Carlos Almeida
Real-time group communications, QoS adaptability, quasi-synchronous systems, fault-tolerance
Most distributed communication infrastructures are not fully synchronous. They are at most quasi-synchronous: only a small part of the system can be considered as synchronous; the rest presents a more dynamic behaviour where the duration of a given activity (processing or communication) is usually relatively small, but there is a non-null probability of taking a much higher value occasionally. In these scenarios, it is very difficult (or even impossible) to always offer the desired timeliness properties. However, some applications need more than just a best-effort policy. They need at least the ability to adapt the offered quality of service (QoS) in a timely and consistent way. In the quasi-synchronous approach, the small synchronous part is used to build components able to control and validate the other parts of the system. This approach does not solve all timeliness problems per se, but can be used by applications to reach a safe state before stopping, or switch in a controlled manner between several different QoS. In this paper we explain how this architecture can be used to build communication protocols that are able to dynamically adapt their offered QoS in a timely and dependable way.
Important Links:
Go Back
