BANGLADESH RESEARCH PUBLICATIONS JOURNAL ISSN: 1998-2003, Volume: 5, Issue: 2, Page: 117-124, March -April, 2011
PROXIMITY BASED MOBILE LOCAL SEARCHING FOR PRODUCTS AND SERVICES OVER GPRS / EDGE Kamruddin Md. Nur1 and Md. Wahidur Rahman2 Kamruddin Md. Nur and Md. Wahidur Rahman (2011). Proximity Based Mobile Local Searching for Products and Services over GPRS / EDGE. Bangladesh Res. Pub. J. 5(2): 117-124. Retrieve from http://www.bdresearchpublications.com/admin/journal/upload/09205/09205.pdf
Abstract Tracking a mobile phone location and finding locations nearby of a mobile phone using GPS (Global Positioning System) already had existed where infrastructure is present. GPS enabled mobile phones are usually expensive and require GIS (Geographic Information System) mobile application. Searching a product or service which is not tagged with GPS mobile applications can not be found even if a user has a GPS enabled phone. In many developing countries, many locations as well as products and services are yet to be included in GPS mobile map applications. Thus, many products and services which are only known locally, may not be important globally within a particular location of a developing country where GPS phones are not abundant, demands a low cost alternative application using GPRS (General Packet Radio Service) which is accessible by almost any GSM mobile operator in the world. Proximity based mobile local products and service searching system only based on GPRS internet is quite challenging for a mobile application developers point of view. This paper addresses these above mentioned issues and presents a tolerant proximity based mobile searching system using mobile Internet (GPRS/EDGE) connectivity only.
Key Words: Proximity Based Mobile Searching; Location Based Searching using GPRS/EDGE.
Introduction With the growing use of mobile data service and continuous development of mobile phone usability, mobile oriented products and services are being at the reach for developed to developing countries. Proximity based or location based mobile services have received a significant attention with the growing use of wireless data service recently (figure 1.) (Wong and Choy 2007, Pressman and Lowe 2008, CartouCHe 2007, Tang 2009). Proximity based or location based mobile search system aims at offering users search product and service location available nearby. GPS enabled mobile map applications can find a location by its name, can show directions from a source to a destination, and can also show places which are already tagged with it quite efficiently. Although, determining users location automatically using GPS helps creating attractive and compelling user experience (Virrantaus et. al. 2001), it may not be the solution to every situation. Here we present a cost effective, tolerant proximity based mobile local
Corresponding Author’s email:
[email protected] 1. Dept. of Computer Science, Stamford University Bangladesh 2. Dept. of Business Administration, Stamford University Bangladesh
Nur and Rahman
118
search using GPRS/EDGE Internet connectivity. Web contents are hosted and experimented on a LAMP (Linux-Apache-MySql-PHP) localhost.
Figure 1. Web server content access in Mobile / PDA browser through GPRS/EDGE Rather describing a well known LAMP hosting procedure, we concentrate on the limited capability of mobile web browsers as well as slow data connectivity of GPRS / EDGE and mapping techniques suitable for built in WAP browsers.
Methodology At first we explore various mobile phones built in WAP browsers web content rendering capabilities and limitations. Then we explore the practical speed and latency of GPRS as the internet connectivity of these browsers. Then we opt to find an alternative map implementation of applications like popular google maps as it requires good web browsers like opera, skyfire, windows mobile internet explorer etc which also usually requires higher processing powered phones. Then we concentrate on our proximity calculation and location map display technique which can be rendered with built in WAP browser of inexpensive phones. Finally we experiment a case study of our implementation application in a localhost and present results.
Technical Review - Mobile Web Interface Mobile web interface development requires careful selection of web technology deployment due to its low processing power, battery consumption issue, Internet speed and latency, memory load, screen size, and interaction design etc. (Rappaport 2001, CartouCHe 2007, Ginige and Murugesan2001). Typically a web site is constructed with xHTML for basic web content presentation, JavaScript, jQuery, ActionScript, Flash etc for interactivity; php, aspx, jsp, python for serverside processing. As built in mobile wap browsers are still prevalent with inexpensive mobile phones, mobile web sites (as well as mobile version of web sites) are usually constructed with HDML, cHTML, WML, xHTML, JavaScript, and widely used low resolution graphics like JPG and PNG. Thus selecting xHTML, JavaScript, and graphics (.jpg, .png) enables developers create a single web interface for both desktop and mobile browsers. Several different standards exists: HDML, WML, cHTML, and xHTML. Older phones support HDML and WML. Most new phones support WML and xHTML. A comparison table of Web contents rendering in Mobile Internet Browsers also known as micro browser, mini browser, and Wireless Internet Browser (WIB) are presented in (Table. 1, x represents feature present).
