Computer Science Technical Report Archive

USC Computer Science Technical Reports, no. 871 (2005)

(USC DC Other)

USC Computer Science Technical Reports, no. 871 (2005)

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content A P2P Simulation Model to Support Mobile, Scalable Nearest Neighbor Queries
for Location-based Services
∗
Wei-Shinn Ku, Roger Zimmermann, Chi-Ngai Wan and Haojun Wang
Computer Science Department, University of Southern California, Los Angeles, CA 90089
[wku, rzimmerm, cwan, haojunwa]@usc.edu
Abstract
With the increasing capabilities of mobile devices, there
has been a growing interest in location-based services. Here
we present MAPLE, a scalable peer-to-peer nearest neigh-
bor (NN) query simulation system for mobile environments.
MAPLE is designed for the efficient sharing of query re-
sults cached in the local storage of mobile peers. The most
innovative aspect of the MAPLE model is its ability to ei-
ther fully or partially compute location-dependent nearest
neighbor objects on each host. Our novel verification al-
gorithms enable a mobile host to locally determine whether
the points of interest obtained from its peers are relevant
to its own NN query solution set. Our simulation system
emulates all aspects of the MAPLE model including the
movement of mobile hosts on a road network and the launch
and execution of location-based nearest neighbor queries.
The simulation illustrates how cooperative data sharing and
distributed processing among mobile peers results in a con-
siderable reduction of the load on remote spatial database
servers. Through its peer-to-peer topology, MAPLE ex-
hibits great scalability: the higher the mobile host density,
the more queries can be answered by peers.
1 Introduction
Ubiquitous and untethered information access is stim-
ulated by the growing capabilities of mobile devices.
Location-based services are especially of interest to many
users. An example query might be to “find the nearest
gas station from my current location.” The combination of
Global Positioning Systems (GPS), wireless communication
technology (e.g., 802.11x), and peer-to-peer overlays (P2P)
offers an exciting environment and opens unique opportu-
nities to provide location based services. By leveraging
ad-hoc networks, information can be shared in a P2P man-
ner among mobile clients to answer spatial searches (e.g.,
nearest neighbor queries). Importantly, in cases where ac-
∗
This research has been funded in part by NSF grants EEC-9529152
(IMSC ERC), CMS-0219463 (ITR), and equipment gifts from the Intel
Corporation, Hewlett-Packard, Sun Microsystems and Raptor Networks
Technology.
cess to the remote database servers is not always guaranteed
– such as during a natural disaster – P2P sharing can provide
a robust alternative where fault-resilience is naturally built
into the design.
Recently we designed a novel technique that utilizes pre-
vious query results cached in the local storage of mobile
peers to compute nearest neighbor query results of mobile
hosts [3]. We introduced a search algorithm termed Sharing-
based Nearest Neighbor (SN
2
) to locally verify whether
points of interest (POI) received from peers provide a com-
plete, partial or irrelevant answer to a submitted location-
based nearest neighbor query. If the POI search can be fully
satisfied from the peers, no access to any remote spatial
database servers is necessary, resulting in a decreased work-
load. Under some circumstances only a partial result can
be algorithmically verified (details are provided later) and
in that case, the query must be forwarded to a remote spatial
database server. However, the server load can be diminished
by utilizing the partial result to constrain the search space.
In this simulation system we present MAPLE (Mobile
scAlable Peer-to-peer nearest neighbor query model for
Location-based sErvices), a prototype of the design in [3]
that implements a sharing-based NN query model in con-
junction with a road network environment. In particular,
the MAPLE system exhibits the following distinguishing
characteristics:
• Sharing-based nearest neighbor query execution.
The MAPLE system performs and visualizes a novel
sharing-based nearest neighbor (SN
2
) query algorithm
among peers in a step-by-step manner.
• Scalability. The MAPLE system leverages P2P shar-
ing to achieve scalability in terms of the number of
peers and to reduce the access frequency to remote
spatial database servers. A higher density of peers
improves its efficiency.
• Realistic movement on road network. The move-
ment of hosts in the MAPLE system is constrained to
real world road networks. Mobile hosts in MAPLE au-
tonomously proceed on road networks and the velocity
of the movement is determined by the speed limit of
underlying road segments.
1
2 The MAPLE Infrastructure
Figure 1 illustrates the infrastructure of MAPLE. We are
focusing on mobile peers, such as cars, that are instrumented
with a GPS to provide continuous location information. Fur-
thermore, we assume that two tiers of wireless connections
are available on future automobiles. Traditional, cellular-
based networks (such as utilized by the OnStar service)
allow medium range connections to base-stations that inter-
face with the wired Internet infrastructure. A second type of
