Comments on 802.16 WirelessHUMAN PAR from within Other Working Groups

From: Stuart J. Kerry [mailto:Stuart.Kerry@philips.com]
Sent: Tuesday, September 26, 2000 8:47 PM
To: stds-802-11@ieee.org
Subject: WLAN/ 802.16 New PAR

Colleagues,

As you are aware from last weeks meeting in Scottsdale, Roger Marks (Chair,
IEEE 802.16) asked dot 11, and dot 5 to look at the PAR for his group
regarding WirelessHUMAN SG.

I know that some of you have looked at this already, especially those folks
in 5GHz, and I would like to have the perspective from the group, so we have
a consensus for the Ex-com where the PAR will be reviewed.

Please pass your comments back to me via the reflector.

Regards,

Stuart

Roger wrote: "At last week's 802.16 Session, the Working Group approved a
draft PAR created by our WirelessHUMAN Study Group. The current version is
at .

As you know, we have coordinated with 802.11 and 802.15 on this project. The
PAR demonstrates this coordination by virtue of the fact that the PHY is
specified to be a modification of 802.11a (and/or HIPERLAN/2).

-----------------------------------------------------------------------------------------

Date: Wed, 27 Sep 2000 12:39:37 -0400
From: "Bob Heile" 
To: stds-802-wpan@ieee.org
Subject: wpan/ Fwd: please review 802.16's WirelessHUMAN PAR

Attached is an email from Roger regarding the wireless HUMAN PAR.  This will be
on the 802 Executive Committee agenda for action at the upcoming meeting in
Tampa.  Any comments we may have as a WG must be submitted by 5pm on Tuesday at
the meeting.  This will be on the WG agenda for Monday.  Please review as we
will determine at that time if we wish to submit comments as a WG.  If you do
have comments please circulate them via the reflector ahead of the meeting.

Thanks

Bob


-----------------------------------------------------------------------------------------

Folks,

Disclaimer:  This is my own opinion.  It is not necessarily shared nor
endorsed by any of my clients.

I think it is time that we stop pretending that the airwaves are just like
Cat 5 cable, that we can run independently developed standards over the air
without worrying about the fact that they share the same medium,
simultaneously.  The situation is bad enough with disparate WLAN standards
developed by 802.11 and ETSI BRAN.  I understand that there is a group of
companies that will attempt to deploy point to multipoint systems in the
unlicensed bands, regardless of whether 802 creates such a standard.  I
don't believe that we should support that effort.  In fact, I believe that
without 802 support the effort is doomed to ultimate failure.

My opinion is that we should not develop such a standard.  It will only
serve to turn another unlicensed band into a trash band, as the 915 MHz band
has become and the 2.4 GHz band is rapidly becoming.  I believe that it is
time for 802 to take an activist position with the regulatory bodies, as
ETSI has, and lobby for spectrum etiquette in the 5GHz UNII band (and its
corollaries elsewhere) that will require users of the band to detect other
users and share access on an equal basis.  This may make it impossible to
offer certain services, such as fixed point to multipoint distribution.
That will be fine with me.  There are plenty of licensed bands for that type
of service (and .16 is already developing standards for them).  Oh yeah, and
there are no such bands for licensed WLAN service.  The least we should do
is to ensure that the unlicensed bands remain usable for WLANs.

The other action that I believe 802.11 and 802.15 should take is to propose
a change to the 802 operating rules.  The rule I would suggest be added
would require that all 802 standards that may operate on the same medium
simultaneously shall incorporate deterministic means to detect other 802
standards operating on the same medium and share access to the medium on a
fair and equitable basis.  This change will then require the WirelessHUMAN,
should its PAR be approved, to share the unlicensed bands with all the WLANs
in its service area.  It will also place the same requirements on 802.11 and
802.15.  Without such a requirement for standards development in 802, the
LMSC stands a fair chance of being seen as a schizophrenic body that has no
clear direction in the wireless arena, standardizing anything and everything
without regard to their impact on each other.

I welcome your comments.

 -Bob

Bob O'Hara
Informed Technology, Inc.

