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

Re: [802.21] SAP semantics



Title: Re: [802.21] SAP semantics
Salut Ajay,

Here are my personal opinions...

I think the way the text is written could be misleading. The problem is the second sentence with the use of the word “similarly”. It leads to the mix-up of the different notions of SAP and protocols.

When the SAP is described, we should focus on description of services accessible at the SAP. It is not the moment to describe how services are supported (by various functions and protocols), which should be specified in a different section. Saying more than necessary, though not necessarily incorrect, is always at the expense of clarity. A standards specification is concise and not verbose.

“MIHF entity may also send commands to another remote (peer) MIHF entity” is not wrong. This is a feature of the MIH protocol, it is not  a description of the SAP.

“Upper layers may directly send commands to MIHF” is a description of a feature of the MIH SAP. (The word “directly” is redundant and could only be misleading. “MIH-user” is defined and is best used systemically rather than “upper layers”.) At the SAP, an MIH-user can send a command to the MIHF entity (I never understand why we name it MIHF entity rather than a straightforward MIH entity). It is the MIHF entity which determines how to support that command. The word “command” here refers to a service.

The specification of the MIH functions (in the MIHF entity) describes how the MIH SAP is supported. In case that by supporting the command by the MIH-user, a remote MIHF entity needs to be contacted (for whatever reason), the “MIHF entity may also send commands to another remote (peer) MIHF entity” (it should be written as “MIHF entity will send commands to its peer remote MIHF entity for further processing (something like that)”. The word “command” here refers to a protocol function.

OK, to make it short, in that text, the second sentence is just not necessary.

Cheers,
Hong-Yon



From: Ajay Rajkumar <ajayrajkumar@lucent.com>
Organization: Bell Labs, Lucent Technologies Inc.
Reply-To: <ajayrajkumar@lucent.com>
Date: Wed, 06 Sep 2006 15:40:51 -0400
To: Hong-Yon Lach <hong-yon.lach@motorola.com>
Cc: <STDS-802-21@listserv.ieee.org>
Subject: Re: [802.21] SAP semantics

Hong-Yon,

I agree that SAP is defined to provide an abstraction between a service provider entity and a user entity.

Also, MIH protocol would be used to provide communication between the two MIHF entities.

The problem arises from the current text in the draft in Section 5.6

"Upper layers may directly send commands to MIHF. Similarly MIHF entity may also
send commands to another remote (peer) MIHF entity. Primitives corresponding to all these services
described above are within the scope of MIH_SAP."

This implies that MIH_SAP is being used both by the MIH User as well as MIHF to communicate with the remote entity, which is my view is incorrect.

Regards,
-ajay

Hong-Yon Lach wrote:
Re: [802.21] SAP semantics Salut all,
 
A SAP is defined to provide an abstraction of service between a service provider entity and its user entities in the local system. A protocol provides a specification of interactions and operations between “peer” entities in different (sometimes virtual) systems with well-defined PDUs exchanges over a communications transport.
 
If a local MIH-user needs a remote services by a remote MIHF, it makes its request to the local MIHF through the MIH SAP. The local MIHF determines that it needs the support of a remote MIHF, it thus performs the necessary operations with the remote MIHF using the MIH protocol. Upon the end of the MIH protocol operation, the local MIHF provides a response to the MIH-user via the MIH SAP.
 
An entity can initiate the execution of its function and/or protocol by many different triggers: timeout of a timer, request by its users, detection of certain system conditions, reception of a PDU from its peer, etc. The entity can initiate its function and/or protocol without going through its SAP. A SAP is not the only means to have its serving entity to initiate a function and/or protocol.
 
I hopes this addresses your concerns.
 
Cheers,
Hong-Yon
 
 
Message Classification:
[ ] General Business Use Only
[ ] Motorola Internal Use Only
[ ] Motorola Confidential Proprietary
 
Hong-Yon Lach
Lab Manager, Edge Mobile Networking Lab (EMNL)
Office: +33 (0)169352536; Mobile: +33 (0)607590268
 

 

From: Cheng Hong <Hong.Cheng@SG.PANASONIC.COM> <mailto:Hong.Cheng@SG.PANASONIC.COM>
 Reply-To: Cheng Hong <Hong.Cheng@SG.PANASONIC.COM> <mailto:Hong.Cheng@SG.PANASONIC.COM>
 Date: Mon, 4 Sep 2006 15:40:35 +0800
 To: <STDS-802-21@listserv.ieee.org> <mailto:STDS-802-21@listserv.ieee.org>
 Conversation: [802.21] SAP semantics
 Subject: Re: [802.21] SAP semantics
 
 Hi Adrian and all,
 

 What you described is generally true. However, with regard to the original question, I don't think it is necessary to have two separate SAPs.
 