short-range networks allows ad-hoc connections with neigh-
boring mobile clients. Technologies that enable short range
communication include, for example, IEEE 802.11x. Ben-
efiting from the power capacities of vehicles, we assume
that each mobile host has a significant transmission range
and virtually unlimited lifetime. The architecture can also
support hand-held mobile devices. However, then power
consumption becomes an additional parameter which we
are not currently considering.
In our design, previous query results can be cached in the
local storage of mobile peers. Such peers move on road net-
works and autonomously launch sharing-based NN queries
or exchange cached query results with adjacent peers. The
SN
2
algorithm is then applied to verify whether data items
received from neighboring peers provide a complete, partial,
or irrelevant answer to the posed query. If only partial or
irrelevant data items are collected, the query is forwarded to
a spatial database server [4]. The complete query result is
then also cached in the local memory.
Data Station
Mobile Host
Mobile Host
Mobile Host
Mobile Host
Mobile Host
Service Area of a
Data Station
Spatial Database
Mobile Host
Transmission
Range
Mobile Host
Mobile Host
Mobile Host
Peer-to-Peer
Channel
Point-to-Point Channel Figure 1. The MAPLE infrastructure.
3 The MAPLE Components
In this section we describe the modular implementation
of the MAPLE system and the simulation parameter set such
as the points of interest and mobile host density.
3.1 Modules
The MAPLE system consists of three components: (1)
the multiple peer simulation module, (2) the server module,
and (3) the sharing-based nearest neighbor query visualiza-
tion module. Sharing-based NN queries are demonstrated
by visualizing the interactions and data exchange processes
among the peers on the road network.
• The multiple peer simulation module concurrently
models a predefined number of mobile hosts. It imple-
ments all the functionality of a single mobile host and
provides the communication facilities among peers and
from peers to remote spatial database servers.
• The server module is responsible for storing points
of interest indexed by an R-tree structure [2]. It per-
forms NN queries from peers with pruning bounds and
records the I/O load and access frequency of the spatial
database server.
• The sharing-based nearest neighbor query visual-
ization module provides a rendering of the verification
process of a sharing-based NN query in a step-by-step
manner. Users can arbitrarily select a mobile host and
launch a location-based NN query within the simula-
tion region.
3.2 Road Network Generation
We generated the underlying road network from the
TIGER/LINE street vector data set available from the
U.S. Census Bureau. The current MAPLE system stores the
road network of several Southern California counties. The
road segments are differentiated into several road classes,
such as freeways, primary highways, secondary and con-
necting roads, and rural roads. Road segments of differ-
ent road classes are associated with different driving speed
limits. Mobile peers monitor the speed limit on the road
that they are currently traveling on and adjust their veloc-
ity accordingly. One of the challenges when integrating
road segments into a complete road network is to isolate
paths that cross and determine if they indeed represent in-
tersections. For example, freeways generally project many
intersections in two-dimensional space, however, many of
them are over-passes or bridges. Our solution is to detect in-
tersection points with the help of their endpoint coordinates.
In addition, differing road classes allow us to distinguish
over-passes from intersections.
3.3 Interest Objects and Mobile Peers
MAPLE models the density of POIs (currently gas sta-
tions and restaurants) in the Greater Los Angeles area via
data available from two online sites: GasPriceWatch.com
1
and CNN/Money. MAPLE also imports vehicle statistics
of the Greater Los Angeles area from the Federal Statistics
web site to initialize the mobile host density. However, users
1
http://www.gaspricewatch.com
2
are not limited to utilizing these density presets. Many pa-
rameters – the density of interest objects and mobile peers
among them – can be changed via the multiple peer simula-
tion module.
3.4 Sharing-based Nearest Neighbor Queries and
Pruning Bounds
Within the system infrastructure shown in Figure 1, a
mobile hostQ collects NN data from peers to harvest these
existing results for completing its own k nearest neighbor
(kNN) search. If data items collected from peers only en-
sure a partial answer, then queries need to be forwarded to
the spatial database server to retrieve the complete result.
However, the partial result can be used to bound and hence
speed up the server search process.
There are two approaches to process NN information ob-
tained from peers. The single peer NN verification process,
also called kNN
single
, attempts to verify whether a point
of interest n
i
obtained from a peer is a valid (i.e., top k)
nearest neighbor of a mobile hostQ. To this end we utilize
the spatial relationship between mobile hosts and their POIs
as follows.
Theorem 3.1 LetQ andP
1
be two mobile hosts, and letP
