|
Posted by Manfred Kwiatkowski on August 25, 2005, 9:56 am
If you were Registered and logged in, you could reply and use other advanced thread options
mark.boolootian@gmail.com writes:
>Manfred Kwiatkowski wrote:
>> As there is no autonegotiation defined on a 100BaseFX any possible
>> negotiation is local to the switch and its adjacent converter.
>
>This is exactly my question. Jay's yielding to Murphy aside, if the
>immediately adjacent converter is not autonegotiating, isn't it true
>that the autonegotiation would then happen with the remote switch?
>Autonegotiation consists of sending a 16-bit link code word
>adverstising
>capabilities. I'm assuming, and have some empirical evidence
>confirming,
>that if the adjacent converter isn't autonegotiating, it is going to
>just pass the link code word on, making its way to the remote switch.
>
No. The code word cannot make it to the other side as there is no
FLP defined on fiber.
>One possible problem, which may have to do with Jay's observations,
>is that if the remote converter is set to autonegotiate, the remote
>switch will be receiving link code words from the near switch *and*
>the remote converter. Worse, the link code words sent by the remote
>switch will be consumed by the remote converter, leaving the near
>switch to assume the neighbor isn't autonegotiating (which ends up
>being true).
The autonegotiation is independent of the fiber side and soley
a matter of the hardware configuration of the MC.
There may be difficulties with link establishment (oszillations)
due to "missing link" features and/or auto-mdix specially if the
capaciators in the MC start to degrade.
--
Manfred Kwiatkowski kwiatkowski@zrz.tu-berlin.de
| 
Yahoo! 
Google 
Windows Live 
del.icio.us 
digg 
Netscape 
XML site map