TTEthernet Startup and Synchronization Agreement. With this implementation of TTEthernet being constrained to only asymmetric benign faults, relatively simple startup and synchronization protocols can be used with high confidence. Startup consists of any one of a designated class of nodes sending out a unique coldstart frame. Given that switches can only be benignly faulty, this coldstart frame cannot be successfully forged in a way that can adversely affect the protocol. All good end systems that hear a coldstart frame acknowledge that frame and the time line that enforces the time-triggered traffic class begins. During execution of the TTEthernet timeline, periodic synchronization events are scheduled. In each synchronization event, Synchronization Masters transmit protocol control frames at the time that each of these end systems thinks it is correct to do so. The self-checking pair switches contain the Compression Master function that acts as a majority voter for the synchronization frames. Given that faulty switches are only asymmetrically benign, the only failure behavior they exhibit is a failure to deliver the synchronization frames to all end systems. That is, there are no possible failures affecting integrity, only availability. The required availability is provided by the use of multiple switches. Each end system is directly connected to a number of switches that is sufficient to meet the system’s availability requirements. The TTEthernet standard [1] establishes the algorithms and the quorum criteria required to meet the system fault tolerance requirements for the synchronization protocol.
Appears in 5 contracts
Samples: ntrs.nasa.gov, ntrs.nasa.gov, ntrs.nasa.gov
