Transcript A Smart Decision Model for Vertical Handoff

A Smart Decision Model for
Vertical Handoff
Ling-Jyh Chen*, Tony Sun*, Benny Chen*,
Venkatesh Rajendran†, Mario Gerla*
Department of Computer Science, University of California at Los Angeles, Los Angeles, CA 90095, USA
† Department of Computer Engineering, University of California at Santa Cruz, Santa Cruz, CA 95064, USA
*
{cclljj, tonysun, cudokido, gerla}@cs.ucla.edu, [email protected]
Presentation Outline





Introduction
Related Work
Smart Decision Model
Smart Decision Example
Conclusion
Introduction
Problems and Solutions

Problem


Solution


Universal Seamless Handoff Architecture (USHA) and
Handoff Servers (HS).
Problem


Mobile devices with multiple network interfaces today cannot
perform handoff among devices without losing existing
internet connection—due to change of IP addresses.
Determining when to handoff to another interface is a
complex decision.
Solution

Smart Decision Model.
Definition of Handoff

Horizontal Handoff



Occurs when the user switches between different
network access points of the same kind.
e.g. Handoff among 802.11 APs.
Vertical Handoff


Involves two different network interfaces which
usually represent different technologies.
e.g. Handoff from 802.11 to 1xRTT (CDMA 2000).
Handoff Illustration
Seamless Handoff



Defined as a handoff scheme that maintains the
connectivity of all applications on the mobile device
when the handoff occurs.
Aims to provide continuous end-to-end data service
in the face of any link outages or handoff events.
Design Goal:


low latency
Minimal packet loss
Related Work
Related Work—
Handoff Decision-making

In “Policy-Enabled Handoffs across Heterogeneous
Wireless Networks,” Proc. of ACM WMCSA, 1999, by
H.J. Wang, R. H. Katz, and J. Giese:



Designed a cost function to decide the best moment and
interface for vertical handoff.
Cost functions presented in this paper is very preliminary
and not able to handle more sophisticated configurations.
Logarithmic functions used in the cost function will also have
difficulty in representing the cost value while the value of
the constraint factor is zero (e.g. the connection is free of
charge).
Related Work—
Handoff Decision-making

In “Cost Metrics For Decision Problem In Wireless Ad
Hoc Networking,” IEEE CAS Workshop on Wireless
Communications and Networking, 2002, by M.
Angermann and J. Kammann:


Modeled handoffs with HTTP traffic.
May have problems with other types of traffic, such as video
and audio streaming, where the bandwidth demand is much
higher than HTTP traffic.
Smart Decision Model
Testbed: Universal Seamless
Handoff Architecture (USHA)
NAT server
All packets are encapsulated
and transmitted using UDP
Applications are bound to the tunnel
and transparent to the handoff.
1xRTT
Smart Decision Model


With USHA, mobile hosts are able to select any
network interface for its connection at any time.
However, still need a model that knows which
interface to use based on various factors such as




Link Capacity (speed)
Cost
Power Consumption
Solution: Smart Decision Model
Smart Decision Model

HCC—Handoff Control Center


Provides connection between
network interfaces and the
upper layer applications.
Composed of




DM (Device Monitor)
SM (System Monitor)
SD (Smart Decision)
HE (Handoff Executor)
Smart Decision Model

DM (Device Monitor)

Monitors and reports the
status of each network
interface:




Signal strength
Link capacity
Power consumption
SM (System Monitor)

Monitors and reports system
information (e.g. current
remaining battery)
Smart Decision Model

SD (Smart Decision)


Integrates user preferences
and all other available
information provided by DM,
SM to achieve a “Smart
Decision”
HE (Handoff Executor)

Performs a device handoff if
current network interface
differs from the one
determined by SD.
Smart Decision Process

Priority Phase:
1.
2.
3.
4.

Add all available interfaces into candidate list.
Remove user specified devices from the candidate list.
If candidate list is empty, add back removed devices from
step 1.
Continue with Normal Phase.
Normal Phase:
1.
2.
3.
4.
Collect information on every wireless interface in the
candidate list from the DM component.
Collect current system status from SM component.
Use the score function to obtain the score of every wireless
interface in the candidate list.
Handoff all current transmissions to the interface with the
highest score if different from current device.
Priority / Normal Phases

Necessary in SD to accommodate user-specific
preferences regarding the usage of network
interfaces.


For instance, a user may decide not to use a device when it
causes undesirable interferences to other devices (e.g.
802.11b and 2.4GHz cordless phones).
With priority and normal phases in place, the SD
module provides flexibility in controlling the desired
network interface to the user.
Score Function



SD deploys a Score Function to calculate a score for every
wireless interface
Handoff target device is the network interface with the highest
score.
Score Function:




wj = weight of factor k
fj,i = normalized score of interface i of factor j
The equation is thus equivalent to:

where e = Expense, c = Link Capacity, p = Power
Consumption.
Score Function Breakdown

Expense:


Link Capacity:


Power Consumption:

Note:




The coefficients α , β , γ are determined by user preference.
Inverse functions are used to bound results from 0 to 1.
M = Maximum bandwidth requirement demanded by the user.
Link capacity is calculated using CapProbe—because advertised link
speed is seldom achieved due to link congestion or bad link quality.
Smart Decision Example
Smart Decision Example—Scenario


A mobile user currently using 1xRTT on his laptop
enters a café.
HCC immediately discovers a usable 802.11b access
point inside the café and conducts the following
comparisons:

Expense/Cost:



Link Capacity:



1xRTT: 1￠/min
802.11b: 10￠/min
1xRTT: 100 Kbps
802.11b: 5 Mbps
Power Consumption (battery time):


1xRTT: 4 hours
802.11b: 2 hours
Smart Decision Example—Result

The mobile user prefers to spend more time in the café and
feels that cost and connection speed are equally important to
her, thus


Coefficient obtained from her preference previously were:




αi = xi / 20 ,
xi :￠/min
βi = Min(yi, M)/M , M = 2Mbps
γi = 2 / zi ,
zi : hours
Scores calculated using the score function are:



wp=0.4, we = 0.3, wc = 0.3
S1xRTT = 0.83
S802.11b = 0.44.
Since S1xRTT > S802.11b , HCC decides to continue using the 1xRTT
interface.
Conclusion
Conclusion


Smart Decision Model provides a solution for
determining the right time to perform handoffs.
Our proposed model is able to make smart decisions
based on



Available network interfaces and properties (e.g. link
capacity, power consumption, and link cost).
System information (e.g. remaining battery).
User preferences.