Announcement of the new work : Link Characteristic Information for Mobility

Soohong Daniel Park <soohong.park@xxxxxxxxxxx> · Wed, 10 May 2006 11:03:03 +0900

Hi

The following is the work description of LCI (Link Characteristic Information for Mobility). Jouni Korhonen, Hannes Tschofenig and I are thinkig of having a new BOF in the IETF-66 on this subject (Target Area is TBD, but presumably, TSV or RTA&Infra). 

The problem statement is available via the link below before IETF repository.
http://daniel.vsix.net/lci/draft-korhonen-lci-link-characteristics-ps-00.txt

Tip: To illustrate what we are trying to achieve in conjunction with LCI, a simple flash demo is attached below. Look at the undesirable disruption of Non-LCI mobile terminal comparing with LCI mobile terminal (service quality is scalable) carefully. Note: It is JUST for your information, so please don't consider it seriously...:-)
http://daniel.vsix.net/lci/lci_concept.html

Our mailing list is http://eeca16.sogang.ac.kr/mailman/listinfo/lci 

==================================
Link Characteristic Information for Mobility (LCI)

Updated: 2006-05-10
Version: 0.9

Description:

Recently more and more mobile terminals are equipped with multiple 
interfaces for different L2 technologies. These mobile terminals make 
it possible to communicate through different wireless networks at 
the same time, or allow the most appropriate interface to be selected 
according to current conditions. In the latter case, transitions 
between heterogeneous links (vertical handovers) occur. Vertical 
handovers often cause an ongoing connection to experience sudden 
path characteristic changes (e.g. available bandwidth and delay). 

Although some transport protocols and application mechanisms provide 
congestion/flow control mechanisms, they are unable to detect and adapt 
quickly, and require to send a number of probes to determine the new 
network characteristics some time after the handover. The network 
capacity may have already been misused during the probing process, 
and the user experience can be disrupted. In some cases, handovers 
between the same type of links (horizontal handovers) may also lead 
to abrupt link characteristic changes, due to the different traffic 
loads on the old and the new networks. Moreover, even if handovers 
do not occur, the access link characteristics may change significantly 
due to the variations of the traffic load on current link. Both of 
these situations can lead to similar adverse effects as those on 
vertical handovers. 

As a matter of fact, the wireless access links are most likely the 
bottlenecks for wireless internet connections. Therefore, it would 
be ideal for mobile terminals to have the capability of sharing their 
access link characteristic information (LCI) with their relevant 
remote network nodes (including remote peers, mobility agents, and 
any other network nodes that may consider this information useful 
for optimizing network capacity usage and user experience). In case 
the bottleneck of a peer-to-peer connection locates in the middle 
of its path rather than its wireless access link (e.g. in the WLAN+ADSL 
access scenario, the ADSL link can be the bottleneck, instead of the 
WLAN), the access LCI would still be informational and the access LCI 
delivery mechansim can be extended to support path characteristics 
discovery. Sometimes, mobile terminals may have difficulties to obtain 
precise access LCI at any time, however, it is also important and 
heuristic to know the magnitude of change even without exact values, 
since this can act as a timely trigger to other mechansims at the 
relevant network nodes to re-investigate and renew their network 
capacity usage status. 

Existing IP mobility enabling technologies, however, do not provide a
method to deliver the access LCI to the relevant remote network nodes.
The principal objective of this work is to explore the possible
signaling solutions for delivering the access LCI at the IP layer or
above. Apparently, existing IP mobility protocols and transport protocols
could be extended to support this useful feature, while the potential
benefits and limitations need serious investigation. A new generic
lightweight signaling protocol may need to be designed for carrying the
LCI to tackle the limitations caused by using other protocol extensions.
Importantly, the adoptable LCI delivery mechanism(s) must be secured,
middlebox traversable, and must avoid significantly increasing the amount
of signaling traffic load, especially over wireless links. At the same
time, the tradeoff between the added LCI delivery and computation load
and gained advantages is also an issue that needs careful examination.

In multihoming scenarios, when multiple interfaces on the mobile terminal 
are used for one application for load sharing, it is desired that the 
LCI of each interface can be delivered simultaneously to the relevant 
remote network nodes. However, the methods of collecting the access LCI 
as accurate and timely as possible are out of the scope of this work.

The proposed work will also cooperate with the working groups that may 
consider the access LCI useful, in order to facilitate the LCI utilization 
by them. Especially, it is expected that the transport area may benefit
from the LCI delivery. It is also expected that real-time streaming 
services can be enhanced based on the availability of the LCI signaling. 
For example, SVC (Scalable Video Coding or H.264 Extended Profile) and 
BSAC-Bit Sliced Arithmetic Coding are designed to support a flexible 
control in terms of video and audio coder respectively following the 
receiver's network qualities, while their functions are limited at the 
moment due to the lack of dynamic signaling from the receiver when the
link characteristic changes.    

Goals:

- Produce "Link Characteristic Information for Mobility Problem 
  Statement" to describe the problem and limitation of the current  
  mobility solutions without link characteristic information delivery, 
  and clarify the motivation of designing the LCI signaling.

- Produce "Link Characteristic Information Description" to describe the 
  required link characteristic information for delivery.

- Evaluate a set of candidate proposals for Link Characteristic Information 
  Delivery (probably multiple documents required).

- Produce "A lightweight signaling protocol for carrying Link Characteristic 
  Information" to design a new signaling mechanism for carrying Link 
  Characteristic Information including middlebox traversal and security 
  soluctions.