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

Re: Please help to clarify some things!




Dae,

There was an attempt to decoupage the signaling encoding from the MAC data rate because the
9.584 MAC rate would work for a block encoded LAN only PHY in addition to a WAN compatible
PHY.  The MAC rate of 10.00 may work for the block encoded LAN only PHY, but it will not
work for a WAN compatible PHY.  The only objective that was related to these issues was the
MAC rate.   The issue of a standardizing on a MAC that can support a WAN compatible PHY as
well as a LAN only PHY instead of standardizing on a MAC that can not support WAN
compatibility continues to be related to cost, processing overhead, and other issues
related to the encoding.

Thank you,
Roy Bynum
MCI WorldCom

Dae Young KIM wrote:

> Roy,
>
> I'm sorry, but when did I assert "WAN compatible PHY should be more expensive than a
> block encoded LAN only PHYZ"? When did I ever comment on the cost?
>
> You're suddenly jumping into the cost which I did not mention.
>
> My point is the issue of line speed and line coding should be decoupled; if not
> completely, then as far as possible.
>
> In fact, you seem to mix up three or four different issue into a bowl.
>
>     - line speed
>     - line code
>     - managemet
>     - cost
>
> Although for any PHY, all the four issues are involved, in pushing your WAN idea, I
> think it's better tackle each issue one-by-one. One of the reason people get confused
> and the discussion takes too long is that, IMHO, you jump among these rather orthogonal
> issues in establishing your logic. You even seem to try to mix up the things and catch
> three (or even four) birds with a single stone.
>
> Yet, I don't think it's a good strategy. You do have a lot of appealing logics, but why
> don't you try to tackle the issues one by one? By jumping among different issues and
> confusing people, you might rather lose more than win something.
>
> I do like and respect your concern and affection on WAN aspects, but let's do the job
> one by one.
>
> Roy Bynum wrote:
>
> > Dae,
> >
> > Where does the idea come from that developing a new technology that operates as a
> > higher signaling rate, with more active bytes per data traffic payload, will be less
> > expensive than a technology that has already been developed, operates at a lower
> > signaling speed, and has 4 bytes of active overhead per 15000 bytes of data traffic
> > payload?  I am confused, when does lower processing, and lower speed become more
> > expensive?  Unless the interface price is artificially inflated, a native data 10GbE
> > WAN compatible PHY should not be more expensive than a block encoded LAN only PHY at
> > the same laser output.
>
> --
> Dae Young KIM
> http://ccl.chungnam.ac.kr/~dykim/