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

Re: why two different coding schemes ?


As Rich mentions 8B/10B is very established in the industry.  More
important we need to look at the attribute it provide such as DC
run length, disparity, and error detection.  

The draw back of the 8B/10B has been argued is the 20% coding overhead.
When an 8B/10B coded data operating 20% faster compared to PRBS 2n31-1
for long copper traces, 8B/10B is comparable even though it was running 
20% faster.
AC coupling signal with 8B/10B only require modest size
capacitors, but running dense backplane with 64/66 require large 
capacitors often not practical.

At the request of the optical manufactures, which did not want to 
operate 20% faster the 64/66  code was selected and the jury is 
still out.




Rich Taborek wrote:
> Jagmit,
> 8B/10B is the most widely deployed transmission code for datacom
> fiber-optic link operating at 200 Mbps and above. It is the only code
> used in 1000BASE-X (Gigabit Ethernet) and would likely have been be the
> only code used for the 10GE LAN PHY if it weren't for the difficulty in
> obtaining 12.5 Gbps opto-electronic components at this point in time.
> 8B/10B is very "optics friendly".
> 64B/66B is employed in both the LAN PHY to keep the line rate of the
> opto-electronics down to 10.3125 Gbps while supporting the 10GE data
> rate of 10.0 Gbps.
> Note that 10GE supports scrambled encoding for the WAN PHY which is a
> third coding scheme.
> I strongly disagree that coding schemes complicate life for anyone. A
> typical path from the central processor of a PC to the media platter of
> a disk drive in a low cost system is likely to employ many more than two
> coding schemes and perhaps 10 or more. Every coding scheme employed is
> typically optimized for its particular application.
> It is inappropriate to use this reflector or any other IEEE 802.3 forum
> to discuss products and timings. Please take any such discussions and
> questions off line.
> Best Regards,
> Rich
> --
> Jagmit Sandhu wrote:
> >
> > Hi, I'm new to 10 Gb/s ethernet. Could somebody explain why two
> > different coding schemes are being contemplated for 10GE ? i.e.
> > 8B/10B and 64B/66B? Why not stick with just one to simplify life for
> > everybody ? Also, are there any vendors with tentative 10GE products
> > coming out soon?
> >
> > thanks
> > jagmit sandhu
> -------------------------------------------------------
> 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  
org:Broadcom;Optical Transport
adr:;;3151 Zanker Rd	;San Jose;CA;95134;
fn:Ali Ghiasi