|Title:||Synchronizing IEEE 1588 clocks under the presence of significant stochastic network delays|
|Authors :||Iantosca, Carmelo|
|Conference details:||2005 Conference on IEEE 1588: Precise Networked Clock Synchronization WG, NISTIR 7302, Institut für Energiesysteme und Fluid-Engineering (IEFE), ZHAW, Winterthur, 10–12 Oktober 2005|
|License (according to publishing contract) :||Licence according to publishing contract|
|Type of review:||Not specified|
|Subjects :||Estimation; Synchronisation; Real-time|
|Subject (DDC) :||621.3: Electrical engineering and electronics|
|Abstract:||The clock synchronization method described in IEEE 1588 is often used in closed and application specific networks, e.g. for distributed measurement or automation and control applications. These networks can be equipped with special switches working as boundary or transparent clocks. In non IEEE 1588 aware switched networks such as Intranets, Metro Ethernets, MPLS or EFM Networks, the method has to cope with uncorrected stochastic message transit delays. We investigate the problem of synchronizing clocks in the presence of stochastic delays of PTP messages caused by queuing in switches. Especially in the case of heavy network traffic, queues may be filled and can lead to substantial delay of PTP messages. The aim of the presented approach is to ensure synchronization even with such large perturbations. Measurements in a simple laboratory switched fast Ethernet network showed that transit delays in the order of magnitude of some milliseconds may arise under heavy traffic. A simple synchronization algorithm fails under these circumstances. A synchronization algorithm was developed that is based on a generic statistical approach and explicitly takes into account the stochastic properties of the transit delay of PTP messages both in the case of delay and in the case of unhindered message flow. By using robust statistical estimation methods it is possible to estimate the drift and the time offset between master and slave reliably, even under strong stochastic perturbations. With this approach, time synchronization in already existing networks can be improved significantly.|
|Departement:||School of Engineering|
|Organisational Unit:||Institute of Data Analysis and Process Design (IDP)|
|Publication type:||Conference Other|
|Appears in Collections:||Publikationen School of Engineering|
Files in This Item:
There are no files associated with this item.
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.