s four walls, it is typically built in-house. If the technology is outside a core competency, it may be safer to employ expert assistance or outsource the project.

Outsourcing can elicit either a cringe or a smile, depending on whom you speak to, but it has become a mainstay as global economics and consumer demand shift the way the electronics industry creates the latest and greatest chips. Add to that the design complexity of today’s leading edge ICs, which require careful coordination of the complete design process, particularly as 90nm process technology hits the mainstream.

More than anything else, complexity and outsourcing are the power behind the rising wave of collaborative design. With the semiconductor supply chain now consisting of a distributed worldwide network of focused entities, design outsourcing is clearly a part of that, said Jim Kupec, COO of eSilicon Corp.

“While design implementation is outsourced all the time, design intent is usually not outsourced. Intent remains a key part of the architectural design team. It is therefore critical that the team have tight interaction with the outsourced design implementation team to ensure design intent is met,” Kupec said.

Further, with design teams becoming increasingly dispersed due to available talent pools and economic factors, both geographically and over multiple time zones, collaborative design activity is skyrocketing.

Whether it is design groups within a single company or working with third parties, collaborative design can be characterized by highly integrated devices that are time-to-market driven, have tight size/weight/power budgets and market price points, embedded software, and a mixture of digital and memories with RF, memories and analog, said Steven Schulz, president and CEO of industry standards organization Silicon Integration Initiative (SI2) Inc.

“We see this in portable retail products: cell phones, PDAs, MP3/video, camcorders, and so on. The key here is that the Herculean task of IC design is now pressured along more dimensions concurrently -- size, cost, power, weight, seasonal market windows, etc.,” Schulz said.

John McCallum, high-technology industry manager at Microsoft, agrees. “The products are becoming so complex. A cell phone is not just a cell phone. It is a camera, a PDA and a cell phone. Many companies don’t have all the expertise in-house and have to reach out for expertise and create partnerships you never saw before,” he said, citing the recently announced Sony-Samsung licensing agreement as an example.

“[Complexity] is changing the way companies are viewing the high-tech market. We are now looking to the influence of the consumer for the way companies design products and their ability to manufacture for a global market,” McCallum added. “The game has changed and if you only focus on your core competencies, you will leave opportunities on the table.”

As a result, bringing the right players to your game is critical. “The engineering talent pool has undergone a transition over many years,” said SI2’s Schulz. “This has been building over 15 to 20 years from graduate school trends, and is no different than any other economic efficiency balancing adjustments. However, the post dot-com bubble and the post-9/11 economy have caused many design companies to take the risk of lower-cost options.”

With those lower-cost outsourced design teams providing value, they have become permanent and integral parts of the new design team structure at large corporations. Even small companies are getting in on the act. “Small companies and start-ups find that they are challenged by the rising NRE costs of designing and manufacturing chips and need to partner for breadth and for economic flexibility. The broadband trend makes networked remote access easy for even small firms and even consultants to share design data,” Schulz added.

Technology Dynamics Present Unique Challenges

Not to be overlooked, the technology dynamics driving collaborative design also cause unique issues.

“Advanced scaling requirements have driven device designs to 0.15-micron feature sizes and below. These advanced technologies utilize low on-chip voltages, making power integrity even more critical. Very high-speed I/O [chip-to-package, package-to-PCB, PCB-to-system] has caused the signal integrity of integrated devices to deteriorate in new designs,” explained Len Perham, CEO of Optimal Corp.

“Addressing these challenges is exacerbated because the problems are too sophisticated for designers to solve without the adoption of highly advanced software into their end-to-end design flow. These problems can only be solved in a team effort,” Perham continued. “Vertically integrated semiconductor manufacturers always had all of necessary disciplines in house. But today, the chip design team might be located in the U.S., the semiconductor foundry in Taiwan, and the packaging house in Korea or the Philippines. The signal integrity and power integrity problems are caused by the integration of all of these disciplines and can only be solved using an integrated solution where all of the unknowns generated by the various technologies are loaded into a sophisticated software solving engine and simulated, analyzed and solved simultaneously.”

