Transcript dynamics_ipcn2000_presentation

Increasing communication availability
with signal-based mobile
controlled handoffs
D. Forsberg, J.T. Malinen, J.K. Malinen, H.H. Kari
TSE-Institute
Telecommunications and Software Engineering
Laboratory of Information Processing Science
Helsinki University of Technology
Finland
HELSINKI UNIVERSITY OF TECHNOLOGY
Increasing communication availability with signal-based mobile controlled handoffs
Presentation overview
Introduction
Wireless Mediapoli
Dynamics - HUT Mobile IP
Motivation
Criteria
Improving communication
availability with Mobile IP
Policy based mobility agent selection
and detection with prioritization
Signal quality awareness
Two phase handoff
Signal quality awareness
802.11 link layer modes
Policies and configurability
Two phase, forward and mobile
controlled soft handoff
Enables glitchless handoffs
Transparent to the mobile user
Tests
Handoff times
Packet loss per location update
Configurable monitor testing
Data throughput
Conclusions
Future work
Dynamic and automatic policy
configurations
Dynamic advertising of access network
services and capabilities
Questions?
HELSINKI UNIVERSITY OF TECHNOLOGY
Increasing communication availability with signal-based mobile controlled handoffs
Introduction
• communication availability problem
– CA is readiness for usage
– we solved the problem under signal quality based handoff
management with Mobile IP in WLAN
• developed system is currently in use and evolving
– Wireless Mediapoli
http://www.mediapoli.com/wireless
– Dynamics - HUT Mobile IP
http://www.cs.hut.fi/Research/Dynamics/
• this presentation assumes that you are familiar with
basic mobile IP
HELSINKI UNIVERSITY OF TECHNOLOGY
Increasing communication availability with signal-based mobile controlled handoffs
Motivation
• clients require mobility in the Internet
• wireless mobile computers and portable digital
assisstants (PDA) are becoming more common
• increased use of multimedia content with mobile
computers (mp3 streams, real video and audio etc.)
– Mobile computers are powerful enough and WLANs have
enough bandwidth to deliver the content
• mobile users want to roam between different link
layer technologies
– Our environment is built on top of 802.11 and Ethernet
• user wants to control the communication parameters
– The user may want to use different service providers based
on his needs and current resources
HELSINKI UNIVERSITY OF TECHNOLOGY
Increasing communication availability with signal-based mobile controlled handoffs
Criteria
•
•
•
•
mobility should not affect the data transmission
tolerance for congestion
efficiency
the mobile node should use the mobility agent that
offers the best communication availability
– the word best may not mean the same between two
different mobile users (cost, bandwidth, services etc.)
• independence of the underlying radio technology
– roaming between different wireless networks
HELSINKI UNIVERSITY OF TECHNOLOGY
Increasing communication availability with signal-based mobile controlled handoffs
Improving communication
availability with Mobile IP
signal quality awareness
- every available MA gets its own
priority based on the signal quality
(SNR) value
two phase, soft handoff
policy based mobility agent selection
and detection with prioritization
- MN can hear possibly many MAs
- priority modification techniques
HELSINKI UNIVERSITY OF TECHNOLOGY
Increasing communication availability with signal-based mobile controlled handoffs
Signal quality awareness
• comparisons and selections are based on dynamic
node priorities
• the signal quality can be seen as a meaningful
distance for the mobile node between the access
point and mobile node
– Meaningful in the sense that the communication
availability is highly dependant on the signal strength in
the communication path
• priorities are based on signal quality values
received from the mobility agents
– In our system the mobility agent is running in an access
point and thus the advertisements have different signal
quality source
HELSINKI UNIVERSITY OF TECHNOLOGY
Increasing communication availability with signal-based mobile controlled handoffs
Policies and configurability
• policy is a set of rules that affect the node
selection process in the mobile node
• four different policies implemented:
– default-policy: uses signal quality history and threshold
when selecting or comparing the current mobility agent
– newest-FA: the mobile node selects the most recently
detected mobility agent
– eager-switching: neither signal quality history nor threshold
is used during mobility agent selection
– early-expire: the mobile node expires the mobility agent
from the list of available mobility agents more eagerly
HELSINKI UNIVERSITY OF TECHNOLOGY
Increasing communication availability with signal-based mobile controlled handoffs
Policies and configurability
• configuration of parameters fine tune the handoff
procedures and policies
– average length: number of signal quality values taken into
the priority calculation. Average is calculated.
– threshold: mobile node will not change the current mobility
agent if the signal quality difference is below the threshold
– min-balance: do not use threshold if the signal quality goes
below this level
– expirepercent: degrade the mobility agent priority with this
percentage if the advertisement from the mobility agent is
old enough
• policies and configuration variables together form a
adjustable environment for mobile users
HELSINKI UNIVERSITY OF TECHNOLOGY
Increasing communication availability with signal-based mobile controlled handoffs
Hierarchical Mobile IP
Home Network
CN
HA
Internet
HFA1
SFA
FA1
Foreign Network
FA2
WLAN
FA3
FA4
Mobile Node
Mobile Node
HELSINKI UNIVERSITY OF TECHNOLOGY
FA5
FA6
Increasing communication availability with signal-based mobile controlled handoffs
Two phase, forward and mobile
controlled soft handoff
• enables glitchless handoffs
• transparent to the mobile user
MN sends location
update request to the
new FA
1.
