Fault-detection, retransmission and fault-tolerant router unit based on triple modular redundancy

Abstract

A fault-detection, retransmission and fault-tolerant router unit based on triple modular redundancy comprises an arbitration unit, a switch, a decoding module, four port input units and a local input unit. The arbitration unit performs output sequence control on input data packets and acquires a control signal for outputting to the switch. The switch successively outputs the input data packets according to the control signal. The decoding module performs decoding verification on data which arrive at a target node. Each port input unit is used for receiving input data packets in one direction, performing error verification, calculating a routing path and outputting the data packets according to a link state and a retransmission request which are fed back by a next-grade router. The local input unit receives data which are input through a local port, performing packing processing, performing error verification, calculating the routing path, and outputting the data packets according to the link state and the retransmission request of the target node. The fault-detection, retransmission and fault-tolerant router unit has advantages of improving transmission reliability of the data packets, and satisfying performance requirements of transmission time delay, area, power consumption, etc.

Description

technical field [0001] The invention relates to a routing unit, in particular to an error-detection-retransmission fault-tolerant routing unit based on triple-mode redundancy, and belongs to the field of integrated circuit multi-core design. Background technique [0002] With the development of integrated circuit manufacturing technology, SoC (System-on-Chip) based on traditional bus architecture has serious problems in scalability and bandwidth limitations, so NoC (Network-on-Chip) was proposed. Among them, the routing unit is an important part of NoC and the key to data transmission. However, as the process size decreases, the data transmission circuit in the routing unit becomes more and more sensitive to interference, which eventually leads to transient faults gradually becoming the main source of data transmission errors. Therefore, adopting a reasonable fault-tolerant algorithm to design a fault-tolerant routing unit has become a new research focus and difficulty. ...

AI Technical Summary

Problems solved by technology

Traditional fault-tolerant routing units generally use end-to-end error detection and retransmission or point-to-point error detection and retransmission to control the error of data packets. This method cannot take into account performance requirements such as transmission delay, area, and power consumption.

In addition, in applications where faults are more serious and error rates are higher, Baotou microchips are not specially protected, which further leads to performance degradation and even higher probability of erroneous transmissions

Images