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

Re: [STDS-802-16-MOBILE] [Handoff] Some additional thoughts on Handover process



Title: [STDS-802-16-MOBILE] [Handoff] Comments / contribs deferred/rejected/withdrawn in Shenzhen to be considered by handoff adhoc
See my comments in-line.
 
Thanks,
Phil
 
----- Original Message -----
From: Mary Chion
Sent: Tuesday, June 08, 2004 10:29 AM
Subject: Re: [STDS-802-16-MOBILE] [Handoff] Some additional thoughts on Handover process

Hi All,
 
I have the following thoughts on HO process:
 
1. Since we are considering the possibility of allowing SHO and fast BS switching HO, we adopt an acitve set maintenance procedure regardless of the HO type (BBM, MBB Hard HO, SHO or fast BS swiitching).
Accomodating these features can be as easy as a few flags in the RNG-RSP.
 
2. I propose to change the current HO process defined in 16e to make it more deterministic. With the current definition, the HO decision point could be at the BS or the MSS, and any of the messages maybe or may not be sent during a HO. I think this allows for too many possibilities, it's hard to implement and maintain a state machine with so many undeterminstic events. When operating and optimizing a larget network, it would be easier to have a more defined HO procedure. I propose we do the following:
a) The HO decision point should be at the BS. The MSS can assist in the HO process by reporting a list of recommaned target BS, but the BS directs to the MSS to handover to a specific active set (the list of serving BS).
Absolutely not.  I agree that it would be nice to simplify the option sets, but not by sacrificing core functionality.  Distributed decision topologies will be the lowest cost, first systems deployed.  Mandating BS control of HO decision guarantees higher cost and complexity--and longer development cycles.  Ultimately, the more complex systems with their higher functionality will prevail.  But smaller, early deployments should not be constrained.  In fact we should do everyting possible to increase the likely successful solution models, especially early penetration models, not restrict the solution set to our own preferred models. 
b) There is a minimum pair of messages have to be sent during a HO: BS Handover Command (or if we don't want to add a new message, we can just use BS Handover Request to direct the MSS) and MS Handover Indication. When the BS decides a HO should take place and chooses the new active set, the BS should send BS Handover Command  to the MSS. When the MSS has performed the HO, the MSS should send MS Handover Indication to indicate the status of the handover. The MSS can use this message to Reject the HO if it has to.
HO information and decision messaging needs some clean-up.  But I don't believe wholesale change is required.
c) For the MSS assisted HO, the HO can be triggered by the MSS sending MS Handover Request. However, this message doesn't have to result in HO. The BS should send BS Handover Response just to stop possible retransmission of MS Handover Request message. For BS triggered HO, the BS just need to send BS Handover Command to start the HO process.
You can't stop MSS HO.  A BS can always refuse to allow a MSS to HO.  And an MSS can go right ahead and drop that Serving BS and commence network re-entry with a Target BS, with valid MSS context collected from the Serving BS.  And there is nothing the Serving BS can do about it.  So BS ability to restrict MSS HO is marginal at best.  We reflect that reality in our current HO processes.
d) I believe no matter what type of HO we are supporting, we should use the same HO process as defined above with possiblity different parameters included in the message. i.e. For a HO with Level 3 backbone communication, the BS Handover Command message can indicate skipping certain steps when the MSS performing network entry at the target BS (maybe just ranging necessary).
Forget Serving BS telling MSS what type of HO to expect at the Target BS.  The Target BS can tell the MSS itself in RNG-RSP.  Better recovery mechanism regardless.
 
Just some of my thought. Thanks.
 
Best Regards,
 
Mary