Re: bit ordering on XSBI vs SFI-4
Jscquake@xxxxxxx wrote:
> <snip>
> As the clause editor for clause 51, I have received feedback that the
> present
> bit ordering for the XSBI interface, i.e. LSB sent first, will cause
> confusion to people that have used the SFI-4/SONET modules.
>
> Background on this topic ... an OC192 serial stream has the MSB bit
> first
> while Ethernet networks have the LSB bit sent first. Early 802.3ae
> discussions, we decided to leave the LSB bit order as is so that the
> PCS
<snip>
To clearify the problem space there is an incompatibility between
the XSBI specification and the SFI-4 (OIF1999.102.8) spec.
XSBI transmits the bit at location Tx_Data[0]
(called the LSB) first. Similarly it receives Rx_Data[0] (called
the LSB) first and Rx_Data[15] (MSB) last. (ref page 171-173,
page 337-338, 369, 375, 383, 408 of 802.3ae_D2.1)
SFI-4 transmits Tx_Data[15] (called the MSB)first. it Receives
Rx_Data[15] (called the MSB) first and Rx_Data[0] (LSB) last.
(Ref table 6-1 note 1).
Justin's proposal is to change the XSBI specification so that it
is pin compatible with the SFI-4 spec. This is performed by
remapping the LSB into XSBI data bit 15, and the MSB into
XSBI data bit 0. That is, a bit swap at the PMA client.
When this is done, the LSB is transmitted first if an SFI-4
compliant transponder is used.
Justin would then modify the PMA so that Tx_Data[15] is serialized
first. Now the LSB is transmitted first by the PMA device.
Essentially there are two bit ordering swaps which occur. The
first is by the PMA client (Clause 49 and 50) the second is
within the PMA (clause 51).
The advantage of this "editorial" change is that SFI-4 compliant
transponders may be used in place of XSBI transponders with
no re-engineering. This is particularly attractive for early
adoptors who may be targetting SFI-4 transponders while
waiting for XSBI compliant parts.
--
Tim Warland P.Eng.
Hardware Design Engineer Broadband Products
High Performance Optical Component Solutions
Nortel Networks (613)765-6634