RE: Higher layer requirem ents for IETF: IS per technology
Michael,
These are great thoughts and in general this seems like a good line of
thinking. Yes, I have not seen any new issues raised either as well..
As for IEs, yes initially a client may just want to know list of
networks available and other high level information which is completely
media independent (location, etc.). This part can be a very small list
of IEs.
Thereafter there could be a basic set of IEs specific for each access
network which could provide specifics w.r.t high level details of that
access network (security, QoS, operator, cost, data rate, etc.) and this
could be followed by more detailed neighbor reports, etc. As for
neighbor reports IMO 802.21 would be better off defining a new media
independent format which has all relevant information from handover
perspective and can also apply to different access technologies.
Also as long as there is a way to retrieve individual IEs (or even a
group) through some TLV based mechanism, aspects related to specific
grouping of these IEs do not seem that important.
Best Regards,
-Vivek
-----Original Message-----
From: stds-802-21@ieee.org [mailto:stds-802-21@ieee.org] On Behalf Of
Michael.G.Williams@nokia.com
Sent: Monday, August 22, 2005 8:39 PM
To: STDS-802-21@listserv.ieee.org
Subject: Higher layer requirem ents for IETF: IS per technology
Colleagues,
Some thoughts to narrow the discussion...
Regarding scenarios/use cases and reference model...
This seems to be a topic that won't die :) At least let's hope no one
brings up tight vs. loose coupling again!
What is the issue, the list of technologes to support? Isn't it 3GPP,
3GPP2, .3, .11, .15, .16, and if they want .20?
Which combinations are supported in handovers?
We agreed not to favor one handover algorithm over another so
network or device can initiate, w/ any mobility protocol.
We agreed not to worry about intra technology handovers, but
otherwise to facilitate handover between any supported types. So any
network type can be the source for IS.
The location of the MIH function serving the IS can be more than
a link away and also might be accessed by passing through legacy devices
without MIH function, so we must support L3+ access in addition to L2.
Where does the MIH function live in the system? I thought we agreed in
any device. I can certainly see the use for it in VPN gateways, FW's,
NATs, routers, wireless PoAs (e.g. BS, AP etc), wired switches, mobile
devices, etc.
But this is all old news. Have new issues been raised here?
Colleagues,
Regarding what is being handed over...
Colleagues,
Regarding an IS basic schema per-technology...
Hopefully there will be a breadth of IS IE's which are common across
media types, which will be a core value of the standard, along with
providing the semantic agreement across the media types, and
standardizing the organizational criterion that no terms are to be
reused with differing semantics.
It doesn't surprise me if the basic set for each technology might have
*some* differences from the other technologies. Not sure why that would
be a problem actually? This could be split into basic MI (independent)
and basic MD (dependent) to accommodate that need.
"Basic" would mean that set of semantic agreements and terms that are
guaranteed to be in the MIH capable device. Basic MI would be common to
all media; Basic MD would be for IS IE's different per media type. Basic
MD could be required to use unique terms.
One of the hopes for the XML approach was the ease of updating the
mobile device on the fly (after initial deployment at any rate) This is
certainly true for the extended set, but why not for the basic set as
well?
If so, maintaining separate basic MD sets for each technology would be
small but nice benefit since updating one (e.g. the .11 basic MD set)
wouldn't have to affect the others (e.g. the .3 basic set)?
.21 standard would require the base semantic values which are common
across multiple basic sets are the same, and to belong to the basic MI
set.
This approach might fit well if there were the desire for 802.11u for
example to define IE's unique to .11 yet need to be in every .11u
device.
What do you think?
Best Regards,
Michael
ss