Invited Speakers


Three eminent keynote speakers have been selected to open each day of the core conference program:

Theodore Baker (Florida State University, USA), a leading researcher in Ada and Real-Time systems, will examine the state of the art in multiprocessor real-time scheduling in his talk What to Make of Multicore Processors for Reliable Real-Time Systems?

Pedro Albertos(Universidad Politécnica de Valencia, Spain), a most authoritative member of the Automatic Control community, will explore the relationship between implementation and performance of control algorithms in a talk entitled Control Co-Design: Algorithms and their Implementation.

James Sutton (Lockheed Martin, USA), a worldwide expert in software systems architecture, will explore how Ada is prepared for the so-called 3.0 World, a world that makes peace with complexity and chaos, and learns to use them to its advantage. That will be in his talk entitled: Ada: Made for the 3.0 World.



What to Make of Multicore Processors for Reliable Real-Time Systems?

Pedro Albertos

Theodore P. Baker,
Florida State University, USA
Tuesday 15 June

 

Now that multi-core microprocessors have become a commodity, it is natural to think about employing them in all kinds of computing, including embedded real-time systems. Appealing aspects of this development include the ability to process more instructions per second than is possible with a single processor, and execute more instructions per Watt than with a single fast processor capable of the same net processing speed. However, making effective use of a multi-core processor is not simple. Not all problems are amenable to parallel decomposition, and for those that are, designing a correct scalable multi-threaded solution can be difficult. If one is also
required to guarantee deadlines will be met, the difficulty becomes much greater.
This talk will review some of what is known about multiprocessor scheduling of task systems with deadlines, including recent advances in the analysis of arbitrary sporadic task systems under fixed-priority and earliestdeadline first scheduling polices. It will also examine critically the foundations of these theoretical results, including assumptions about task independence and worst-case execution time estimates, with a view toward their practical applicability
.

Presenter
Ted Baker is a professor at the Florida State University. Starting in 1979, he led a group that produced one of the first validated Ada cross-compilers for an embedded microprocessor system. Ted was active in the Ada Runtime Environment Working group and the workshops that led to the Ravenscar profile. He served as domain expert for real-time systems and systems programming on the Ada 95 Mapping-Revision Team, and later led the FSU team that developed the first Ada 95 multi-tasking implementation, for GNAT. He chaired the POSIX Ada language binding working group, and led the development of Florist, the first implementation of POSIX 1003.5b-c, including the FSU Pthreads library.
Concurrent with his work on Ada, Professor Baker has done research on real-time scheduling, worst-case execution time prediction, and real-time concurrency control. His contributions include the Stack Resource Protocol and the Deadline Sporadic Server. Ted's recent work has focused on incorporating device drivers into real-time schedulability analysis, and on verifying the ability of task systems to meet deadlines under various scheduling policies on multiprocessor platforms.


 

Control Co-Design: Algorithms and their Implementation

Ted Baker
Pedro Albertos
Universidad Politécnica de Valencia, Spain
Wednesday 16 June

Traditionally, control designers and real-time computer experts work separately. The former conceive the control algorithms based on the required performance and the process knowledge, regardless of their subsequent implementation. Computer experts instead deal with the control code without paying much attention to the impact of the code execution on the control performance. However both issues are strictly interlaced and both designs should be jointly treated, mainly if the control tasks share resources with some other activities and these resources are limited. In this talk, the real-time control design and implementation will be reviewed from both perspectives.
Global requirements in control applications in time-critical environments, such as automobile, aerospace or flight control, where multiple interactive control loops are implemented, are reviewed. Special attention is devoted to new and widespread control scenarios where the controller is no longer implemented in a fully dedicated computer without resources constraints, but sharing and competing for computing, storage and communication facilities with several other tasks. Embedded control systems, networked control systems and event-based control systems challenge the design of the control and its implementation where architectural issues play a relevant role in the controlled system performance. In this context, new software development models and middleware paradigms emerge, which strive to assure quality of service of control performance as well as computing, communications and power resources availability. The final goal of this technology is to allow the separation of complex control systems design from the real time tasks dynamic deployment. Some key concepts interacting with both control performance and control implementation, such as the control effort or the control kernel, are emphasized and some general directions in the co-design are summarized.

