Arizona State Pushes Embedded Education -

Arizona State Pushes Embedded Education

Have you noted the recent formation of a new “embedded communications” consortium between Arizona State University in Tempe and the nearby networking and embedded hardware powerhouses Intel Corp. and Motorola Computer Group?

The consortium comes at a time when a number of serious problems are emerging which could hamstring company efforts to move into markets with significant new growth opportunities, not only in traditional embedded applications in telecom and networking, industrial and building automation, but in the greatly expanded consumer space with its variety of network-connected embedded devices. . The most serious and chronic one is the shortage of academically trained professionals which leaves the industry dependent on a hard core of aging engineers and software developers who developed their skills as embedded has grown and evolved over the past 30 years.

Because embedded in the past did not have the kind of visibility in the marketplace that desktop computing has had, companies in this segment have always had to depend on the ability of its professionals to re-educate themselves in new techniques and tools as the technology and the market has changed. Common complaints I have heard consistently over the years include: “Not enough new people are coming in to replace the ones leaving.” “Not trained in the skills I need.” “They are good at programming, but have know idea of how to design a system.” “They can build what I want, but they don't know how to debug it or test it.” “I have to invest in a lot of on-the-job training and education of the people I hire.”

With the stagnation of the desktop computing market and the emergence of embedded as the new centerpiece of mainstream computing, companies and the programmers and developers there have tried to shift their expertise to the embedded space. Potential employers now have plenty of candidates, but only a few have the skills they require. And, I am told, virtually none have the deep understanding of the underlying hardware or the application environment in which they are expected to program.

Numerous educational efforts are underway within corporations and among entrepreneurs who are looked to provide appropriate educational expertise through trade schools, on-line “embedded universities,” conferences such as the Embedded Systems Conferences, and numerous specialized seminars targeted at telecomm, networking, embedded internet devices and so on

With net-centric connectivity the problems are getting worse. While the skills developed for the earlier embedded environment are still important, they must be applied to a “problem space” that is no longer closed. That is, they must be applied in a open environment in which reliability depends not just on the embedded device, but on everything to which an embedded processor or controller is connected.

But as far as I know, there has been no all-encompassing effort to address the broader educational and research needs of this new environment where communications and embedded computing collide..

However, this is the focus of the new consortium formed recently by ASU's College of Engineering and Applied Sciences, Intel and Motorola. According to Ben Huey, associate dean of planning for the College of Engineering and Applied Sciences, the Consortium for Embedded and Internet Working Technologies will fuel Intel's and Motorola's goal of attracting students from ASU who are trained to create the next generation of embedded communications systems.

“Together we've crafted a collaboration designed to address a national shortage of engineers with embedded systems expertise necessary to develop the technologies for embedded systems in general, and those required for internetworking and communications systems in particular,” said Huey. Initially, as much effort will be put into building an appropriate infrastructure of faculty, researchers, financial programs and teachers to address the problems

Because they view the problem as industry-wide, the participants hope that the consortium can be expanded beyond the current participants. “We are already looking to broaden participation, first from companies and educational institutions locally, and then more broadly,” said Shlomo Pri-tal, director of international business development at the Motorola Computer Group, who is the company's representative to the consortium.

The program will include a number of elements, beginning with a focused embedded computing systems curriculum being introduced within ASU's College of Engineering. A key component of the curriculum is a new, credit-bearing internship program in which students and faculty members will work alongside industry on projects that build the products and technologies for the next generation. “There have always been internships and fellowships and financial aid programs,” said Huey. “The difference here is that rather than limit such focused aid to seniors and graduate students we are going to reach down deep into the undergraduate level to sophomore or junior level students.”

According to Vince Liggett, chief scientist at the Motorola Computer Group, the reason for this is simple: “The need has been so critical in the areas of embedded software development that by the junior year many college students with experience in program development in Java and C are already being seduced by salaries that they did not expect to get for years after they graduated.”