-----------------------------------------------------------------------------------------

Stuart,
Having the WirelessHUMAN system collocated with WLANS is like having cell phones
and cordless phones share the same band.  I don't like the idea one bit.  The
potential for interference is great whether it is in 2.4 or 5 GHz.  If we are
forced to live with them, the only thing to do is allocate a part of 5 GHz for
WLANS (5.15 to 5.35) and a different part (5.7-5.8) for W-HUMAN.  The 2.4 GHz
band has too little room for allocating bandwidth to W-HUMAN for proper spectrum
management and reuse for WLANS.  
Current FH radios could stand the extra channels, but everyone wants to go to
higher data rates and that reduces the number of simultaneously operating
channels proportionately.  Thus there is no way that W-HUMAN can avoid harm to
WLANS if it shares the same band.  
Therefore, I would express concern if the IEEE starts a new PAR for a project
that will interfere with existing IEEE WLANS.

		Carl Andren
		Intersil

-----------------------------------------------------------------------------------------

Here we ride again! IEEE802.11 develops a standard, and from now on rather
than acting proactively of medium sharing mechanisms (or just acknowledging
that anything can happen in unlicensed bands), we yell "let's convert the
unlicensed band into 802.11 band!"

IEEE802.11 looses time after time an opportunity to collaborate in early
phase with other projects, rather than licking the wounds in the aftermath.
As an example, 802.11 turned a blind eye to my requests to widen the scope
of 802.11e and of the 802.11e-HIPERLAN/2 collaboration to include both LAN
and MAN needs. Wireless HUMAN could use the outcome!! The result: Wireless
HUMAN decided in the last meeting that it will base the project on 802.16.1
MAC, which is neither compatible with 802.11 MAC nor is very considerate to
other users of the band. Look at the bright side of it, however: They
decided to base the standard on 802.11a/HIPERLAN2 PHY.

I think that 802.11 missed the train. The decision will pass the exec, just
as WPAN passed the exec in spite of 802.11 opposition, and even if not,
there will be (and already are) Wireless Access systems in the band. If you
want some control over the outcome, suggest to embrace the  project and
propose that 802.11 will be the sponsor and the root project of the of the
Wireless HUMAN Task Group.

Naftali Chayat,
BreezeCOM

-----------------------------------------------------------------------------------------

It may already be too late for the 2.4 band.  But let's not let happen at 5
GHz what is happening now at 2.4.  Let's actively crusade for all users of
this valuable new spectrum to have mandatory coordination mechanisms.   The
"coexistence" group set up by WirelessHUMAN must also address coordination.
A WirelessHUMAN standard should not be approved by 802, unless coordination
mechanisms are in place that are satisfactory to 802.11, and 802.15 as well
(should they chose to introduce a standard that operates in this band).

As for Bob's suggestion of changing the operating rules, we largely agree,
except that we would probably have to grandfather the work going on at 2.4,
as we don't see how they can be made to cooperatively share the spectrum.
The decision to support multiple incompatible PHYs has virtually ensured
that active coordination is impossible.  We must not let that happen at 5
GHz as well.

Best Regards,

Matthew Sherman &
Bob Miller 

AT&T Labs - Research

-----------------------------------------------------------------------------------------

Naftali Chayat wrote:
> Wireless HUMAN could use the outcome!! The result: Wireless
> HUMAN decided in the last meeting that it will base the project on 802.16.1
> MAC, which is neither compatible with 802.11 MAC nor is very considerate to
> other users of the band. Look at the bright side of it, however: They
> decided to base the standard on 802.11a/HIPERLAN2 PHY.
> 
> I think that 802.11 missed the train. The decision will pass the exec, just
> as WPAN passed the exec in spite of 802.11 opposition, and even if not,
> there will be (and already are) Wireless Access systems in the band.

I think the exec will also consider our Spectrum Managed IEEE 802.11a
PAR at the same meeting. It is
very likely that Dynamic Frequency Selection will be MAC specific, not
so much PHY specific. A .16 MAC would use a different apparatus than a
.11 MAC because of different framing and signaling.

It would be useful for the 802 exec to have a crib sheet of bands,
coverage, capacity, regulatory rules:

		2.45	5	11+
WPAN
WLAN
WMAN

 for Europe, North America and Japan,
 now, after FCC 99-231 Processing Gain ruling (which also may affect the
5.725-5.85GHz ISM band),
 and after the 5GHz Worldwide Single Standard meeting in late October.

Maybe the chairs of .11, .15, and .16 could work together to create such
a chart, and address
the upcoming overlap between outdoor WLAN and WMAN, and on the other
hand indoor WPAN and WLAN.

> If you
> want some control over the outcome, suggest to embrace the  project and
> propose that 802.11 will be the sponsor and the root project of the of the
> Wireless HUMAN Task Group.