1
havek nearest neighbors,n
1
,n
2
,...,n
k
, which are sorted
in ascending order according to their distance to P
1
. For
any nearest neighborn
i
ofP
1
, if Dist(Q,n
i
) +δ≤ Dist(P
1
,
n
k
) thenn
i
is one of the topk nearest neighbors ofQ.
In Theorem 3.1, Dist(Q, n
i
) is the Euclidean distance be-
tweenQ andn
i
,δ is the Euclidean distance betweenQ and
P
1
, and Dist(P
1
, n
k
) is the Euclidean distance betweenP
1
and its cached farthest nearest neighborn
k
. An illustration
of Theorem 3.1 is shown in Figure 2. The nearest neighbor
n
2
of mobile hostP
1
, which is a peer of mobile hostQ, can
be verified as the nearest neighbor ofQ.
Q
P1
n1
n2
n3
Dist (Q, n2) + < Dist (P1, n3)
Figure 2. Verification of a point of interest.
Under some conditions the kNN
single
method may not
be able to verify allk nearest neighbors. Therefore, we ex-
tend the verification process to include results from multiple
peers simultaneously. The kNN
multiple
method combines
the area of all the peers, each bounded by the outermost
NN circle, into a certain regionR
c
. ThekNN
multiple
ver-
ification technique is executed based on R
c
similarly to
kNN
single
. Theorem 3.2 provides the rules for verifying
nearest neighbors with multiple peers.
Theorem 3.2 If the nearest neighbor data setNN
P
is com-
posed of data from j peers, the certain region R
c
can be
represented as:
R
c
=P
1−area
∪P
2−area
∪···∪P
j−area
.
For any interest objectn
i
inR
c
, the distance betweenQ and
n
i
is used as a radius to draw a circle C
ni
. If C
ni
is fully
covered byR
c
, thenn
i
is a valid NN ofQ.
There will be cases when neither kNN
single
nor
kNN
multiple
can fulfill a kNN query. Hence a set with
some unverified elements is returned. If the response time
is critical, a user may agree to accept such akNN data set,
where the objects are not guaranteed to be the topk nearest
neighbors. Otherwise thekNN query must be forwarded to
a spatial database server. The partial results can be used to
bound – and hence speed up – the server search process [3].
3.5 System Demonstration
MAPLE is implemented in Java and Figure 3 shows its
multiple peer simulator interface. The left window frame
shows the simulated service region of MAPLE visualizing
all mobile hosts and POIs. Mobile hosts move on the road
network autonomously [1] while observing the speed limit
constraints. On the right pane, the simulator displays the
configuration parameters of current simulation, such as the
service region dimensions, and the number of mobile hosts.
Users are able to select (via mouse click) any mobile host
to launch a kNN query. This action activates the sharing-
based NN query algorithm visualization interface shown in
Figure 4. Each POI retrieved from peers within the com-
munication range is first verified in a step-by-step manner
using thekNN
single
method. In the case that thekNN
single
method cannot verify allk nearest neighbors, the simulator
automatically launches thekNN
multiple
method, illustrated
in Figure 5. Figure 6 depicts the server interface of MAPLE.
The server module executes the spatial queries received from
peers. It also records the page access frequency within the
spatial database server when performing spatial queries.
4 Conclusions
We have described MAPLE, a system to aid in the study
of scalable peer-to-peer data sharing for location-based ser-
3
Figure 3. The multiple peer simulator interface
of MAPLE.
Figure 4. The sharing-based NN query algo-
rithm visualization interface of MAPLE illus-
trating a single peer verification example.
vices in mobile environments. We implemented localized
verification algorithms to ascertain whether data items re-
ceived from peers provide a relevant answer to posed queries.
We also implemented a road network with realistic speed
limits to constrain the movement of peers in the mobile en-
vironment. MAPLE is interactive and provides step-by-step
visualizations to understand peer communication and verifi-
cation procedures. The objective of MAPLE is to provide a
platform for the evaluation of our ongoing research into the
novel design of P2P sharing techniques for location-based
services. MAPLE demonstrates the excellent scalability and
Figure 5. The sharing-based NN query algo-
rithm visualization interface of MAPLE illus-
trating a multiple peer verification example.
Figure 6. The server interface of MAPLE.
effectiveness of our current algorithm in high density mobile
environments.
References
[1] J. Broch, D. A. Maltz, D. B. Johnson, Y .-C. Hu, and J. Jetcheva.
A Performance Comparison of Multi-Hop Wireless Ad Hoc
Network Routing Protocols. In Proceedings of the 4
th
ACM/IEEE MobiCom, pages 85–97, 1998.
[2] A. Guttman. R-Trees: A Dynamic Index Structure for Spatial
Searching. In Proceedings of the ACM SIGMOD International
Conference on Management of Data, pages 47–57, Boston,
Massachusetts, June 18-21, 1984.
[3] W.-S. Ku, R. Zimmermann, and C.-N. Wan. Location-based
Spatial Queries with Data Sharing in Mobile Environment.
Technical Report USC-CS-TR05-843, University of Southern
California, 2005.
[4] R. Zimmermann, W.-S. Ku, and W.-C. Chu. Efficient Query
Routing in Distributed Spatial Databases. In 12th ACM In-
ternational Workshop on Geographic Information Systems.,
pages 176–183, 2004.
4