 The question regarding the remote MIH communication is different from a local use case. In a remote scenario, conceptually, the local MIHF becomes a user of the remote MIHF. The command is sent via the MIH Protocol (either via L2 or L3), and should be processed by the remote MIHF. In this case, there should be no problem for the remote MIHF to make use of the MIH_SAP for the commands come in via the MIH Protocol, since functions supported over the MIH protocol should not be different from that of the local MIHF. Therefore, it is rather an optimized design.
 

 cheers
 

 Cheng Hong
 

 
  

 
 

From: stds-802-21@ieee.org  [mailto:stds-802-21@ieee.org] <mailto:stds-802-21@ieee.org%5D>  On Behalf Of Stephens, Adrian  P
 Sent: Monday, September 04, 2006 3:09 PM
 To: Phillip  Barber; Peretz Feder
 Cc: Andrea Francini;  STDS-802-21@listserv.ieee.org
 Subject: RE: [802.21] SAP  semantics
 

 
 
Hello All,
 

 
 
 
While Phillip's answer is correct,  I'd say it's not  optimal.
 

 
 
 
The SAP should group together all the primitives that are  required
 

 
for one instance of a particular entity to provide  services to an instance
 

 
of another entity.
 

 
 
 
If each entity is providing services to the other,   I'd respectfully
 

 
suggest that two SAPs should be  provided.   This should make the
 

 
description of the purpose of the SAP  clearer.
 

 
 
 
Also,  architecturally,  it also makes it  possible to reroute the
 

 
communications,  e.g. to insert a  new
 

 
entity that relates to just one SAP and has no effect on  the other
 

 
such as a new fragmentation/reassembly  sublayer.
 

 
 
 
 
 
Best Regards,
 

 
 
 
Adrian
 

  

 
 

From: stds-802-21@ieee.org  [mailto:stds-802-21@ieee.org] <mailto:stds-802-21@ieee.org%5D>  On Behalf Of Phillip  Barber
 Sent: 04 September 2006 07:07
 To: Peretz  Feder
 Cc: Andrea Francini;  STDS-802-21@listserv.ieee.org
 Subject: Re: [802.21] SAP  semantics
 

 
My quick answer is 'Yes'.
 
 
 
The SAP is just a focal point for  communications in the model.
 
 
 
While for any specific message exchange through  a SAP the roles of the participants are set, the roles of the participants  may change from message exchange-to-message exchange. In one exchange the MS  may be a service user. In another exchange the MS may be a service  provider.
 
 
 
Thanks,
Phillip Barber
Chief Scientist
Broadband Wireless  Solutions
Huawei Technologies Co., LTD.
 
----- Original Message -----
 
  

 From:  Peretz  Feder <mailto:pfeder@lucent.com>  
 
 To: phillip Barber <mailto:pbarber@BROADBANDMOBILETECH.COM>  
 
 Cc: Andrea Francini <mailto:francini@lucent.com>  ; STDS-802-21@listserv.ieee.org  
 
 Sent: Sunday, September 03, 2006 8:31  PM
 
 Subject: Re: [802.21] SAP  semantics
 
 
Phil, as the 802.16g SAP expert, what is your take on the  following?
 
 Can an entity (MIHF 802.21 in this case) that provides services through a SAP (MIH_SAP)
also obtain services (transport and delivery of commands to a remote MIHF)
through the same SAP?
 
Andrea wrote: I always thought that at the two ends of a SAP the roles of service user and
service provider are fixed. Is this not the case? Can the same entity (the MIHF
in our case) be at the same time a service provider and a service user with
respect to the same SAP?
 
Peretz Feder
 
 
On 9/3/2006 12:56 AM, Andrea Francini wrote:
 
  
Hello Everyone,
 
I have a very basic question from reading the latest 802.21 draft
(P802-21-D01-09):
 
In section 5.6 (page 29, line 34 through 40) I find the following statement:
 
"The MIH_SAP and associated primitives provide the interface from MIHF to the
upper layers of the mobility-management stack. Upper layers need to register
with MIHF as users to receive MIHF generated events and also for link layer
events that originate at layers below the MIHF but may be passed on to upper
layers through MIHF. Upper layers may directly send commands to MIHF. Similarly
MIHF entity may also send commands to another remote (peer) MIHF entity.
Primitives corresponding to all these services described above are within the
scope of MIH_SAP."
 
The statement indicates that MIHF provides services to a number of upper layers
(MIH users) through the MIH_SAP. Then the statement also seems to indicate (the
wording is not totally explicit) that MIHF can use MIH_SAP to send commands to a
remote MIHF entity.
 
My question is about the general SAP semantics:
 
Can an entity (MIHF in this case) that provides services through a SAP (MIH_SAP)
also obtain services (transport and delivery of commands to a remote MIHF)
through the same SAP?
 
I always thought that at the two ends of a SAP the roles of service user and
service provider are fixed. Is this not the case? Can the same entity (the MIHF
in our case) be at the same time a service provider and a service user with
respect to the same SAP?
 
Thanks a lot,
 
Andrea