A self-reconfiguring IEEE 1687 network for fault monitoring

Abstract

Abstract-Efficient handling of faults during operation is highly dependent on the interval (latency) from the time embedded instruments detect errors to the time when the fault manager localizes the errors. In this paper, we propose a self-reconfiguring IEEE 1687 network in which all instruments that have detected errors are automatically included in the scan path. To enable self-reconfiguration, we propose a modified segment insertion bit (SIB) compliant to IEEE 1687. We provide time analyses on error detection and fault localization for single and multiple faults, and we suggest how the self-reconfiguring IEEE 1687 network should be designed such that time for error detection and fault localization is kept low and deterministic. For validation, we implemented and performed post-layout simulations for one self-reconfiguring network. We show that compared to previous schemes, our proposed network significantly reduces the fault localization time.

Test standards reuse for structured and cost-efficient dependability management of system-on-chips

“…A tree-based structural model of SIB-based hierarchical networks is proposed in [Jut13], [Zad16b] and [Zad11b], where each node in the tree corresponds to a single SIB. In [Tše16] another node type was added for registers in order to allow for register placement in the hierarchical levels.…”

“…indivisible) and not as a pair of ScanMux and ScanRegister. Consequently, constructing an efficient hierarchical organization of the network could be carried out for optimizing the instruments' access time [Zad11b] or for optimizing the average localization time of fault detecting instruments by balancing the network hierarchy for instance by using binary or ternary trees [Jut13] [Zad16b].…”

“…IJTAG networks were also considered by recent works in performing dependability operations, however, the focus was only on connecting fault detecting instruments and performing system-level fault diagnosis [Jut11], [Jut13], [Shi14], [Jut16], [Shi16], [Pet14], [Zad16b] and [Zad17].…”

“…In [Zad16b] and [Zad17] the concept of a self reconfiguring network was introduced as a solution for fast fault localization in hierarchical IJTAG networks. Whenever an error is produced as a result of a certain fault, and is subsequently detected by an error-detecting instrument, the network is automatically configured to include a specific instrument register that holds a produced error code.…”

“…rollback and re-execute [Shi16] or core isolation). Self reconfiguring networks [Zad16b] and [Zad17] address this by providing a fast localization and fault diagnosis methodology for hierarchical IJTAG networks while maintaining the modularity of the network which is essential for its scalability.…”