SFA changes the downstream
route and sends the location
update request reply
every FA between the LFA
and the SFA prepares the
downstream route for the MN
phase I
HELSINKI UNIVERSITY OF TECHNOLOGY
2.
MN receives the reply and
changes the upstream
route to the new FA
3.
every FA between the SFA
and the LFA prepares the
upstream route for the MN
phase II
Increasing communication availability with signal-based mobile controlled handoffs
Tests
• tests
–
–
–
–
handoff times
packet loss per location update
configurable monitor testing
data throughput
• forced location updates
• handoff latency measured in the MN
• data throughput measured with netperf
– maximum TCP throughput and paced UDP streams (about
1.4 Mbps)
HELSINKI UNIVERSITY OF TECHNOLOGY
Increasing communication availability with signal-based mobile controlled handoffs
Handoff times in hierarchy
CN
HA
Location update latencies
for some transitions
OLD
FA
FA11
FA13
FA31
HFA
FA11
FA1
FA13
FA3
FA13
FA31
FA29
FA12
FA2
FA14
FA32
HELSINKI UNIVERSITY OF TECHNOLOGY
FA15
NEW Average
FA
in ms
FA12 19,1
FA14 30,4
FA32 41,4
Increasing communication availability with signal-based mobile controlled handoffs
Configurable monitor testing
Setting 1
Threshold
50
Min-balance
10
Old-FA-factor
50
Worst-min-time
50
Worst-max-time
10
Average-length
1
Early-expire
OFF
Newest-FA
OFF
Eager-switching
ON
Lost Packets
Location updates
Setting 2
1
13
50
50
10
3
OFF
OFF
OFF
Plain MIP
2179
8
HELSINKI UNIVERSITY OF TECHNOLOGY
• lost packets and location
updates measurement with
different settings of the
monitor
• testbed emulated with signal
quality environment
recorder/replayer
• packet dropping based on
low signal quality levels
Setting 1
66
63
Setting 2
117
9
Increasing communication availability with signal-based mobile controlled handoffs
Lost packets/ location update
Data stream:
100kB/s, 1kB packets
CN
HA
HFA
FA11
FA13
FA3
FA31
FA12
FA29
FA14
OLD
FA
NEW
FA
Lost packets/
update
FA11
FA31
FA29
FA31
FA12
FA15
FA32
FA13
FA31
FA29
FA32
FA13
FA15
FA31
FA11
FA12
0,00
0,00
0,00
0,00
0,00
0,03
0,07
0,10
FA15
FA32
Data stream
CN --> MN
OLD
FA
NEW
FA
Lost packets/
update
FA11
FA31
FA29
FA31
FA12
FA15
FA32
FA13
FA31
FA29
FA32
FA13
FA15
FA31
FA11
FA12
0,27
0,27
0,00
0,15
0,14
0,00
0,00
0,00
Data stream
MN --> CN
HELSINKI UNIVERSITY OF TECHNOLOGY
Increasing communication availability with signal-based mobile controlled handoffs
Data throughput
1,4
1,2
Mbps
1,0
0,8
0,6
0,4
TCP
UDP
0,2
0,0
0,1
1,0
10,0
locupds/s
HELSINKI UNIVERSITY OF TECHNOLOGY
100,0
Increasing communication availability with signal-based mobile controlled handoffs
Conclusions
• the mobile user can change the policy and handoff
parameters dynamically without disturbing the
communication sessions
• with soft handoffs neither buffering nor multicasting
is required to achieve seamless handoffs
• the solution is not dependent on the handoff
management below the network layer
• the prioritiy-based FA comparison is feasible
because it is not bound to the signal quality values
and, thus, not only to the WLANs
HELSINKI UNIVERSITY OF TECHNOLOGY
Increasing communication availability with signal-based mobile controlled handoffs
Conclusions
• signal quality awareness is a simple way to improve
the communication availability without extending
mobility protocol
• the tests showed that hierarchical Mobile IP with
signa quality awareness and two-phase handoff
supports micro mobility
• signal quality awareness is a simple way to improve
the communication availability without extending
mobility protocols
HELSINKI UNIVERSITY OF TECHNOLOGY
Increasing communication availability with signal-based mobile controlled handoffs
Future work
• scalability with multiple mobile nodes under the same
FA hierarchy and an HA should be analysed
• hard handoff management is required whenchannel
switching occurs in an ad hoc mode WLAN. Its use
with the presented system should be evaluated.
• dynamic parameters in tha FA and agent
advertisements to support mobility agent selection in
the MN should be studied.
HELSINKI UNIVERSITY OF TECHNOLOGY
Increasing communication availability with signal-based mobile controlled handoffs
Questions?
D. Forsberg, J.T. Malinen, J.K. Malinen, H.H. Kari
Email
{dforsber, jtm, jkmaline, hhk}@cs.hut.fi
WWW
http://www.cs.hut.fi/Research/Dynamics/
HELSINKI UNIVERSITY OF TECHNOLOGY