glen@aecom.yu.edu (Glen M. Marianko) (10/25/89)
During most casual conversations about ethernet I've heard "you can
have a maximum of four repeaters between any two nodes on an
ethernet". I took this to mean that you can have five segments
max with nodes on each segment. I recently read a short blurb
that said something like some of the segments cannot have nodes
on them - the three in the middle, I guess it meant. Is this so?
Say it isn't ...
I also heard of something called a "link segment" or "inter-repeater link"
that someone quickly and poorly explained as a longer-than usual segment
between two repeaters w/o nodes on it. I guess with one of these
links you can exceed 185m thin and 500m thick?
What if I had a maxed out network - repeater-wise - and I just wanted
to make sure that the farthest workstations could collide with eachother
properly within the spec. Is there a device that I can put on the
beginning of segment 1 and the end of segment 5 that will test this?
--
-- Glen M. Marianko Manager, LAN Services Glasgal Communications, Inc.
151 Veterans Drive Northvale, New Jersey 07647 201-768-8082
glen@aecom.yu.edu - {uunet}!aecom!glen (Courtesy of AECOM & unaffiliated)paul@aucs.uucp (Paul Steele) (10/25/89)
glen@aecom.yu.edu (Glen M. Marianko) writes: >During most casual conversations about ethernet I've heard "you can >have a maximum of four repeaters between any two nodes on an >ethernet". I took this to mean that you can have five segments >max with nodes on each segment. I recently read a short blurb >that said something like some of the segments cannot have nodes >on them - the three in the middle, I guess it meant. Is this so? >Say it isn't ... In our ethernet segment, we have a maximum of four repeaters in a row and we also have nodes hung off of every repeater without problems. IN our situation, we use multiport repeaters at the each end of an optical star, which means in general we only have 2 repeaters between every node. However, in some buildings we have a Western Digital lattisnet repeater connected to one of the thin segments from a multiport repeater, thus giving us 4 repeaters in a row in some situations. We aren't really vvery happy with our design but it works. -- Paul H. Steele UUCP: {uunet|watmath|utai}!cs.dal.ca!aucs!Paul Acadia University BITNET: Paul@Acadia or PHS@Acadia (preferred) Wolfville, NS Internet: Paul@AcadiaU.CA CANADA B0P 1X0 (902) 542-2201x587
illgen@hq.af.mil (Keneth..Illgen) (10/25/89)
In article <2549@aecom.yu.edu> glen@aecom.yu.edu (Glen M. Marianko) writes: >What if I had a maxed out network - repeater-wise - and I just wanted >to make sure that the farthest workstations could collide with eachother >properly within the spec. Is there a device that I can put on the >beginning of segment 1 and the end of segment 5 that will test this? There are a couple of devices that come to mind. On the expensive side there is The Sniffer by Network General. You can have it max out your network and measure the results, in collisions, and also determine your problem areas. Additionally you can get a better feel for where you can place add nodes and how large you can let your network get. The other device that comes to mind is Cabletron's Node Emulator/LAN Specialist. The LAN Specialist does many of the same things as the Sniffer but is designed for the technician to lug around as a troubleshooting tool for the integrity of the cable plant. The Sniffer is more for the software types that are interested in system analysis and counting packets. I have much more information on these and other pieces of equipment. We use the Sniffer, the Node Emulator and the LAN Specialist here. I've been satisfied with all of them. I'm sure I only touched on one area of your inquiry but if you have other questions about test equipment I'd be happy to try to help. ken
ted@blia.BLI.COM (Ted Marshall) (10/27/89)
The four repeater rule only applies if you are using fiber optic or similar links between pairs of repeaters. I.e. Ethernet ------------------------------------------ | repeater O : fiber opt : O | ------------------------------------------ | O : : O | ------------------------------------------ These repeaters use to be called "half-repeaters". If you use normal "local" repeaters to directly connect together two Ethernet (thick or thin) segments, that counts as TWO out of the four allowed. Thus, two nodes cannot be more than three coax segments (500M each for thick) plus two fiber segments (1000M TOTAL) plus 6 transceiver cables (50M each) = 2800M apart. Like most of the standard configuration rules, you can break this one if you are conservative on other rules or have a light network load. if the segments are short, you may be able to add more repeaters. You cannot get the same effect of 1000M of fiber link with the same length of coax because the signal propigates slower through the coax. If you have five coax segments, 500M each, strung in sequence between 4 full repeaters, the problem that you will see is a higher incedence of collisions because it takes longer for a signal to travel from one end to the other. More importantly, you'll get collisions that were not detected by one of the transmitting stations because it finished transmitting before the colliding packet reached it. Thus, the packet is not retransmitted by the data link layer. The bottom line in this configuration is that it will work fine under a light load, but as the traffic goes up, your performance will go to hell! -- Ted Marshall ...!uunet!blia!ted <or> ted@blia.bli.com ShareBase Corp., 14600 Winchester Blvd, Los Gatos, Ca 95030 (408)378-7000 The opinions expressed above are those of the poster and not his employer.
goodloe@b11.ingr.com (Tony Goodloe) (10/30/89)
In article <9850@blia.BLI.COM>, ted@blia.BLI.COM (Ted Marshall) writes: > of coax because the signal propigates slower through the coax. From the standard: "The propogation velocity of the coaxial cable is assumed to be 0.77c minimum ..." and "A link segment .. end-to-end propogation delay of 2570 ns and shall be terminated in a repeater set at each end." I thought the prop velocity through fiber was ~.65c, but I can't find it anywhere. Can comebody help? tony