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

Re: [8023-POEP] [802.1 - 4934] AVB & Power Management

Geoff,

I fully agree. Thank you for your inputs

/Ph

From: Geoff Thompson [mailto:gthompso@xxxxxxxxxx]
Sent: Wednesday, December 24, 2008 19:05
To: Philippe Klein
Cc: STDS-802-1-L@xxxxxxxxxxxxxxxxx; STDS-802-3-POEP@xxxxxxxxxxxxxxxxx; STDS-802-3-EEE@xxxxxxxxxxxxxxxxx
Subject: Re: [802.1 - 4934] AVB & Power Management

Philippe-

See me response mixed in with your text.

At 06:40 AM 12/24/2008 , Philippe Klein wrote:
Ballot on P802.1AR/D1.9 due Jan. 1
  Also in ballot: P802.1Qay/D4.6, P802.1Qaw/D4.0 (both Jan. 5)
    P802.1aq/d1.5 (Jan. 9), P802.1X-rev/D3.2 (Jan. 14)
For particulars see www.ieee802.org/1/email-pages/ballots.html
-----
For subscription help, including changes, see
               www.ieee802.org/1/email-pages/rqjv1007.html
-----

Geoff, John,

Thank you for your inputs to my initial mail. I apologize for my lack of responsiveness but I was kept away from my desk for a few days.
I should confess that I did not track the power management developments for many years but will learn the 802,3 standard.

My mention to  WoL was certainly not a proposition to incorporate this particular protocol in the standard. I should have use a more generic description:
My point here was rather or not AVB will require a sleeping device to stay connected thru any wake-up mechanism (which might be out of the scope of AVB) or will use a proxy scheme instead.

Very good point and one that is entire germane to the subject under discussion.
My point is that both WoL and PoE are viable with different plusses and minuses, but they DO have a basic difference:
  - WoL does "require a sleeping device to stay connected" as described above
  - PoE requires no active connection whatsoever but rather uses a "proxy scheme" since the power-on action has to take place in the PSE of the powering/controlling device rather than the end device itself.


The main point of raising the power management issue in AVB is to discuss how a sleeping AV device should be handled and prevent the AVB protocols to impact its power saving budget.

I would assert that the sentence above should be rewritten slightly to say:
"The main point of raising the power management issue in AVB is to discuss how powering AV sub-system devices should be handled and to include AVB protocols to maximize power savings when the AVB system is not in active use."



I am looking forward for a very interactive discussion on the subject at the next interim meeting.

Me too.

Best regards,

Geoff


/Philippe


-----Original Message-----
From: IEEE 802.1 list HELP only [mailto:hdk-918.xb9n_m63sj@xxxxxxx] On Behalf Of Geoff Thompson
Sent: Friday, December 12, 2008 21:55
To: STDS-802-1-L@xxxxxxxxxxxxxxxxx
Subject: Re: [802.1 - 4859] AVB & Power Management

Ballots due early January:
  P802.1AR/D1.9 (Jan. 1), P802.1Qay/D4.6 (Jan. 5)
For particulars see www.ieee802.org/1/email-pages/ballots.html
-----
Didn't see  your contribution?  DON'T PANIC.  Before sending  a duplicate
post, see "Sending->retrying" at
               www.ieee802.org/1/email-pages/rqjv1007.html
-----

John-

It is true that there are higher level protocol issues that must be
addressed. Any standards based software mechanisms that want to do power
control has to define an interface somewhere between those mechanisms and
the underlying hardware. My point was that this interface is already
defined in the approved standards space within the MIB and mechanisms of PoE.

As to your second point:
 >  Also note, that we cannot limit ourselves to devices that
 > have no wall plug.  I would hope that your mechanism
 > address these self-powered devices as well,

It is not "my mechanism", it is a very widely adopted IEEE 802.3 standard.
While it admittedly does require wires to work (There have been many
requests for an 802.11 version of PoE but the proposers have not made it
through the Technical and Economic Feasibility aspects of the 5 Criteria
yet) it is not limited to devices that are not "self powered".  The power
control aspects of PoE can be effectively utilized where PoE only provides
a small portion of the target device power and is present largely for the
presence of the power control feature.

Several presentations were made to the Energy Efficient Ethernet group on
transition to and from a power down state. In particular, see my
presentation
<http://www.ieee802.org/3/eee_study/public/jul07/thompson_2_0707.pdf> which
was followed up by Bob Grow's presentation a meeting later
<http://www.ieee802.org/3/eee_study/public/sep07/grow_1_0907.pdf>.

My point is not that there are no higher level protocol issues, quite the
contrary. Lower level mechanisms are in place, either as the already
standardized PoE or the widely adopted (but not yet standardized) WoL.

My only fear is that the entrenched proprietary aspect of WoL would be an
impediment to the standardization required for its inclusion.

Best regards,

        Geoff

At 08:50 PM 12/11/2008 , John Nels Fuller wrote:
-----
 >
 > Geoff,
 >
 > Your points are well taken.  However, regardless of the
 > mechanism of waking a sleeping device (or for that matter
 > putting it to sleep) there are higher level protocol issues
 > to be addressed.  Also note, that we cannot limit ourselves
 > to devices that have no wall plug.  I would hope that your
 > mechanism address these self-powered devices as well, but, as
 > I said, it does not change the higher level protocol issues.
 >
 > jnf
 > _________________________
 > John Nels Fuller
 > Interconnect Solutions Architect
 > 48 NE Mulberry Ct.
 > College Place, WA   99324-2101   USA
 > Mobile phone: +1 206 409 0338
 > Home: +1 509 526 4977
 > Fax: +1 509 529 2073
 > email: jfuller@xxxxxxxxxxxx