Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: 64/66 on SONET (RE: AIS (RE: WIS OH))


First I did not say that block coding (64B/66B or 8B/10B) itself 
performs the trouble reporting function.

Instead, we are proposing another Link Signaling (LS) mechanism that 
can be implemented in LAN-PHY either on 8B/10B for WWDM or 64B/66B 
for Serial.  Therefore, if you need, you will be able to implement 
the LS mechanism even for 64B/66B on SONET, while I do NOT think we 
need it.

We are proposing the Link Signaling Sublayer (LSS) that enables us 
to transmit 4-byte LS Control Code LS/ID1/ID2/Info by using 
InterPacket Gap (IPG) period or during idles.  This is a Layer-1 
entity and its required [r] code carries its local Station Management 
(STA) register bit value, such as Link Status (Remote Fault) and 
Isolate (Break Link), to its Link Partner's STA.  The forth byte 
in LS Control Code (Info) can carry 8-bit row data or 3-bit data 
protected by minimum 4-bit Hamming distance.  The second and 
third bytes (ID1,ID2) are used to identify what kind of information 
the code is carrying; we are proposing optional three control 
codes those are compatible to WAN-PHY OAM&P.

Therefore, if you need, you may define your own LSS control code 
for your three trouble reporting functions.  I think it's a kind of 
Vendor/Carrier specific part and it is beyond the scope of 802.3ae, 
similar to the additional SONET overhead bytes in 802.3ae WAN-PHY, 
such as M1, K1, K2, D1-12, N1, ........

In a bottom line, I still DO support 64B66B within the WAN PHY.

Best Regards,

At 0:15 PM -0500 00.6.11, Roy Bynum wrote:
> Please give details on how you think that you would be able to 
> implement the three trouble reporting functions that I have 
> recommended through the use of the original proposal for the WAN 
> compatible PHY.  There are a potential of additional functions 
> within that original proposal that have been ignored by the block 
> coding cadre, mainly because they can not replicate that level of
> functionality.
> 8B/10B was invented by the same company that invented SDLC, IBM, 
> specifically for short, self contained communications that did not
> have need of remote reporting functionality.  Some functionality 
> has been added, but block coding remains severely handicapped,
> because it was not invented to be used outside of a limited 
> distance environment.  For extended distance communications IBM had
> SDLC.  Granted, SDLC was designed to be used over unreliable 
> analog transmission systems.  But the principal of reserved bytes for
> command and control has remained.  Block coding does not have any 
> place for reserved bytes for command and control.  Unless you want
> to put all of the functionality back into the SONET overhead, 
> I suggest that you re-think your support for block coding within the
> Thank you,
> Roy Bynum
> ----- Original Message -----
> From: "Osamu ISHIDA" <ishida@xxxxxxxxxxxxxxxx>
> To: "Roy Bynum" <rabynum@xxxxxxxxxxxxxx>
> Cc: <stds-802-3-hssg@xxxxxxxx>
> Sent: Sunday, June 11, 2000 3:42 AM
> Subject: 64/66 on SONET (RE: AIS (RE: WIS OH))

NTT Network Innovation Laboratories
TEL +81-468-59-3263  FAX +81-468-55-1282