Examined closely, complexity and outsourcing as key motivators of collaborative design seem to act as opposing forces: Outsourcing is driving a distributed, loosely coupled environment, while complexity is driving the need for tighter coordination at every step of the design and manufacturing process, noted eSilicon’s Kupec.

“Both forces are strong, and their opposing natures cause something we call, 'The Complexity Paradox.' The combination of these two forces has made successful IC design and manufacturing even more challenging than before. Only the most advanced organizations can successfully deal with this phenomena,” he concluded.

Read Part Two of this series next week, which addresses the real-world challenges of making collaborative design work, including case studies and tools available.

Special Market Focus: Collaborative Design, Part Two
By Ann Steffora Mutschler -- Electronic News, 2/3/2005

It is well known -- and expected -- that many aspects of business are impacted by global economic forces, and design collaboration is certainly one of them. Particularly in the creation of multi-million gate chips that combine IP from various sources, designed by teams on as many as three continents, the challenges are unique and varied.

For the most part, challenges lie in the detailed communication and interaction between experts, along with the management of the data they are trying to use, said Steve Chidester, product marketing director for Cadence Design System’s PCB tools.

Specifically, communication challenges take many forms, ranging from language barriers, time zone limitations and tool compatibility to network bandwidth, which all add different dimensions to the problem.

Technology challenges also have the potential to make successful collaborative design a monumental task. Ease of use, the ability to use tools externally through firewalls, security of IP, cost and integration with other systems all pose serious challenges, said John McCallum, Microsoft’s high tech industry manager.

And if that isn’t enough, consider the massive explosion in design data due to increasing system complexity at the high end for more functions and more integration of IP. There’s also increasing silicon complexity at the low end because of exponential transistor counts, OPC correction for sub-wavelength processes, new materials and manufacturing techniques, device physics effects from the fab that can no longer be ignored, and on and on. Taken together, this goes well beyond the predictions on the SIA’s road maps, said Steven Schulz, president and CEO of industry organization Silicon Integration Initiative Inc. (Si2).

“The data complexity problem can best be addressed by applying the four fundamental engineering principles of complexity management across the design flows: Hierarchy, Abstraction, Incrementalism, and Iteration,” said Schulz. “These concepts need to be applied throughout the flows used by the collaborative teams, or the data problem resurfaces.”

In addition, data file formats are often a problem given proprietary tools, flows, and formats for exchange, in addition to different versions of so-called standard file formats being used. The net effect of these issues, Schulz believes, is that sharing data is becoming the largest challenge to shared design.

To address this, Si2 has driven projects such as the OpenAccess standard, which aims to scale better with design complexity, manage the volume of data efficiently, and permit multi-vendor tool flows to “plug and play” without traditional file format barriers. Si2 reports its members have implemented OpenAccess in large advanced IC design efforts across the globe and across corporate boundaries.

Another take on the data sharing issue is the actual point when the data needs to be passed. “One of the core problems with communication between dispersed design teams within the same or an outsourced organization is the data hand-off point. This invisible line defines the separation of responsibilities and is always the point of contention when collaboration issues arise,” noted Adam Traidman, president of Giga Scale IC. “Unfortunately, this ‘hand-off’ point is all too often viewed as the place where one team throws a specification or design over the wall.”

Traidman pointed to an outsourced ASIC design flow as an example. “Specifications and general architecture are defined by ‘company A’, while ‘company B’ is hired on a contracting or outsourcing basis to implement ‘company A’s’ design and see it through to manufacturing and production. Problems arise because the tossing the spec over the wall mentality creates a virtual disconnect between design and implementation.”

“While EDA tools have come a long way in helping to optimize a design during the implementation flow, few if any have taken aim at solving the problem of keeping initial specifications tied directly to final implementation,” Traidman said.

Clearly, communication is the key to successful collaborative design. But what’s the best way to share information with all members of the design team, either internally or at an outside partner?

“Beyond a well-defined and accurate system specification, clear work division is critical,” said Jim Kupec, COO of eSilicon. “It is very difficult, if not impossible, to partition a single functional block between locations. On the other hand, partitioning functional blocks on a location basis, and clearly defining the interface requirements usually works quite well. This is typically in a hierarchical design.”