Intel and Motorola faced many of the same problems when they managed to hire such nascent and still-developing talent: experience limited to one language; no broad education in related topics such as testing and debug; programming experience limited to just one platform or API–usually the desk top; and no understanding of CPUs other than the X86 or operating systems other than Windows. “If we were really lucky we got someone with a beginner's knowledge of Unix, which meant that at least they would be able to transfer that to Linux or to one of the POSIX-compliant or -compatible real-time operating systems,” said Jeff Harris, director of research and development at Motorola Computer Group.

The introduction of ubiquitous connectivity into the equation has only made the problem more severe. “Now we not only need people with training in all of the traditional disciplines that have made embedded what it is, but we also need to understand the impact of the convergence between communications and computing on all aspects of how we build, program, test, debug and maintain them,” said Pri-tal.

A number of questions, he said, have been identified by the consortium that need further investigation: (1) to what degree will students have to be trained in parallel programming to reflect developments at the device level, such as C extensions for circuit design, and at the systems level in the arrays of network processors in many telecom and networking applications? (2) How will hardware and software development tools have to be changed to work in network environments that are by nature non-deterministic and not always real-time? (3) What alternate processor architectures will emerge to satisfy the new requirements of the embedded net-centric computing environment, or can existing architectures be modified and adapted? And (4) how will operating systems have to change in this new computing environment?

“These are all questions that deserve intensive research in and of themselves,” said Huey. “But we also need the answers if we are going to stay ahead of the game educationally. The speed with which the computing and communications environment is changing is so fast that we cannot afford to just keep up with current requirements in industry, but must stay ahead of them to some degree.”

What training and manpower issues you are facing? How are you dealing with them? Where is the problem most acute: in software or hardware development? What programs are underway at your company or educational institution?

Reader Feedback:

Nowadays embedded systems are flooding into all fields of engineering, say, industrial process control,robotics, automotive, medical, Avionics, Defence, telecom.etc…

So it is very difficult to train the students in particular area. Automotive and avionics industry requires people from automobile, automotive electronics, mechanical, control systems, aeronautical eng, etc.

It is better to (set up)have a master degree(MS/MTech) in Real-Time Embedded systems for all engineers with any stream, and in a masters degree we should have close interaction with students and developers in this field. This will solve the problem to some extent.

Mahantappa M Patil
Software Engineer
Rasquares Information Systems, Cimbatore

I am excited about the idea. I've been in the industry for two years working in embedded and networking projects ,most of the time spent in testing.Even today, I am not in aposition to design an embedded system.So,obviously an academic curriculum in embedded systems will help both the people and the industry!


I strongly agree that there should be courses especially in EE (perhaps at junior or senior level) on designing embedded system, RTOS designand digital system design. Couple with C or Java course from CS dept, this option should be good enough for an interested student to be a good embedded system designer.

Shamsul Bahry
Sr. Software Engineer
Motorola Multimedia

I am excited about the idea. I've been in the industry for two years working in embedded and networking projects ,most of the time spent in testing.Even today, I am not in aposition to design an embedded system.So,obviously an academic curriculum in embedded systems will help both the people and the industry!


I suspect that one of the problems is the lack of educators with current embedded experience. When I went to college 14 years ago those educators got their experience from NASA project grants in the 70s.

It must be difficult to keep someone at a university for $70k a year when they could be making $100k in industry.

Personally I might consider a teaching position as an alternative to the “excitement” of industry if I could support my family.

James Munn
Sr. Software Engineer

It's tough for people interested in embedded programming tofind decent training. It's a 'do it yourself' situation.

I am working on a Masters degree in Computer Science. We have nocourses dealing with embedded programming. It appearsyou have to be an EE major to get any training.

Any thoughts?

Donald Goodwin

I could not agree more … in fact the problem is worse than you think. As both a continuing education instructor and a RTOS sales manager

1) even embedded RTOS houses do not train their sales force in the fundamentals of SDK's and tool chains etc.


2) colleges feel the topic is too technical for the level of students they attract and leave the “problem” to universities.

Peter Wheeler
Sales Manager

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.