Sharif Digital Repository

Choreography description languages aim at specifying from a global point of view interactions among a set of services involved in a new system. From this specification, local implementations or peers can be automatically generated. Generation of peers that precisely implement the choreography specification is not always possible: This problem is known as realizability. When peers corresponding to this specification are being executed we may want to modify the choreography specification and reconfigure dynamically the system. This is the case for instance if we add or remove interactions due to the addition of functionalities to the system at hand or the loss of a service. In this article, we... 

With the increasing use of pervasive computing environments (PCEs), developing dynamic reconfigurable software in such environments becomes an important issue. The ability to change software components in running systems has advantages such as building adaptive, long-life, and self-reconfigurable software as well as increasing invisibility in PCEs. As dynamic reconfiguration is performed in error-prone wireless mobile systems frequently, it can threaten system safety. Therefore, a mechanism to support assured dynamic reconfiguration at run-time for PCEs is required. In this paper, we propose an Assured Dynamic Reconfiguration Framework (ADRF) with emphasis on assurance analysis. The... 

In order to be adapted to changes in user requirements and/or the environment, many software systems need to run continuously while they evolve. Most current approaches for such dynamic reconfiguration assume that the evolved system will behave as expected and thus will be reliable if the reconfiguration is consistent. This assumption may not correspond to reality because the delivered quality estimated previously could vary due to parameter changes at runtime. To ensure that the system acts correctly in the field after the reconfiguration, reliability of changes has to be checked at runtime. Existing approaches, however, are not applicable in highly available systems due to the possibility... 

Forthcoming networked systems require mechanisms for on-the-fiy reconfiguration in their protocol stacks to be able to operate in different situations and networks. Since every protocol in a protocol stack has at least one peer protocol in another system, dynamic reconfiguration of a protocol raises the need for reconfiguration in the peer stack. For an assured dynamic (run-time) reconfiguration, executions of two peer protocols are stopped in a safe state, new protocols are initialized, and stacks switch to the new protocols at the same time. This paper proposes a software framework for dynamic reconfiguration of two communicating protocol stacks. A distributed algorithm is implemented in... 

Growing the need for long-life and high-available programs, dynamic reconfiguration is going to be an important research issue. Dynamic reconfiguration enables the software systems to change at runtime to decrease their down-time in case of any updating, upgrading or in any changes. Any invalid reconfiguration at runtime may lead programs into invalid states. In this paper, we investigate on validity of dynamic reconfiguration for component-based programs and propose validity conditions for it. We show that the problem of validity assurance in general is undecidable and there is no general-purpose algorithm to verify dynamic reconfiguration validity. To have a computable algorithm for... 

High availability of software systems is an essential requirement for pervasive computing environments. In such systems self-adaptation, using dynamic reconfiguration is also a key feature. However, dynamic reconfiguration potentially decreases the system availability by making parts of the system temporary frozen, especially during incomplete or faulty execution of the reconfiguration process. In this paper, we propose Assured Dynamic Reconfiguration Framework (ADRF), consisting of run-time analysis phases, assuring the desired correctness and completeness of dynamic reconfiguration process. We also specify factors that affect availability of reconfigurable software in pervasive computing... 

We propose a new optical reconfigurable Network-on-Chip (NoC), named ReFaT ONoC (Reconfigurable Flat and Tree Optical NoC). ReFaT is a dynamically reconfigurable architecture, which customizes the topology and routing paths based on the application characteristics. ReFaT, as an all-optical NoC, routes optical packets based on their wavelengths. For this purpose, we propose a novel architecture for the optical switch, which eliminates the need for optical resource reservation, and thus avoids the corresponding latency and area overheads. As a key idea for dynamic reconfiguration, each application is mapped to a specific set of wavelengths and utilizes its dedicated routing algorithm. We...