|
Europe has always been sensitive to issues touching upon
system level design given the structure of the European
electronic industry that is much heavier on the system side
than on the semiconductor side. To quote some of the
European initiatives, we looked at the "Technology Roadmap
on Software-Intensive Systems: The Vision of ITEA”. The
major challenges for System Engineering, Software
Engineering and Engineering Process Support noted by ITEA
include:
 |
Increase Reuse
|
|
|
Hardware/Software
co-design
|
|
|
Modeling non-functional
properties
|
|
|
Move from products to
services via Software Components
|
|
|
System and SW
architecture
|
|
|
Validation and
verification at the system level
|
|
|
Adaptation of HW and SW
via re-configurable architectures and
component plug and play
|
|
|
Composable SW systems
using reusable SW components, integration
and certification of components
|
|
|
Support of parallel
development via integration technology
|
|
|
Development of common
workflow and process standards in semantics.
|
ITEA also looked at possible evolution trends, and
underlined:
Formal verification, co-design flows, reuse and
integration of components, (this was highlighted as
a major need), platform/function co-design, unified
HW/SW flows, and domain/architecture co-design. In
particular, "Component reuse standards to allow
intellectual property design to be introduced into
the embedded software design world", a quiet and
subtle nod to the fact that in HW, IP design and
reuse is more common than in embedded SW. Several
key issues in efficient component reuse are
configurability (to allow optimized reuse),
decomposition and modularity, and carefully crafted
interfaces.
Our
vision for ESW is to change radically the way in which ESW
is developed today by:
Linking ESW up-wards in the abstraction layers to system
functionality;
Linking ESW to the programmable platforms that support it
thus providing the much needed means to verify whether the
constraints posed on ES are met.
We
already reasoned about some of the concerns that are
typical of ESW that are not commonly understood in
the software community. In addition, we stress that
the research agenda has to address these very issues
if we wish to come out of the ESW quagmire. The
approach we advocate is holistic: it includes
methodology, supporting tools, IP’s, hardware and
software platforms, and supply chain management.
Only by taking a global, high-level view of the
problem, can we devise solutions that are going to
have a real impact on the design of embedded
systems. The essential issue to resolve is the link
between functionality and programmable platforms. We
need to start from a high-level abstraction of
system functionality that is completely
implementation independent and rests upon solid
theoretical foundations that will allow formal
analysis. We also need to select the platform that
can support the functionality meeting the physical
constraints placed on the final implementation. We
need to implement the functionality onto the
platform so that its properties of interest are
maintained and the physical constraints are indeed
met. The following work packages are intended to
address these issues. |
