A proposal for the "Five Criteria"
In order to perform productive work at the Power Over MDI
Study group meeting next week in York, I?m going to float
the following proposed text. I think that if we have a good
starting point, then we can address the "Five Criteria" (see
http://grouper.ieee.org/groups/802//3/rules/rules.html#P72 ) next week
we stand a chance at getting a PAR.
1) Address the "Broad Market" criteria:
- Board set of applications:
- Multiple vendors, multiple users.
- Balance cost, LAN vs. attached stations.
Currently three groups that have presented information,
telephone stations (estimated 8M nodes in 2001,)
wireless access points (estimated 700K nodes 2001,) and
theatre/lighting control (no numbers presented.)
Additional applications have been identified such as
surveillance cameras, card keys and office controls.
34 companies/organizations sent attendees to the call for
interest meeting in Montreal.
All attendees (44 people/34 organizations) to the call for
interest voted in favor of forming a study group
20 companies/organizations indicated that they would work on
developing such a standard
All devices presented currently require a power source, which
is burdens the attached station. The proposal to provide
power over the MDI interface does shift that cost to the LAN;
however, the centralization of power should reduce the
over-all system cost to the end users.
** Strictly speaking this presents the hardest criteria to
justify. **
2) Address the "Compatibility" criteria:
- Conformance with CSMA/ CD MAC, PLS.
- Conformance with 802.2.
- Conformance with 802 Functional Requirements.
The proposed standard will be compatible with the CSMA/CD
MAC and PLS, with currently authorized extensions. Changes
to the PLS will be required as the proposed standard
development is a PLS change; however, compatibility with
existing installations will be maintained.
The proposed standard will conform with the 802.2 LLC
interface, it does not effect it.
The proposed standard will conform with the 802 Functional
Requirements
3) Address the "Distinct Identity" criteria:
- Substantially different from other 802.3 specifications/
solutions.
- Unique solution for problem (not two alternatives/ problem).
- Easy for document reader to select relevant spec.
The proposed standard will be an optional component to any
of the twisted pair interfaces. It differs from the
existing 802.3 physical layer definitions as nothing in
the 802.3 specification address powering devices, locally
or through the MDI interface.
Only a single powering system will be proposed. Implementers
will only have the option of producing equipment that either
include or does not include this clause. Optional inclusion
is similar to 802.3ad, no implementor is required to
implement link aggregation; however, if an implementor
chooses to provide the functionality only one method is
specified.
The proposed standard will be the only 802.3 component that
will address power. The standard will be formatted as new
clause. Together, these should make it easy for a reader to
find relevant information.
4) Address the "Technical Feasibility" criteria:
- Demonstrated feasibility; reports - - working models.
- Proven technology, reasonable testing.
- Confidence in reliability.
Presently, several proprietary implementation of the
technology exist in the market place. 802.9 had developed
significant information regarding power over MDI
configurations during a study for their standard.
Both the telephone network and 802.5 have always provided
power over the same cable plant as the data signals.
5) Address the "Economic Feasibility" criteria:
- Cost factors known, reliable data.
- Reasonable cost for performance expected.
- Total Installation costs considered.
The standard proposes to move cost from the end station to
the LAN infrastructure. Inclusion of power sourcing
capabilities will not be a zero cost addition to hubs or
switches, it is expected that both power sourcing and
non-power sourcing units will be available in the
marketplace. Costs for customers who are not interested in
providing power over their cable plant will not be
adversely effected.
The target areas are low power devices where the cost of the
front end of the power supply is generally attributed to the
number of components, not the current draw. A centralized
power system should be more economical than a distributed one.
Should users be interested in backup power, centralization of
the power source will drastically reduce power over separate
backup power systems per station.
Installation costs for two wiring systems, one for power one
for data are at least twice the cost of a single wiring plant.
While some installations would already have installed power
systems, their would be no additional cost added to the
installation of twisted pair cable.