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

RE: Signal vs. Idle debate (was: Here's a new idea)


Please see below your message:

At 02:22 AM 5/5/00 -0600, Paul Yew wrote:

>I've just joined the group, so this may be out of line.
>IMHO, it seems that what IEEE decides may not matter. While IEEE debates 
>the method that "should" be used to provide power down Ethernet cabling, 
>Cisco, the leader in the industry, has announced what the solution is. 
>Long before IEEE ever gets a standard agreed to and published, the 
>following products will be widely used and any standard that doesn't 
>support this de facto implementation will simply be ignore.
>If you refer to 
>you will find the basic specifications for Cisco's Catalyst Inline power 
>Patch Panel.  This device does mid span power insertion on pins 4&5 and 7&8.
>If you refer to 
>_ov.htm you will find the basic specifications for Cisco's Catalyst 6000 
>switch family which powers endpoints over pins 1&2 and 3&6.
>Based on these products, Cisco has determined that endpoints will have to 
>be able to be powered from the signal and/or idle wire pairs.
>Obviously Cisco has determined a viable solution so why not accept it?
>Paul Yew

Yes I believe that you are out of line.

We are fully aware of the impact that proprietary implementations have on 
the marketplace.

We are also aware of the imapct that true consensus Standards can have on 
the marketplace.

If one were to follow your line of reasoning one would have to ask "Why are 
you participating or for that matter why is Cisco?" The logical extension 
of your philosophy would be that there is no place for industry consensus 
for setting technical standards. We should all just yield to the most 
dominant company for any particular implementation. If we followed that 
line of reasoning then:
         1) There would have been no Ethernet,
           we would have all rolled over when
           IBM proposed Token Ring.

         2) There would have been no 10BASE-F,
            we would have all rolled over when
            Codenoll proposed CodeStar.

         3) There would have been no 10BASE-T,
            we would have all rolled over when
            SynOptics proposed LattisNet UTP

And then, because the specs would have been owned by proposing company, 
there would not have been plug compatible competition (take a look at the 
mainframe disk wars of everybody vs IBM in the early 70s) and we wouldn't 
have seen the price erosion that has served the customers so well.

Companies grow and shrink, technologies exert profound changes on market 
dominance. Market sizes are drastically changed by new features. It is 
entirely possible that the new markets created by this standard will be 
served by vendors who are not currently a dominant presence in networking.

To support your side, the fact that a major force in the industry has 
announced and/or delivered product is very much a part of the standards 
equation in both the pro and con direction. Trust me, we do notice what is 
going on around us.

Some standards groups do a bad job on this and the market ends up ignoring 
their work. Other groups through skill or luck manage to do a better job. 
They actually set standards that serve and expand the market and nurture 
competition which results in broader availability and lower cost to the 
customer. We hope we are the latter.

Best regards,

Geoff Thompson, Chair, IEEE 802.3