I disagree, and think their group must address outdoor CEPT requirements
in a different
fashion than WLANs do. You have a notion "control over the outcome",
which I 
haven't experienced in several years as liaison from .3 to .1, .9, and
.12. What is
your definition of "control"? and how would .11 exercise it?

I don't see the value of a .16 MAC in the band when FDD radios will
drive the high end, and 802.11 MACs will drive the low end. A distinct
middle ground
for them will be a small place in the market.

Peter Ecclesine

-----------------------------------------------------------------------------------------

Stuart:

I doubt strongly that an etiquette or any other form of dynamic sharing of a
channel in the 5GHz (or 2.4GHz) band will be capable of providing either the
performance or quality of service those deploying and using both the WLAN and
broadband access systems expect.  Calling for coordination among the various
groups and/or the development of a coexistence plan may be a politically
desirable solution, but the only approach to such a coexistence solution that
leaves the existing work of 802.11 & ETSI-BRAN able to operate in the
environment that those (ALREADY APPROVED) standards were developed for is
subdivision of the band into sections for various applications, as was
suggested by Carl Andren.  I do not regard this as a good solution -- the
proper approach is to reject this PAR until a suitable band can be identified
(or allocated ...).  There seems to be a tendency these days to propose
projects to standardize wireless MAC/PHY protocols for anything not prohibited
by technical or regulatory limits, without concern for the limitations of the
shared medium.  During the interim meeting 802.11 had the good sense not to
initiate work toward yet another 2.4GHz PHY.  Unfortunately, 802.16 did not
exercise the same restraint relative to the 5GHz band.

While I agree with the intent of Bob O'Hara's suggestion regarding a change to
the rules to require 802 groups using common media to develop deterministic
means to detect each other and share the medium fairly; I do not believe this
is sufficient.  What is needed is a change to the definition of "scope" in item
12 of the PAR form to define projects that propose to use the same INSTANCE of
a physical medium as INHERENTLY having the same scope, independent of the
purpose of the intended communication and/or the categories of anticipated
users thereof.  Would a PAR be approved for an incompatible, but simultaneous,
communication activity over an existing Ethernet cable plant?  I doubt it
(unless said activity were restricted to a non-overlapping subset of the
spectrum, as with 10broad36), yet that is exactly equivalent to what 802.16 is
proposing for Wireless HUMAN.  With this change to the definition of scope, a
project to serve a new application or user community has 3 alternatives -- use
or extend the preexisting MAC/PHY for a new purpose, find a different PHY
instance (e.g. another frequency band), or demonstrate the lack of overlapping
"scope" (e.g. that the spatial extent of the radiation they emit in the common
band will be disjoint from that expected in actual, continued deployment of the
preexisting standard -- which is implausible in the case of 802.11 & 802.16).

Is it possible to develop and specify a means to share fairly a channel in the
5GHz band between LAN and MAN applications?  Probably -- but will the result
allow performance that is acceptable to the users of EITHER network?  Probably
not!  This is certainly true for the users of 802.11(e), where the fundamental
MAC & PHY mechanisms were defined in advance of, and without consideration for,
a channel sharing mechanism with a MAC protocol such as 802.16 is defining (and
is likely to be no better for 802.16 users if  their protocol is not modified
to take account of the sharing mechanism -- but 802.11 is already finalized,
whereas 802.16 is not).  Imposing a post-facto etiquette and/or time-allocation
requirement on an existing protocol is unreasonable, and sets a dangerous
precedent against the longevity, quality, and value of any future IEEE 802
wireless standards work.  In particular, MAC functions specifically developed
to achieve acceptable service over an extremely unfriendly PHY, as are present
in 802.11 and HiperLAN-2, are likely to be rendered useless (or, at "best"
inappropriate and inefficient) by an effective scheme for dynamic deferral or
negotiated time-division sharing of a channel with a protocol intended for
fixed point-to-multipoint wireless access.  Furthermore, in the case of 802.11,
to accept sharing of the channel and cooperate in the development of a fair
sharing mechanism in conjunction with 802.16 (while continuing to be able to
offer conformant products which satisfy the purposes for which 802.11 was
chartered) requires reversal of one of the most fundamental concepts of
802.11:  To date all 802.11 projects have been based on the concept of a common
MAC over a plurality of PHYs, but an appropriate MAC for this sort of
dynamically shared channel most likely looks nothing like the existing 802.11
MAC -- unless Wireless HUMAN were to adopt the 802.11 MAC, in which case their
project does not belong in 802.16...    (To get a feeling about the complexity
of dynamic channel-sharing schemes, even when all the stations are using the
SAME basic protocol, look at some of the recent submissions proposing BSS
overlap mitigation mechanisms for 802.11e.  A non-exhaustive list of such
documents includes 00/071r1, 00/108, 00/110, 00/120r1, and 00/194.)