Cadence approaches data sharing by using a data vault as a repository, whereby a design “owner” will post project data, assigning access to team members. Versions of designs are checked in and out with comments and other team members are automatically notified of the changes.

Si2’s Schulz believes the key is in starting well. “I would begin with the overall design methodology required for the design, define a clean mapping of where the expertise exists for aspects of the design, map out where the handoff points will be across teams -- starting with written requirements specs and documentation for all data transfers -- then develop purpose-crafted design tool flows that implement the methodology with the experts. All of this needs to be architected and shared among the teams in advance, along with a process for conflict/ambiguity resolution.”

Giga Scale IC’s Traidman acknowledged that while there is no perfect formula for the most effective communication, there are steps to take, such as frequent communication, to make sure everyone is in sync. “The disconnect between the original design spec and [the taped-out design] is often caused by the original spec being incomplete. By design they are more general than the implementation. The further you are willing to drill down at the architectural level, the closer your implementation will match.”

Automated Tools Could Ease Collaboration - But Where are They?

Although it is a fundamental part of electronic design, the collaborative aspect of the design process has not been an area of great concern. But that is starting to change, said Laurie Balch, principal design and engineering analyst at Gartner Dataquest. “With so many pressing design issues, there were lots of other problems, more urgent to put money towards.”

Although the MCAD industry has long used product lifecycle management (PLM) and product data management tools (PDM), that activity had not carried over to the EDA space. Synchronicity was formerly the most visible tool for collaboration in the EDA space, but has been quiet since being purchased by Matrix One, a big player in collaboration tool market. Cadence has a relationship with Matrix One for data management in its Allegro Design Workbench.

On the bright side, Balch expects the collaborative tool market to take off for electronic design, although in the same way it did with MCAD -- slowly. “We are approaching the tipping point for EDA where more companies begin to recognize that technology can help them communicate more efficiently,” she said.

Finally, a player not widely seen in the electronic design tool space, but one that will increasingly show its presence in the semiconductor industry is Microsoft, which has a variety of tools to facilitate collaboration.

Beyond e-mail and office productivity tools, the software giant also has SharePoint, a team tool for collaborating and sharing data; an enterprise and secure version of Instant Messaging for real-time communication; the NetMeeting Internet conferencing tool with multi-point data conferencing, text chat, whiteboard, file transfer, and point-to-point audio and video; as well as commonly-used Project and Portfolio tools.

Not to be quickly dismissed, Intel, Samsung, ON Semiconductor, Flextronics and other top semiconductor industry company have been open about their uses of Microsoft technology to integrate design and manufacturing processes.

This is part two of a three-part series on collaborative design. Next week the series concludes with a “How-To” guide on setting up teams and partnerships on a global scale.

Special Market Focus: Collaborative Design, Part 3
By Ann Steffora Mutschler -- Electronic News, 2/10/2005

Understanding the challenges of collaborating on a worldwide scale is one thing. But implementing viable strategies can be quite another.

For companies that get it right, the rewards can include unique access to regional markets through partners, cost reduction and resource availability, as well as an inside perspective of business opportunities. The ability for innovations discovered in one location to be communicated to all collaborators is another benefit of successful collaboration.

It’s no surprise that most agree the best place to start is at the beginning -- the foundation of the partnership being the determining factor to success. “Forming a global team takes time and an understanding of the roles and responsibilities of each team,” said Phil Bishop, CEO of Celoxica Ltd. “One of the best ways to build a global team is to start with a consolidated team in one location and to bring parties that will relocate offsite into a standard set of processes. After the party is trained they can then setup an offsite design function that is set up to integrate and communicate with the originating hub.”

That message rings clear at Open-Silicon, a fabless ASIC house that creates moderately complex, high-volume custom ASICs. The company uses a highly predictable and reliable, factory-like approach, according to Naveed Sherwani, co-founder, president and CEO of the company.

“When we start a project, from that point on it is one team, one design,” he said. From there, roles and responsibilities are identified from which an ASIC program manager then builds a project plan.

In Open-Silicon’s approach Sherwani stressed that openness and trust hold the utmost importance because typically in collaborative design, people hide information. “Trust between companies is built by openness,” he said.

