[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference 7.286::fddi

Title:	FDDI - The Next Generation

Moderator:	NETCAD::STEFANI

Created:	Thu Apr 27 1989
Last Modified:	Thu Jun 05 1997
Last Successful Update:	Fri Jun 06 1997
Number of topics:	2259
Total number of notes:	8590

251.0. "Where does the token actually go?" by ZPOVC::HWCHOY (Chicken on fire.) Mon May 06 1991 13:46

    Hi,
    
    I understand that maximum fiber length on an FDDI ring is 200km due to
    the need for the token to come round within an interval. In a tree
    topology, does the token actually traverse the entire tree (depth-first
    traversal?) before leaving the concentrator?
    
    Does that mean I have to add up ALL the fiber length between SAS and
    CONCs in the 200km? Howabout when I'm using SMF fibers? When I use SMF
    between a CONC and SAS, does that eat up the 200 km total?
    
    thanx for any explanation,
    
    heng-wah

T.R	Title	User	Personal Name	Date	Lines
251.1		KONING::KONING	Lietuva laisva!	`Tue May 07 1991 11:05`	12
	Yes, you have to count all the fiber between concentrators and SAS in doing the total -- with one exception. All that counts is ACTIVE fiber. When you use "dual homing", the standby connection does NOT count since it isn't included in the ring. In a dual homing configuration, count only the active connection to the concentrator. (I'm assuming the active and standby connection are the same length. If not, count whichever is longer.) Yes, single mode fiber has to be counted just as much as multimode fiber. paul
251.2		ZPOVC::HWCHOY	Chicken on fire.	`Tue May 07 1991 14:01`	6
	However, the fiber on the ring needs counting twice (as it's the dual ring), and the CONC-to-SAS is only once. This is the right reason to keep the main ring small and use SMF to extend from a CONC to a remote CONC or SAS? thanx again
251.3	cascading tree is depth-traversed?	ZPOVC::HWCHOY	Chicken on fire.	`Tue May 07 1991 14:04`	3
	just to confirm, in a cascading tree of CONC, a frame entering the "root" CONC will depth-traverse the whole tree, before emerging from the "root" again?
251.4		KONING::KONING	Lietuva laisva!	`Tue May 07 1991 18:40`	10
	Yes, depth first. ("Preorder traversal" is what a computer scientist would call it.) The reason the dual ring is counted twice is that it might wrap, which effectively doubles the distance. That's one reason for recommending a small dual ring. Another reason is that the dual ring only handles SINGLE faults: if more than one break occurs, you have a partitioned network. The tree can handle multiple faults. paul