You have a point; although, as IQ indicates in the post below, Docomo's use of brass boards may have more to do with politics than anything else because they certainly have access to advanced EDA software tools like those that, as you pointed out, compliments and enhances the prototype process. Moreover, Docomo's use of bass boards may also have something to do with the development of the right package of services to support wireless data. Right now a carrier can take advantage of the open field approach or a walled garden approach, which I think allows a deregulated wireless carrier to create shareholder's value faster. Anyway, here are 2 excellent posts from IQ.
benahavis, that is a great article about NOK and IDCC.
I used to work in the EDA (electronic design automation) industry, selling top-down system-level design and simulation tools. The approach described in this article is very advanced.
For those who are perhaps not as attuned to this type of technology, let me try to explain how significant this really is.
In the past, hardware designers had to take their concepts to implementation by actually building prototypes and testing the hardware in the lab. You can imagine how this can drag out the development cycle if the first few prototypes don't work as expected. It takes a long time to build and test prototype hardware.
Now as systems get more and more complex it becomes impractical to build prototypes for the whole system. And when you throw software into the mix, it becomes almost impossible to evaluate the performance of the entire system until everything is finished and working.
There has been a move for years from schematic-based design (drawings of circuit elements that are interconnected by lines depicting "wires") to language-based design. Typically, language based-design uses VHDL or Verilog (the two most popular design languages) to specify the design. The advantage here is (1) it is much faster to specify the hardware by typing in "code" than drawing pictures, (2) language-based designs can be easily simulated before actually building prototypes, and (3) language-based designs can support higher levels of design specification that can be fed into automated "synthesizers" that produce the actual gates of logic used in the design. All this speeds up the design process and reduces errors in the finished product.
VHDL and Verilog have been around for some time, and most IC's are developed using one or both of these languages. However, many designers still work down at the "register" level when doing their design. For complex systems it is becoming impossible to design at this level of detail. Instead, system designers now specify the behavior of the system they want to create first, verifying that behavior at a high level, then going about the process of implementing the various building blocks at a more detailed level. This is called top-down design, because the designers don't worry about the details until the computer understands the desired behavior. Then the computer is used to "synthesize" (or automatically generate) the actual circuits that will produce that behavior.
All this is requires very complicated software in order to support the simulation (verification of behavior) and synthesis (automatic generation of circuits that behave as specified at the top level. Companies that produce such software are at the very leading edge of technology.
IDCC has done a lot of work in-house to produce their own proprietary systems and techniques for doing this sort of top-down design. The fact that they not only understand the wireless technologies they are building, but are also breaking new ground in the area top-down design methodologies is very impressive.
In this article we get a glimpse into how their design methodologies work and how they can be used in the industry to transfer their designs to other companies. This is important for NOK, as the article explains, because many of the blocks developed by IDCC must be passed to engineering groups at NOK for inclusion in their ASICs. Many engineers will have to receive and understand IDCC's building blocks, in order to integrate them with NOK's technology.
The work that IDCC has done in this area puts them way ahead of many other companies, in that they not only have the technology and expertise, but can easily transfer their technology and expertise to other companies that may want to license from them.
"For the pilot project, the components were taken through the whole implementation path, from system-level specification, via C-models, to VHDL RTL and synthesis, targeting a prototype gate array technology."
The beauty of such an approach is that desigers can use the optimum language and the optimum technology for their application. It is impressive to see such a design environment where a complex system can be supported in C/C++, VHDL, Verilog, or the emerging SystemC. In fact, it appears this sytem allows interchangability between behavioral and functional levels.
"All these levels can be described in C/C++ using tools such as CoWare N2C or emergent system-design languages such as SystemC. Layers 0.0 and 0.x can also be described using RTL VHDL or Verilog."
As IDCC seeks to expand its revenues from engineering services, capabilities like this will be key. This will also be very important as they move toward ASIC production with one or more semi partners, because their designs will have to be transferred to multiple organizations and verified on different platforms.
ragingbull.com
Dagrinch, thanks. Re: HG's comment, "But, when we sent our (ASIC) design out to the foundry (to produce a single working prototype chip), it came back working perfectly on the very first try! ...we expect design wins by the second half of 2001."
First, he appears to be referring to the BDCMA ASIC that TI produced. Yes, they were successful on the first try, which is a testament to IDCC's system modeling/verification capabilities....
ragingbull.com
Bravo, clarence, excellent post.
This is very key information. As I understood it, Japan (along with many others) submitted WCDMA proposals in the early stages of IMT-2000. Several steps were taken to harmonize the various WCDMA standards prior to arriving at the one final standard adopted. Japan, wanting a leg up as you wrote, decided to deploy early, prior to finalization of the standards. So the "brass-boards" that will be used early on will be based on those aspects that NTT thought were stable at the time they moved forward. Of course, some things changed, so their initial deployment will not be 100% compliant with IMT-2000, but will be similar in many regards.
Eventually they must bring the system into compliance in order to support 3G and global roaming. In the interim they get to make some revenue and position the Japanese wireless industry to capture large shares of the ultimate global market.
And this indicates the importance of the Japan office for IDCC. By making relationships now, they can participate with NTT and other Japanese manufacturers working with NTT in making this variant on WCDMA work. That will obviously involve IDCC IPR and know-how.
siliconinvestor.com
Never mind apples and oranges, the important thing is that IDCC is not trying to sell the wireless equivalent of left-hand drive cars in Japan. Their sales effort in Japan is being brokered by a well-respected Japanese who knows how to get things done especially at a time where there is tremendous pressure being exerted -- by the US primarily -- to change their telecommunications pricing structure.
Have you noticed how it is American companies who are usually the first to exploit the negative vibes that Qualcomm creates when it uses the US State Department or the US Trade Representative as a "marketing arm?" Th latter phrase was commonly thrown around during the run-up to IMT2000 and caused a lot of deep-seated resentment that probably explains some of the events that have happened with QCOM's installed base. IDCC's turn may be coming soon now that Qualcomm has been maneuvered into Asia's haggling zone. |
