From: Jonathan V Hays (jhays@jtan.com)
Date: Thu Sep 25 2003 - 12:01:39 GMT-3
To review, full-duplex mode in Ethernet disables the unneeded collision
detection circuitry. A full duplex-connection is a effectively
point-to-point circuit, with the local device's receive pair of wires
connected to the transmit pair for the remote device, and the local
device's transmit pair connected to the remote devices receive wires.
Collisions are only possible in half-duplex mode.
The Cisco documentation is clear that Gigabit Ethernet interfaces in
Catalyst 6000 series switches is full-duplex only, which precludes
collision detection.
This more general link (below) implies that Gigabit Ethernet collisions
are possible. Admittedly, this reference is pretty high level and does
not get into details.
http://www.cisco.com/warp/public/779/smbiz/netguide/ii1_lan.html
"As noted in the previous section, Ethernet bandwidth or data-carrying
capacity (also called throughput) is 10 Mbps. Fast Ethernet and Gigabit
Ethernet works the same way-through collision detection-but provide 10
times the bandwidth, at 100 Mbps, and 100 times the bandwidth, at 1000
Mbps, respectively."
---This link is about a Gigabit Ethernet Circuit Board for the PIX 525, and includes the mode "1000basesx - forces half duplex operation"
http://www.cisco.com/univercd/cc/td/doc/product/iaabu/pix/pix_v53/inst_5 25/board.htm#xtocid5
---
Although most of the switch documentation I came across says that the Gigabit Ethernet ports cannot use half-duplex, the link above indicates the 2900XL/3500XL Gigabit Ethernet ports "can negotiate full or half duplex with the attached device."
http://www.cisco.com/univercd/cc/td/doc/product/lan/c2900xl/29_35wc3/sc/ swgports.htm#xtocid1
Since it half-duplex mode is possible on some Cisco Gigabit Ethernet devices, I would think that collisions are also possible.
HTH,
Jonathan
-----Original Message----- From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf Of Ralph Simmons Sent: Thursday, September 25, 2003 9:47 AM To: Barney Gaumer; ccielab@groupstudy.com Subject: Re: gig ethernet/ collision debate
so is there any way that you would get collisions then on a gig link other then faulty equipment then? My buddy says he has seen it before and I saw it is not possible. What could of caused a interface to show collisions or did he just not know what he was looking at?
Barney Gaumer <bagaumer@yahoo.com> wrote:Since Duplex is not a settable option (at least on 6509 platform) you should not see collisions on either fiber or copper since it is Full Duplex by default.
Again, 6509 is the only platform I have looked at this on, if other platforms can change Speed/Duplex on Gig someone else will half to speak to that. I think Gig is inherently FDX though.
6509-a (enable) sh port 3/7 Port Name Status Vlan Duplex Speed Type ----- -------------------- ---------- ---------- ------ ----- ------------ 3/7 notconnect 1 full 1000 1000BaseSX
6509-a (enable) set port speed 3/7 1000 Feature not supported on Module 3. 6509-a (enable) set port dup 3/7 full Feature not supported on Port 3/7. 6509-a (enable)
6509-a (enable) sh port 9/1 Port Name Status Vlan Duplex Speed Type ----- -------------------- ---------- ---------- ------ ----- ------------ 9/1 notconnect 1 full 1000 1000BaseT
6509-a (enable) set port speed 9/1 1000 Feature not supported on Module 9. 6509-a (enable) set port dup 9/1 full Feature not supported on Port 9/1. 6509-a (enable) --- Ralph Simmons wrote: > Hoping one of you guys can settle a debate i am with > a coworker and win me a free lunch from him. > > We were debating about wheter of not you can get > collisions on gig ethernet. Is the collision > detection removed from gig ethernet? If not under > what circumstances could you get collisions? Any > documentation of cisco.com would be great to prove > it. > > Thanks, > > > --------------------------------- > Do you Yahoo!? > The New Yahoo! Shopping - with improved product > search > > ***Get your CCIE and a FREE vacation: > Shop.GroupStudy.com*** >
This archive was generated by hypermail 2.1.4 : Wed Oct 01 2003 - 07:24:36 GMT-3