Linked assets

Computer Science Technical Report Archive

Conceptually similar

PDF

USC Computer Science Technical Reports, no. 886 (2006)

PDF

USC Computer Science Technical Reports, no. 843 (2005)

PDF

USC Computer Science Technical Reports, no. 892 (2007)

PDF

USC Computer Science Technical Reports, no. 908 (2009)

PDF

USC Computer Science Technical Reports, no. 909 (2009)

PDF

USC Computer Science Technical Reports, no. 891 (2007)

PDF

USC Computer Science Technical Reports, no. 899 (2008)

PDF

USC Computer Science Technical Reports, no. 628 (1996)

PDF

USC Computer Science Technical Reports, no. 959 (2015)

PDF

USC Computer Science Technical Reports, no. 912 (2009)

PDF

USC Computer Science Technical Reports, no. 734 (2000)

PDF

USC Computer Science Technical Reports, no. 878 (2006)

PDF

USC Computer Science Technical Reports, no. 627 (1996)

PDF

USC Computer Science Technical Reports, no. 846 (2005)

PDF

USC Computer Science Technical Reports, no. 659 (1997)

PDF

USC Computer Science Technical Reports, no. 762 (2002)

PDF

USC Computer Science Technical Reports, no. 893 (2007)

PDF

USC Computer Science Technical Reports, no. 795 (2003)

PDF

USC Computer Science Technical Reports, no. 781 (2002)

PDF

USC Computer Science Technical Reports, no. 685 (1998)

Description

Wei-Shinn Ku, Roger Zimmermann, Chi-Ngai Wan, Haojun Wang. "A P2P simulation model to support mobile scalable nearest neighbor queries for location-based services." Computer Science Technical Reports (Los Angeles, California, USA: University of Southern California. Department of Computer Science) no. 871 (2005).

Asset Metadata

Creator Ku, Wei-Shinn (author), Wan, Chi-Ngai (author), Wang, Haojun (author), Zimmermann, Roger (author)

Core Title USC Computer Science Technical Reports, no. 871 (2005)

Alternative Title A P2P simulation model to support mobile scalable nearest neighbor queries for location-based services (title)

Publisher Department of Computer Science,USC Viterbi School of Engineering, University of Southern California, 3650 McClintock Avenue, Los Angeles, California, 90089, USA (publisher)

Tag OAI-PMH Harvest

Format 4 pages (extent), technical reports (aat)

Language English

Unique identifier UC16269672

Identifier 05-871 A P2P Simulation Model to Support Mobile Scalable Nearest Neighbor Queries for Location-based Services (filename)

Legacy Identifier usc-cstr-05-871

Format 4 pages (extent),technical reports (aat)

Rights Department of Computer Science (University of Southern California) and the author(s).

Internet Media Type application/pdf

Source 20180426-rozan-cstechreports-shoaf (batch), Computer Science Technical Report Archive (collection), University of Southern California. Department of Computer Science. Technical Reports (series)

Access Conditions The author(s) retain rights to their work according to U.S. copyright law. Electronic access is being provided by the USC Libraries, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright.

Repository Name USC Viterbi School of Engineering Department of Computer Science

Repository Location Department of Computer Science. USC Viterbi School of Engineering. Los Angeles\, CA\, 90089

Repository Email csdept@usc.edu

Inherited Values

Title Computer Science Technical Report Archive

Description Archive of computer science technical reports published by the USC Department of Computer Science from 1991 - 2017.

Coverage Temporal 1991/2017

Repository Email csdept@usc.edu

Repository Name USC Viterbi School of Engineering Department of Computer Science

Repository Location Department of Computer Science. USC Viterbi School of Engineering. Los Angeles\, CA\, 90089

Publisher Department of Computer Science,USC Viterbi School of Engineering, University of Southern California, 3650 McClintock Avenue, Los Angeles, California, 90089, USA (publisher)