Real-time communication in FDDI-based networks

Due to the character of the original source materials and the nature of batch digitization, quality control issues may be present in this document. Please report any quality issues you encounter to digital@library.tamu.edu, referencing the URI of the item.Includes bibliographical references.Issued also on microfiche from Lange Micrographics.The Fiber Distributed Data Interface (FDDI) is recommended as a candidate for many hard real-time local-area communication systems. The Timed-Token Medium Access Control (MAC) Protocol of FDDI is capable of guaranteeing real-time traffic on an FDDI network with the synchronous transmission mode. This study addresses the issues in implementing real-time communications on an FDDI network using synchronous mode transmission, and guaranteeing the apphcation-to-application deadlines of multiple message streams. Extensive research has been done on using FDDI for real-time communication. Most previous research had three limitations. Delays and deadlines were only considered at the MAC level. Also, in order to simplify the analysis, only one message stream per node was considered. Furthermore, the analytical results could not be verified due to the commercial unavailability of FDDI products implementing synchronous mode transmission. This work removes these limitations and shows a systematic way for guaranteeing end-to-end message deadlines at the application level on an FDDI network. The concept of a Synchronous Server is introduced to eliminate the restriction of one message stream per node. The subsequent problems faced in bandwidth allocation and network access when multiple streams are allowed on one node are addressed. The design and implementation of such a server are discussed, and a methodology is proposed for estimating application-to-application mes sage delays and guaranteeing deadlines when using such a server. Certain constraints are proposed as necessary conditions for guaranteeing deadlines. It is claimed that deadlines can be guaranteed for synchronous message streams once these constraints are satisfied, i.e., these constraints form a set of sufficient conditions for guaranteeing deadlines. The implementation of the synchronous server allocates synchronous bandwidth to applications based on these constraints. Synchronous mode transmission is implemented to be able to verify the analytical results. Experimental results using this server are provided to support the claims made. Obtaining timing information on different machines also brings up the issue of clock synchronization, and this is also addressed