| 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 |
BEFORE:
Our site LAN was broken up into (3) thickwire segments, corresponding to
quadrants of our building (1st floor South, 2nd floor South, and 2nd floor
North), each individually bridged to a central segment (the Data Center)
with LAN bridge 200s.
Offices... Offices...
| |
| |
DEMPRs... DEMPRs...
| |
| |
DELNI DELNI
| |
| |
------------------------ -----------------------------
1st floor South | 2nd floor South |
| |
LANbridge 200 LANbridge 200
| |
| |
------------------------------------------------------------------
Data Center |
|
LANbridge 200
|
|
-----------------------------
| 2nd floor North
|
DELNI
|
|
DEMPRs...
|
|
Offices...
AFTER:
Now our site has an FDDI dual-ring between the Data Center, and the (3)
SERs where the thinwire runs are concentrated for each quadrant of the
building.
FDDI dual-ring
################DECbridge 620###########################DECbridge 620###########
# | | #
# | | #
# Offices... | Offices... | #
# | | | | #
# | | | | #
# DEMPRs... | DEMPRs... | #
# | | | | #
# | | | | #
# DELNI-----------' DELNI-------' #
# | | #
# | | #
# ------------------------ ----------------------------- #
# 1st floor South | 2nd floor South | #
# | | #
# LANbridge 200 LANbridge 200 #
# | | #
# | | #
# ------------------------------------------------------------------ #
# Data Center | | #
# | | #
# | LANbridge 200 #
# | | #
# | | #
# | ----------------------------- #
# | | 2nd floor North #
# | | #
# | DELNI-------. #
# | | | #
# | | | #
# | DEMPRs... | #
# | | | #
# | | | #
# | Offices... | #
# | | #
# | | #
################DECbridge 620###########################DECbridge 620###########
FDDI dual-ring
COMMENTS:
Perceived (i.e. "seat-of-the-pants") network performance has gone down,
substantially. I don't have quantitative measurements of "before and
after" (except for collision rates), but desktop users are *definitely*
complaining of network "hangs". I have access to a Network General
SNIFFER, and have observed a collision rate in the wiring closets at times
exceeding 20% via FDDI.
QUESTION(s):
Have I *fundamentally* mis-configured something here? Shouldn't I expect
*at least* equal throughput between the FDDI and NI? Can the FDDI and NI
connections co-exist since they are all bridged (i.e. won't the spanning
tree algorithim and port costs take care of appropriate and correct packet
routing)?
I haven't "tweaked" anything, and I have RTFM'ed all about the
FDDI-giblets along with the FDDI configuration guidelines, but I am still
fresh out of clues. Suggestions, recommendations are extremely welcome.
Thanks in advance.
| T.R | Title | User | Personal Name | Date | Lines |
|---|---|---|---|---|---|
| 1002.1 | A few things to check | BRAT::BUKOWSKI | Fri Jul 02 1993 13:40 | 15 | |
Check for CRC errors with the analyzer or ltm. Are you transceivers
that support the dempr/delni combination set with heartbeat disabled?
The spanning tree shouldn't be a problem.
Check phy port 1 of you 620's for errors, especially "Estimated Link Error
Rate". If it is much lower than 15, you could be looking at some problems
with the fiber between the equipment (hi db loss from dirty or poorly
polished fiber ends).
BTW: do you know which bridges/lines are in backup? Are you truely
using the FDDI ring as the backbone?
Mike
| |||||
| 1002.2 | Everything checks out so far... | DECWET::FUJII | Carey J. Fujii, Redmond, WA | Fri Jul 02 1993 14:39 | 22 |
> Check for CRC errors with the analyzer or ltm. Are you transceivers
> that support the dempr/delni combination set with heartbeat disabled?
CRC error rate is averaging around 0.004%. The transceivers have long since
had heartbeat disabled.
> Check phy port 1 of you 620's for errors, especially "Estimated Link Error
> Rate". If it is much lower than 15, you could be looking at some problems
> with the fiber between the equipment (hi db loss from dirty or poorly
> polished fiber ends).
I'm not sure how to do this? I'm using RBMS v2.0 to interrogate the bridges
and I didn't see a counter for "Estimated Link Error Rate". The vendor that
installed the fiber did test each strand for acceptable db loss. However, I
realize the "proof is in the pudding", so I'll check this out once I know how.
> BTW: do you know which bridges/lines are in backup? Are you truely
> using the FDDI ring as the backbone?
Yep. The ports on then LANbridge 200s show up in "BACKUP" mode once the
DECbridge 620s are hooked up.
| |||||
| 1002.3 | MSU/ELMs gave me some answers, but I still have questions... | DECWET::FUJII | Carey J. Fujii, Redmond, WA | Fri Jul 09 1993 21:40 | 9 |
> Check phy port 1 of you 620's for errors, especially "Estimated Link Error > Rate". If it is much lower than 15, you could be looking at some problems > with the fiber between the equipment (hi db loss from dirty or poorly > polished fiber ends). Caught a clue. I use MSU, and used ELMs to take a look at the PHYPORTs. Link error rate on all bridges is right at 15. Other ideas? Now that I have ELMs, I am continuing my investigation a little more indepth. Stay tuned for more news. | |||||
| 1002.4 | Tedious troubleshooting tips | SCHOOL::RLEBLANC | Mon Jul 12 1993 12:09 | 27 | |
If you haven't solved it yet here are some other tedious checking
tips:
o What is the length of your Ethernet Segments both Thick
and Thin in the new configuration ?
o Check the 620 bridges if not all of the Ethernet ports are
in use make sure they have Ethernet Loopback connecters on
those ports?
o Check all your BNC connectors and make sure all the
ends are properly crimped at both ends?
o Of all the people complaining are most of them located in
the same wiring closet by chance? Start there if true.
o Are you running latest firmware on all the FDDI gear?
o Did you recently have a power failure? Maybe you need to
reset the dempers or have some flakey hardware somewhere.
o Check all bridge counters see if one of the bridges is
counting error counters on a magnitude greater than
others?
Rene'
| |||||