Reconfigurability has long been recognized as a way to improve efficiency in the use of system resources, for example, when a system undergoes variable load situations
, when it evolves during its lifetime
or even when faults affect part of its structure
. This means that reconfigurability, in a broad sense, may be beneficial to areas that range from Quality of Service (QoS), e.g., when the number of system users varies, to Dependability, e.g., through graceful degradation.
However, achieving reconfigurability may conflict with several other operational goals such as continued real-time and safe operation, and it becomes more difficult when the system is distributed, requiring adequate support from the network. Therefore, whenever either of those two operational goals were relevant, the typical option has been to rely on a single static configuration. In some cases, reconfigurable solutions have been devised but limited to a few predefined operational modes, thus still with reduced flexibility and efficiency.
Therefore, more flexible approaches to reconfigurability are needed to improve resource efficiency
, exploiting paradigms such as flexible modes, flexible scheduling, dynamic QoS management, stateful schedules
, etc, particularly at the network level. This opens the way to keep costs low as the overall system complexity increases.
Conversely, other domains exist, such as wireless sensor networks, mobile ad-hoc networks and ubiquitous computing, in which reconfigurations are unavoidable thus the problem is how to support real-time and dependable operation under changing topology, asymmetric links and openness to interference.
The goal of this workshop was to bring together people from industry and academia to:
discuss the motivations, interest and challenges
of reconfigurability in distributed real-time embedded systems;
deduce the network requirements
to support flexible reconfigurability under real-time and safe operation;
discuss the adequacy of existing protocols and middlewares
discuss how to provide real-time communication in highly flexible networks
and identify the potential of current protocols;
deduce further network requirements to support real-time communication in highly flexible networks
A report of the workshop is under preparation and will be made available, shortly, in this site.