Presenter
Pedro Albertos, past president of IFAC (the International Federation of Automatic Control) in 1999-2002, and Senior Member of IEEE, is a world recognized expert in real-time control, leading several projects in the field. Full Professor since 1975, he is currently at the Systems Engineering and Control Department of UPV. He is Doctor Honoris Causa from Oulu University (Finland) and Bucharest Polytechnic (Rumania). Invited Professor in more than 20 Universities, he delivered seminars in more than 30 universities and research centers. Authored over 300 papers, book chapters and congress communications, co-editor of 7 books and co-author of “Multivariable Control Systems” (Springer 2004) and “Feedback and Control for Everyone” (Springer 2010). He is also associated editor of Control Engineering Practice and Automatica and Editor in Chief of the Spanish journal RIAI. His research interests include multivariable control and non-conventional sampling control systems, with focus on time delays and multirate sampling patterns, being involved in the ARTIST2 Network of Excellence on Embedded Control Systems.

 


Ada: Made for the 3.0 World

James Sutton James Sutton,
Lockheed Martin, USA
Thursday 17 June

Ada’s best days are ahead… if we, her friends, will embrace the world that is and let go of the world we loved for so long.
Ada was created 30 years ahead of her time. In the 1.0 World in which Ada was created people still believed you could attain certainty in systems development. The big challenge was to find the right corner to look around to find the certainty one just knew would be waiting there. Once the certainty was found, you would be able to exploit it to develop successful systems.
However, around every next corner was… nothing certain, and little even useful. Eventually people walked away; first from the search, and finally from the idea itself that certainty was attainable in systems development. But people still developed systems and had to do something; so, they leapt off a cliff and embraced uncertainty. Service-Oriented Architectures, social networking, mash-ups, and the like are 2.0 manifestations of this leap. We have not yet landed, though where we are headed is finally coming into focus in the mists ahead.
It is the 3.0 World. The 3.0 World makes peace with complexity and chaos, and learns to use them to its advantage. Principles, strategies, and statistics replace rules, procedures, and lists. Lean overturns Mass production and Systems Engineering displaces canned processes. The Cynefin Framework, a way of conceptualizing problem solving that suggests solution strategies for all kinds of situations from simple to chaotic, provides navigation beacons for the unknown. Most of these things are beginning to show up already in software development, and in this world Ada can finally and fully come into her own.
Early Ada advocates spoke to the 1.0 World in 1.0 terms; and appropriately so. “Ada is the best way to implement assured processes.” “Ada provides consistent object functionality regardless of compiler.” These Ada strengths appeal to 1.0 values, but they are not Ada’s only or even greatest virtues. In this talk we will explore the demands of uncertainty, and ways in which Ada’s greatest strengths address them. More importantly, we will lay out a path for how you the Ada community can make people understand that Ada is the most relevant language available today.
As you engage this 3.0 World and make it understand Ada in 3.0 terms, we will see the best days of Ada ahead.

Presenter
James Sutton’s passion is for unleashing the power and joy of human creativity in the development of systems. A current focus is helping preserve the middle class from the migration of well-paying white-collar engineering jobs to lower status and pay through the misapplication of antiquated and counterproductive business models like the Unit-Cost Equation.
James is a chief software-systems architect whose development programs have quadrupled productivity compared to company and industry norms, and at the same time experienced ten-times fewer defects than are typical in industry. His book “Lean Software Strategies” won the 2007 Shingo Prize, which Business Week has called “The Nobel Prize of Manufacturing.” He is an INCOSE (International Council on Systems Engineering) CSEP (Certified Systems Engineering Professional), with a Master’s degree in Systems Engineering from Southern Methodist University.
Most recently, in 2009 he joined with software luminaries such as David Anderson, Dean Leffingwell, Alan Shalloway and Don Reinertsen in co-founding the Lean Software and Systems Consortium. The LeanSSC is dedicated to facilitating the adoption of 3.0-World approaches and mindset in organizations that develop significant software-intensive systems.