Some people are likely to argue that 802.11 is operating in an unlicensed band,
and that uses such as Wireless HUMAN is proposing can and will occur with or
without an IEEE 802 standard.  This is true, but is not a justification for
IEEE 802 LMSC to participate actively in development and promotion of
conflicting standards which, by their very success, will prevent those who
invested in equipment conformant to an IEEE 802 standard from obtaining the
benefits for which that standard was developed and under whose reputation for
interoperability they made their equipment selection.

With wired media the sharing problems can be prevented by techniques as simple
as specifying different connectors.  As a result, LMSC has, on several
occasions, been able to avoid rejecting a partially overlapping PAR by
chartering multiple working groups to publish incompatible solutions that are
targeted at different applications that happen to have similar problems  With
wireless media the situation is totally different:  Wireless media are shared
due to spatial proximity, even in cases where the application requirements are
disjoint.  There is currently a significant attempt underway, initiated by
several major equipment suppliers, to cause adoption of a single, worldwide
standard for wireless LAN communication in the 5GHz band.  At the
just-completed meeting 802.11 voted in favor of participating in this effort.
It is going to be hard enough to achieve the compromises and approvals
necessary to converge the protocols developed for a common purpose, without the
problems of either attempting to incorporate the requirements of a
significantly different protocol intended to serve a different application (or
the uncertainty among the WLAN standards bodies as to the usability of the
result of their convergence in a channel shared with a MAN access protocol).
This is a (reasonably) minor problem as long as MANs in the 5GHz band remain
proprietary solutions proposed by a few parties.  This will likely become a
major problem if IEEE 802 LMSC endorses the conflicting use of the 5GHz band.

Michael Fischer                    voice: +1-210-614-4096 x107
CHOICE Microsystems, Inc.          fax:   +1-210-614-8192

------------------------------------------------------------------

I agree with Steve Shellhammer's comment on the need for a very comprehensive
coexistance effort between .11, .15, & .16 on both the 2.4-2.5GHz  and 5Ghz
bands. The new .16 PAR states "It will focus on the 5-6 GHz range and may be
applied to unlicensed bands between 2 and 11 GHz.". If the system will operate
in the ISM band, customers will have to choose 1 system in a given physical
location (given todays coexistance mechanisms). This will be a major problem in
public space systems, where providers will probably not be able to enforce
client upgrades, and yet will be expected to support both the older clients (ex.
.11b and .11a) and the newer clients (ex. .16).

One plus I see is that a serious coexistance effort would be very useful in
presenting a consolidated position to the WRC meeting, which I belive will be
considering global assignments in the 5GHz band.

Kevin M. Barry
S.I.T.A.

------------------------------------------------------------------

Naftali is right!  There are no options to turn the band into the IEEE band;
it's not going to happen.  There is no FCC rule that says only IEEE protocols
will happen in the ISM or the UNII bands.  Lots of times I have heard the
argument that wireless LAN companies want to get some profits under their belts
before they pursue another standard and that the market is going to get
confused.  That is a false sense of security.  IEEE is  conceding all the new
wideband in 2.4GHz to proprietary solutions, let's pull together a better set of
protocols than anybody else.  The IEEE companies will lose out in the long run
unless we act.

Richard H. Paine

------------------------------------------------------------------

To all our readers ...

It is very interesting to see the various viewpoints that are occurring
regarding this PAR. Please continue to have these discussions. It is
invaluable to have this action for our November meeting position from 802.11
WG. I as chair must reflect this to the ExCom.

Please MAKE SURE you send you comments directly to the reflector so all of
us can see.

We should plan to make a formal position statement to the 802.16 Chair and
his WG, on the first day of our next Plenary meeting. Also, to discuss the
issues within our full membership, and let him with his WG. Under the
con-operation rules of 802.

I do think there is a more fundamental issue here too. That is the overall
co-existent, and positioning of the Wireless groups under 802. This I have
seen raised in some of the replies.

Any volunteers to lead this important effort. (Don't all rush at once
please).

Stuart
Chair IEEE 802.11