Re: [802.21] 答复: [802.21] question on MIH event subscri ption
Junxiang,
guojunxiang wrote:
> Hi Qiaobing,
> What you dig out is really a big issue and the solution sounds good. But I
> am wondering if the MIH user forgets to subscribe some events, maybe
> handover cannot be completed since the handover flow consists of quite a few
> handover messages(and their corresponding .indication events), lack
> subscription of any one of these events will cause the handover flow to be
> aborted.
Yes, I agree with you analysis. But as a matter of fact, no standards
can prevent people from making design mistakes. And if the MIH User code
contains bugs (forgetting to subscribe some necessary events), things
will break for sure and there is really little *we* can do.
regards,
-Qiaobing
>
>
> Best Regards,
> Junxiang Guo
> -----邮件原件-----
> 发件人: Qiaobing Xie [mailto:Qiaobing.Xie@MOTOROLA.COM]
> 发送时间: 2007年3月2日 5:05
> 收件人: STDS-802-21@listserv.ieee.org
> 主题: Re: [802.21] question on MIH event subscription
>
> Hi, Vivek,
>
> Gupta, Vivek G wrote:
>> Hello Qiaobing,
>>
>> I am fine with your proposal.
>> So should MIH_MN_HO_Candidate_Query.indication (and all such other
>> indication messages) be included in Table-4 as well?
>
> Yes. Right now all the MIH events captured in Table-4 a pass-thru link
> events; None of the MIH events originated *within* MIHF (including
> MIH_MN_HO_Candidate_Query indication and all such other indications) are
> captured in Table-4 and I firmly believe they should be captured in Table-4.
>
>> Also even though this is an MIH event, this is a local-only MIH event
>> (one that would never go over the medium). So we don't need to define
>> any MIH protocol messages for these events.
>
> Precisely! But they should follow the same subscription/notification model
> as the link-pass-thru MIH events.
>
> regards,
> -Qiaobing
>
>> Kind Regards
>> -Vivek
>>
>>> -----Original Message-----
>>> From: Qiaobing Xie [mailto:Qiaobing.Xie@motorola.com]
>>> Sent: Thursday, March 01, 2007 12:45 PM
>>> To: Gupta, Vivek G
>>> Cc: STDS-802-21@listserv.ieee.org
>>> Subject: Re: [802.21] question on MIH event subscription
>>>
>>> Hi, Vivek,
>>>
>>> ...
>>>> However consider the case of following primitive:
>>>> MIH_MN_HO_Candidate_Query.indication
>>>>
>>>> It is an example of MIH_xxx.indication type of primitive, however I
>> am
>>>> not sure if this is the case of an MIH event. As such MIH Users may
>> not
>>>> need to subscribe for this.
>>> When a MIH_MN_HO_Candidate_Query request message arrives from MIHF_A
>> to
>>> MIHF_B, the arrival of the message will trigger a
>>> MIH_MN_HO_Candidate_Query indication from MIHF_B to its MIH User.
>>> This part seems quite clearly defined in the spec now.
>>>
>>> But what if there are multiple MIH Users above MIHF_B? They may be
>> very
>>> different applications and it is possible that not all of them are
>>> interested in getting (or designed to handle)
>> MIH_MN_HO_Candidate_Query
>>> indications.
>>>
>>> This can be neatly solved by requiring subscription to receiving
>>> MIH_MN_HO_Candidate_Query indications.
>>>
>>> Of cause, the downside is that if no one at MIHF_B subscribes for
>>> MIH_MN_HO_Candidate_Query indication, the message is going to be
>>> basically ignored by MIHF_B. But that seems ok - just as if some one
>>> sent the message to a wrong destination.
>>>
>>> regards,
>>> -Qiaobing
>>>
>>>> Best Regards
>>>> -Vivek
>>>>
>>>>> -----Original Message-----
>>>>> From: stds-802-21@ieee.org [mailto:stds-802-21@ieee.org] On Behalf
>> Of
>>>>> Qiaobing Xie
>>>>> Sent: Thursday, March 01, 2007 9:40 AM
>>>>> To: 802.21 List
>>>>> Subject: question on MIH event subscription
>>>>>
>>>>> All,
>>>>>
>>>>> I think the following statements should be true:
>>>>>
>>>>> - any MIH .indication primitive shall be considered an MIH event
>>>>> - an MIH User shall be notified about an MIH event only that User
>> has
>>>>> explicit subscription for the event
>>>>>
>>>>> Therefore, we will need to accept the following scenario as
>> perfectly
>>>>> normal:
>>>>>
>>>>> 1. An MIH User invokes MIH_Configure_Link and sets some threshold
>> in
>>>>> order to get a link report when that threshold is crossed;
>>>>>
>>>>> 2. But he forgets to (or intentionally does not) explicitly
>> subscribe
>>>>> MIH_Link_Parameters_Report.indication event;
>>>>>
>>>>> 3. When the event happens, the User will not get the indication.
>>>>>
>>>>> My reasoning is that one may want to build a User whose
>> responsibility
>>>>> is just to set link thresholds but the responsibility for receiving
>>>> and
>>>>> processing the events are delegated to a different User (or Users).
>>>> The
>>>>> latter will subscribe for the event but not the former.
>>>>>
>>>>> regards,
>>>>> -Qiaobing
>