HEX: scaling honeycombs is easier than scaling clock trees (Q269518)

From MaRDI portal
scientific article
Language Label Description Also known as
English
HEX: scaling honeycombs is easier than scaling clock trees
scientific article

    Statements

    HEX: scaling honeycombs is easier than scaling clock trees (English)
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    18 April 2016
    0 references
    In the paper, the authors show that a hexagonal grid can be a viable alternative to buffered clock trees that are typically used in applications and systems requiring accurate time synchronization. Precise clock synchronization is a critical issue in many environments and systems, including VLSI circuits, Globally Asynchronous Locally Synchronized (GALS) architectures, multi-core processors, parallel computing, and communication network hardware devices. It is also required in many new applications, among others, telecommunication protocols, power plants, industrial automation, testing and measurement systems, robotics and automatic control solutions, etc. Accurate clock synchronization is very important for planning, managing, securing and debugging in all the above systems, where precise instants of event occurrence and their execution times need to be known. The paper is well organized, and the presented results are interesting and encouraging. The proposal to use a hexagonal (HEX) grid with simple intermediate nodes is very interesting and promising. This method offers self-stabilization and guarantees Byzantine fault-tolerance, and additionally supports multiple synchronized clock schemes. A theoretical worst-case analysis of a new approach, proposed in the paper, reveals an acceptably small clock (timing) skew. The results obtained via simulation fully confirm this finding. The proposed approach can be applied in the design of network time protocols, VLSI circuits and multi-core processor systems.
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    clock synchronization
    0 references
    clock distribution
    0 references
    Byzantine fault-tolerance
    0 references
    self-stabilization
    0 references
    VLSI circuits
    0 references
    computer systems
    0 references
    GALS architectures
    0 references
    dependable computing simulations
    0 references
    skew analysis
    0 references
    0 references