We use embedded computers everyday - we don’t see most of them because, as their name implies, they’re buried deep inside almost everything that effects our daily lives - from industrial equipment to domestic appliances. Although they don’t receive anything like as much attention from the media as does the ignoble PC, the sheer number of embedded computers and their economic importance is considerable. In fact, around ninety percent of all microprocessors manufactured in any one year are destined to be embedded inside electronic products.
There’s no getting away from them - those clever pieces of silicon that pervade our lives. The electronic radio - alarm arouses us from our slumbers. The central - heating has already turned itself on - just long enough to take the chill off the bedroom. We take a shower with lashings of hot water. A quick look at the early morning TV, station - hopping with the remote - control. Silicon surrounds us as we drive to the office - dozens and dozens of little wafers of silicon, making sure the car starts, the brakes work, the temperature is just right and so it goes on throughout our working day and into our relaxing evening.
In 1997, some 30 million microprocessor chips (or to give them their proper name, Central Processing Units - CPU's) were shipped into the PC World. Compare that with the number of CPU's destined for the embedded computer world - well over 3 Billion. And around one billion of those contained powerful and sophisticated software systems valued at an estimated $10 billion.
Designing of Embeded System
The processes and components that make up an embedded system application are much like those of the general purpose computing world - hardware, software and data, perhaps with networking thrown in. The key difference lies in the speed with which the embedded application must work on the data it receives to produce a result, and the reliability of that result. It has to work very quickly and it has to work every time - just think of the braking system in your car. You would be less than happy if the software controlling those brakes took a few seconds to decide to operate whenever you hit the brake pedal!
An added constraint in the embedded systems world is cost. If your video recorder contained the latest top-of-the-range Pentium microprocessor it would be unaffordable commercially. So, low cost chips (hardware) with very high quality applications (software) and accurate data are the order of the day in the embedded systems world.
Traditionally the approach to designing a product that requires an embedded application has been a two - handed affair. A requirements document is the input used for the specification/partitioning phase of the process. After the design is partitioned into its hardware and software components, the detailed hardware and software design and the implementation are done independently.
When both are completed they come together during the integration and test phase. This is the first chance that the software has to run on the hardware and vice-versa. Invariably, changes will have to be made and so it’s back to the drawing board - usually just for the software as it’s too expensive to redo the hardware at this point in the product cycle. Eventually a second prototype is created and the testing process is repeated. Too often, it’s back to the drawing board once more, or else trade-offs are made of the "it’s not exactly right but it will have to do" variety and the product is shipped to the customer. Frequently the product does not have the full functionality promised and does not run at full speed.
Obviously, this process of designing and developing hardware and software separately is time consuming, expensive and unproductive - three aspects which have a profound effect on the competitiveness of a company. If only the hardware and the software could be designed together and tested at a very early stage in the life-cycle then costs could be reduced, the final product could be brought to market more quickly and the resulting gains in productivity would improve the competitive health of the company.
In this article, we will look at the state-of-play in the embedded system world today, and at a European Union initiative to develop the tools and techniques needed to design and develop embedded hardware and software as a single entity.
The Art of Co - Design
The relatively new technique of designing hardware/software components together for an embedded system is called ‘Co-design’. Co-design focuses on the areas of system specification (co-specification techniques), architectural design and hardware/software partitioning (co-synthesis) and the interaction between the hardware and software as the design progresses (co-verification or co-simulation).
To achieve the goals of successful co-design, state-of-the-art, cost-effective tools are needed to ensure the quality, reliability, manageability and flexibility of embedded application design and development. One of the ways to capitalise on the use of
hardware/software co-design is by taking advantage of advances made by commercial software engineering tool vendors.
The Europian Union
In 1995, EUROPEAN UNION joined forces with the Microelectronics and Computer technology Corporation (MCC) in the USA to undertake a world-wide survey of current practice and the tools needed to achieve successful hardware/software co-design.
OMI, realising from the outset that building embedded systems from third party components has the greatest potential to improve the capability of European industry to meet market needs, has provided the framework in which the necessary tools and techniques can be developed in a cost-effective manner, and rapidly brought to market.
Over 350 European companies have participated in the OMI programme, amongst them some of Europe’s most famous names - Siemens & Daimler Benz from Germany, GEC Plessey & British Aerospace from UK, Thomson & Alcatel from France, Fiat & Olivetti from Italy. These companies have matched EU funding to the tune of around $250 million on projects that have led to the development of world-class products.
For close on five years, EU supported projects have been producing the hardware/software development tools the market wants. Tools that have been designed to work together, which can reduce time-to-market and allow for the re-use of code - three key areas of demand cited by all industrial sectors in many surveys. OMI projects have produced tools that enable the early detection of design flaws through the co-verification of hardware and software from the start of the design process. Tiny operating system software to manage the performance of the embedded application, as well as techniques to standardise the development process, which in turn leads to improved productivity.
These tools are not theoretical - doomed for ever to remain in the libraries of academia, nor are they prototypes - destined to gather dust on the shelves of the R&D lab. They are real products that have been developed specifically to address the needs of the embedded systems market.
For example, one project has evolved an holistic approach to the microelectronics systems development cycle, providing a complete environment from requirements analysis through design, development and testing to run-time. Another is focused on the cost-effective re-use of code, bringing the concept of object-oriented programming to the modelling of how an embedded system will behave in real-life. While the goal of a third project is to achieve the early detection of design errors by the integration of the hardware and software design process. This project focuses on the specific areas of the development cycle key to detecting early design flaws - namely design co-verification, complete (hardware/software) system modelling and simulation, monitoring and performance evaluation as well as design flow traceability. All the tools developed have been embedded into a common framework, allowing the exchange of design data and providing a consistent, easy-to-use interface.
The Way Forward
All the above co-design tools and techniques form only part of the huge effort being supported through OMI to strengthen European industry in all areas of advanced electronic processing and control. Tools that validate and verify the design process for hardware & software systems, helping to significantly improve the achievement of right-first-time-design; Tools which reduce the burden of development and time-to-market by enabling both re-use and improved testing capabilities. State-of-the-art tools which indeed provide European embedded systems designers with a competitive edge. Information on the work, the results and the participants in all the OMI projects can be found at the OMI Web site.
The ever increasing demand for greater complexity and functionality from embedded software will mean that more and more use is made of advanced third-party software to minimise development, testing time and cost of these more powerful applications. Although most embedded software is still developed entirely in house, the International Data Corporation (IDC) estimate that the use of third-party software in the embedded world will continue to grow year on year by around 25% - with an estimate for 1998 of $1 billion. As the cost of the software per chip continues to rise - a major industry concern, especially in the volume market - and in many cases now exceeds the actual cost of the silicon, the need for high quality, third-party co-design tools will become paramount.The European Union R&D programmes run in four-year cycles. The next - known as the Fifth Framework - is scheduled to start at the beginning of 1999. Co-design technologies will continue to be at the forefront of theprogramme - providing European industry with a leading edge to compete effectively in tomorrows world.
Related Online Articles:
No comment yet. Be the first to post a comment.