As collaboration propels us to think in a way that extends our company’s own four walls, how much partners trust one another may be one of the biggest issues, said John McCallum, Microsoft Corp.’s high technology industry manager.

At eSilicon, collaboration starts out on a more traditional course. “Our approach has been to first build a strong team and a solid methodology in the U.S., close to our customers. The proximity of the design team is critical to the high first-time success rate we have enjoyed,” said Jim Kupec, COO at the company.

The company is now in the process of replicating the methodology offshore to manage costs with criteria focused on teams with compatible style and methodology. “If you can get those two items right, the challenges of going offshore are minimized,” he noted. “We will always have a strong local team to interact with our customers, no matter where they are. We will supplement that team with groups in other locations that have lower cost factors.”

Establishing trust is only part of the challenge, though. Once that’s done, cultural and language issues must be addressed. Open-Silicon, for example, uses a fixed design methodology, picks its staff carefully, and trains them using a fixed formula. So far, in its Bangalore, India-based execution group, the approach has worked.

At the same time, Sherwani believes it is a unique situation that likely would not work in the United States for cultural reasons. In Silicon Valley in particular, design engineers are able to learn many different skill sets while on the job. In India, design engineers may stay on one facet of semiconductor design for years, he said.

While it may sound foreign to the U.S. semiconductor industry, this has allowed Open-Silicon to be successful. “We have design center units with experts, organized like a factory. There are different departments of expertise where each engineer is not a jack-of-all-trades. When a design is won, the right experts are chosen for the project,” he said.

Steve Chidester, PCB tools product marketing director for Cadence Design System Inc., which has six offices in India and continues to develop its presence there, says cultural and language differences aren’t easy to overcome. “Some [issues] don’t have easy solutions. Either you speak the language or you don’t,” he said.

Coordinating time schedules with its India-based offices has been a mixed bag for the EDA giant. If a problem comes in at the end of the day to Cadence in the U.S., it can be sent to an Indian office. If they can solve it, they do. But if they need more information, it pushes the solution back another day. Chidester says there is no easy solution, except sometimes to work odd hours.

At system level design tool provider Celoxica, one of the primary challenges it has seen with global collaborative design involves communication. “A specification that defines the work or the partnership in a comprehensive way must exist. The specification must have a firm set of deliverables defined for all the parties collaborating,” said Bishop. “We have found that defining a partnership with a clear demarcation to the work or service to be performed by all groups is critical to global alliances. Interfacing demarcated work segments is much more reasonable then trying to go with too high a degree of integration and interdependence on a particular collaboration.”

Further, the company relies on one common language for communication globally -- English -- for all collaborations and partnerships. However, a local or regional program management presence is essential to help insure that the same message is well understood by all parties.

Nowhere is this more apparent than in Asia, where a huge advantage to being accepted is the ability to speak the language.

Tony Curzon Price, CEO of Arithmatica Inc., says it is absolutely imperative that everyone understands each other, a problem that his company has encountered in the Japanese market. “Whenever you have a complex task to do you must collaborate among teams, cross-culturally and across companies,” he said. And with it, you have a whole new layer of translation to do in order to see the other party’s point of view.”

Layers of translation must take place not just with language, but also across cultures. Without this, projects can get off-track. “Some of the hardest moments have been deep into a project when the perceptions of the parties involved started to diverge. It takes someone from the outside that is prepared to stay neutral to make sure everyone is on the same page,” Price said.

In Price’s experience, particularly in Japan, the distributor is the diplomatic core because they explain any wrong steps and isolate the company from taking certain directions with customers.

What exactly collaborative design of the future will look like is anyone’s guess, but it surely will include the management of every conceivable combination of people, locations and data. “There are talented people everywhere and big companies are taking advantage of that - competition and shorter cycles are pushing it,” he added.

Fortunately, communication technology and global savvy is making it easier to do just that.

Finally, Microsoft’s McCallum foresees global collaborative electronic design taking advantage of technologies already available that are capable of connecting all members of a project team, from marketing to the design chain to the supply chain, so each party can focus on its core strengths.

This is the final article in a 3-part series on collaborative design. For the full series, click here

Choose from over 300 different titles including Computer,
Electronics, and Engineering Magazines for Qualified Professionals.