TTP - A PROTOCOL FOR FAULT-TOLERANT REAL-TIME SYSTEMS

The communication protocol is an essential part of distributed real-time systems. To make it easier to implement a system under stipulated dependability and timeliness requirements, the functionality of this protocol should go beyond a message-exchange service. The authors introduce the Time-Triggered Protocol. an integrated communication protocol for time-triggered architectures. TTP provide the services required to implement a fault-tolerant real-time system: predictable message transmission, message acknowledgment in group communication, clock synchronization, membership, rapid mode change. and redundancy management. The time-triggered approach. where all system activities (such as task activations and message transmissions) are driven by the progression of global time, is preferred by the authors over the event-triggered approach because it provides better system predictability, testability, and analysability. TTP makes optimum use of the properties of a time-triggered system, especially with regard to the system's inherent availability of a priori information about its own future behavior. The protocol uses this information in conjunction with an innovative cyclic-redundancy-check calculation technique to ensure the required dependability and to enforce state agreement between the communication partners It supports different fault-tolerant system configurations with replicated nodes and replicated communication channels. TTP implements these services without extra messages and with only a small overhead in the message size. The authors compare TTP's performance with that of other protocols proposed for automotive real-time applications (such as Jl850 and the token protocol). They also discuss a VLSI TTP controller and its system integration.