Thread Links | Date Links | ||||
---|---|---|---|---|---|
Thread Prev | Thread Next | Thread Index | Date Prev | Date Next | Date Index |
Liang, Please see comments below. Regards, Ulises From: Liang Chen
[mailto:leo.liang.chen@gmail.com] Dear Ulises, It is my pleasure to open an interesting topic in 802.21 if it will
introduce some contributions to the working group. Your explaination on the discovery of new 802.11 AP sounds reasonable
to some extent. But how does an 802.21 STA in power management mode fisrt
discover an unknown 802.11 AP if it totally depend on the network
topology information? According to the DNA schemas proposed by IETF( http://pec.etri.re.kr/~yghong/draft-hong-dna-if-l2-00.txt),
the authors seem quite pessimistic on the make-before-break handover on
multiple interfaces with different link-layer technologies. I guess
802.21 will experience the similiar problem if STA wants to perform a
make-before-break handover without the help of network topology
information. So I am interested to what you think to the problem? [UO] One of the DNA author’s main assumptions is that “host does not have any information about a new link until a new link-layer connection is established”. However If the STA is an 802.21 capable STA/UE and it also supports multiple interfaces, then it is possible to get link layer event information from these interfaces before the HO has to take place. A mobility management application could use this information to establish a link layer connection on a suitable interface before the current link layer connection is dropped. Note that link layer event reporting is one of the main services provided by 802.21; handover execution and handover control however is outside the scope of 802.21. Best regards, Liang
On 11/9/05, Olvera-Hernandez,
Ulises < Ulises.Olvera-Hernandez@interdigital.com> wrote:
Liang, Thanks for the good discussion. Please see comments below. Regards, Ulises From: Liang Chen [mailto: leo.liang.chen@gmail.com]
Dear
Ulises, Plese
check my in-line reply. Thanks! -Liang On
11/9/05, Olvera-Hernandez, Ulises
<
Ulises.Olvera-Hernandez@interdigital.com> wrote: Dear Liang Chen, The model described in Section 6.3 is founded on some
assumptions (such continuous network topology update, either manual or
automatic) and it addresses 802.21 Information Service only. The use case that
you described below is supported in principle, although the initial AP
discovery procedure might relay on 802.21 Event Service and Command service and
on certain degree of scanning. I feel
the network topology update is a big problem, because 802.11 AP can not locate
itself in the existing network topology map with the reference of other
networks unless the 802.11 AP installs other radio to collect the topology
information. [UO] The AP is discovered by the STA(s) that attempts to
associate to it. If the STA is 802.21 capable this information can be sent to
the MIIS over the existing radio connection. Note that this would only be
required the first time the STA discovers the new AP.
If 802.21
STA could detect the newly deployed AP before handover, the network topology
information is not necessary. [UO] Only the initial
discovery is required. Once the network topology map is updated, relevant
information might be delivered over the current connection ( e.g., a cellular
network). The benefit becomes apparent when the population of STAs attempting
to access the AP increases (e.g., enterprise environment) . This definitely
adds value, in particular considering this specific use case that we are
discussion where the AP has no 802.21 functionality and it is completely
oblivious to its surroundings. However,
it might introduce another problem if we want to support the make-before-break
handover. [UO] Could you please
elaborate on the problem you see? Does
802.21 has any proposal for the problem? [UO] 802.21 does not
address handover policies. Therefore whether a handover is executed using MBB
or BBM is up to the application controlling the handover. 802.21 simply
provides services that allow both modes of operation.
|