Transmit Disable and Rate Select
Transmit Disable and Rate Select may be important signals, but their
location in singular. Specifically, these signals are located solely
within the PMD. The requirements for these signals is significantly
different than those for Link Status signals like signal_detect and
sync_detect since the latter signals may be generated by multiple PHY
sublayers, and in some cases, for multiple physical instantiations
within the same sublayer. The transport requirements for these latter
signals is not so simple as it is for Transmit Disable and Rate Select.
There is no reason for sublayers up the stack from the PMD to set or
check the value of either Transmit Disable and Rate Select. Control and
monitoring of both Transmit Disable and Rate Select may be performed by
station management. Pins = Bad. Both Transmit Disable and Rate Select
are ideally suited for transport via MDIO. MDIO is already available at
the PMD for P802.3ae. I propose that both be specified as optional
signals in clause 30.
> In the same sense as Rich did, I would like to throw in my 10 cents.
> Just as the link status needs to be defined, I see also the need to define two additional signals in the other direction. Namely transmit disable and rate select.
> Transmit Disable
> laser transmitters are designed to always transmit light at a predefined average level. So even if the transmitter gets all zeroes at the input, after a certain amount of time it will start transmitting static light. From the fiber Channel the experience has been, that in some instances the cross talk from neighboring transmitters could be modulated onto this static light signal.
> The receiver on the other side has gone into maximum amplification, because it is not receiving a signal, and the outputs are "chattering" on the input noise. If the input noise, to the receiver, contains some valid signals, the signals might be interpreted by the upper layers as valid signals, and the system will oscillate between initialization and shutoff.
> To avoid this, the new transceivers have a transmit disable pin, that will shutoff the light coming out of the transceiver. This signal has to be provided by the upper layer.
> Rate Select
> This is also a new single on transceivers and in 10 GbE will be optional. But I can see for some implementations that this signal could be useful to switch between different receive filters for different data rates, or different PMD's, if they are implemented.
> I strongly believe that we should also define these signals.
> Have fun
Richard Taborek Sr. Phone: 408-845-6102
Chief Technology Officer Cell: 408-832-3957
nSerial Corporation Fax: 408-845-6114
2500-5 Augustine Dr. mailto:rtaborek@xxxxxxxxxxx
Santa Clara, CA 95054 http://www.nSerial.com