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

Re: [8023-POEP] Call for Presentations



Jose-

(I have included your 802.3ah Sponsor Ballot comments below for the information of the others on this distribution list.)

Your referenced paper "uets-eng.pdf" is not appropriate for consideration by this group as it is watermarked "CONFIDENTIAL". Consideration of material with such restrictions is contrary to IEEE-SA policy.

I don't think that the groups is headed in the direction that you are looking for. The consensus (so far) seems to be for the fewest changes possible with maximum consideration given to forward and backward compatibility with 802.3af and existing Cat. 5 and 5e standardized cabling. It is true that the PAR is not yet finalized nor approved by 802.3 so there is still time for the focus of the project to move. You are welcome to attend the meeting and try to shift the sense of the group.

My personal opinion with respect to your comment #121 is that this is a systems management issue. It should be handled through higher layer protocols on a system-to-system basis and thus it actually has nothing to do with Ethernet. By that measure, it is out of scope for 802.3. Whether it would be more appropriate for elsewhere in 802 or the IETF is for discussion outside this group.

My own opinion regarding your comment #120, my opinion is that an entirely different standard is needed for power in the EFM domain. Many of the considerations are quite different. My personal belief is that the greater value here would be in a backfeed system that would be capable of providing power to outside distribution switches (e.g. those with fiber optic backhaul to the Central Office) from customer premise demarcation devices.

I believe the group wishes to have a small (hopefully rapid) project with limited scope. Under that philosophy, I believe it would be  more appropriate for your suggestions and efforts to be directed to new projects.

With best regards,

Geoff Thompson

At 02:59 PM 12/27/2004 +0100, Jose Morales wrote:
Michael,
 
During the final voting phase (Balloting Sponsors) of the IEEE 802.3ah EFM project, I sent two comments related to power supply and managment that  were favorably received. The response was: "The suggestion may make a great new project, however, extending af is out of the scope of EFM's objectives."

As the objective of 8023-POEP is to extend 802.3af, I believe that it would be interesting now to consider my proposals, that are in the first page of the document:

http://www.ieee802.org/3/efm/public/comments/d3_1/D3_1_proposed_responses.pdf

With more than 1,200 million telephone lines over the world, and 802.3ah, there is a broad market potential for a solution like this, as I describe in a paper you will see in

http://www.lmdata.es/uets-eng.pdf
 

Thanks,

Jose Morales Barroso
jmb@xxxxxxxx

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Cl 00           SC      P       L       #121
Morales Barroso, Jose           L&M Data Communica

Comment Type T
The large number of connections based on EFM that will exist in the future makes it very advisable to apply power management procedures (copper & optical fiber) in order to eliminate "ghost power", because the average use of this connections is less than 5 hours/day (< 20% of the total time). For example, with 200 million users, the energy saving would be of the order of 14 TWh/year, equivalent to 1,4 Billion Euros(<>1,75 Billion $).
SuggestedRemedy
There is a power management specified in Std 802.11-1999, Clause 11, Subclause 11.2, that will serve as a basis to implement the control via the OAM protocol or with a specific procedure. In order to reduce the power consumed by the equipment, diverse components of these equipment can become disconnected during periods of inactivity.

Applying power management to all the Ethernet equipment (not only EFM) would result in a huge energy saving, due to the high number of devices that use this technology.
Cl 00           SC      P       L       #120
Morales Barroso, Jose           L&M Data Communica

Comment Type T
It is very important for Ethernet over voice-grade copper connections to supply power either from the Central Office (CO) like in the European ISDN, or from the switch/hub in LANs. This feature would have many advantages, not only for service providers, but also for users of 10PASS-TS or 2BASE-TL in LANs and campus networks; since it enables users to connect remote devices. For example, the cameras used in video-monitoring systems would need only one pair to transport the signal and the power, rather than the four pairs specified now in 802.3af. This would increase the link length from 100 m to 2 km.
SuggestedRemedy
The specifications of 802.3af should be extended to include voice-grade copper, in order to make the changes required.

Clause 31 should be modified to something like this: "DTE powering is intended to provide both data transfer and power feed to 10BASE-T, 100BASE-TX, 1000BASE-T, 10PASS-TS or 2BASE-TL devices".

As a reference, the ISDN and HDSL connections in Europe supply power from CO to the CPEs through POTS cables.