http://www.bdresearchpublications.com/journal/
Proximity Based Mobile Local Searching
119
Table 1. Comparison web content rendering in different mobile/PDA internet browsers Web Contents
Oper NetFront Open Nokia SEMC- Motorola Pocket IE Obigo Vagabon a Wave Series 60 Browser Native d Mini Browser
cHTML
x
x
x
x
x
x
x
x
x
xHTML
x
x
x
x
x
x
x
x
x
WML
x
x
x
x
x
x
x
x
x
JavaScript
x
x
x
x
x
x
x
x
x
.jpg .png
x
x
x
x
x
x
x
x
x
CSS
x
x
x
x
x
x
SSL
x
x
x
x
x
Flash Lite
x
x
x
x
x
x
Mobile Internet Speed & Latency An unavoidable important issue of GPRS is Internet speed (and latency). A standard class 10 GPRS with (4+1, 3+2) time slot connection offers max total speed of 48 kbps (6KB) and where a standard class 10 EDGE (Enhanced Data Rate for Global Evolution) with (4+1, 3+2) time slot connection offers max speed of 236 kbps (29.5 KB). In reality, the speed is rather 4-6 Kbps for GPRS and 10-15 Kbps for EDGE on average. Both GPRS and EDGE has minimum of (1+1+1)= 3 sec of handshake delay plus latency. Internet connection latency depends on a number of factors including service provider’s setup and device location distance from the nearest mobile base station (Mobileisgood, 2007). Thus a careful good web design is substantially important for mobile / PDA access and use. Thus we set our agenda to a proximity based mobile local searching system which is -
• • • • • •
accessible from mobile internet browser light weight server side processing light weight client side processing low network bandwidth savvy a solution that does not require more bandwidth than GPRS / EDGE internet speed flexible and user friendly
System Analysis - Proximity Based Mobile Search Basic functionalities of the proximity based mobile local searching could be a location (e.g. city, town, suburb etc.) search, products or service (e.g. shopping mall, medicine store, petrol station etc.) location search nearby, using mobile browser (Steinfield, 2004). Surfer can select an interested item from the search result list and view or print selected item list, approximate distance from surfers location and view location map. All information found in the database is stored willingly by clients during clients sign up process. The System Administrator varies the sign up information and make information available to surfers. (Figure. 2) illustrates the basic functionalities of the system as use case diagram and (Figure. 3) illustrates the basic system interactions as sequence diagram.
http://www.bdresearchpublications.com/journal/
Nur and Rahman
120
Figure 2. Basic Usecase Diagram of Proximity Based Mobile Search System Although, automatic detection of user’s current location enables a compelling experience, due to the limitations of mini-browser functionalities in mobile phones and absence of proper technologies in developing countries for automatic location detection, this system aims to find an alternative way of offering a tolerant proximity based mobile local search. Considering Disjkstra algorithm for single source shortest path O (V (log V+E)) and Floyed Warshall’s O (V3) all pair shortest path algorithm, none of them are quite practical due to the mobile browser limitations, Internet speed and latency (Horowitz et. al. 2007). The total geographic region (the map) is divided into rectangular area slices as it found in a map. Each area has a name, locationId, locationImage. Each location knows its neighbor locations and distance from its neighbor locations in the database. User selects users’ current location and optionally selects the find in location of the information of interest and then types query to search. Proximity Service module finds neighbour nodes and calculates the distance between locations and displays area map to the user. The neighbour tables and distance matrix can be stored in the database temporary tables or just forgotten after proximity calculation. A simple implementation technique of map implementation and proximity calculation is shown in subsequent sections.
Figure 3. Basic Interaction Diagram of Proximity Based Mobile Search System http://www.bdresearchpublications.com/journal/
Proximity Based Mobile Local Searching
121
Location Mapping A number of map services are being offered by Google and Yahoo Inc., however these maps are bandwidth intensive and not optimized for GPRS / EDGE users. Moreover, many locations of developing countries have not been mapped and added by Google or Yahoo map service. In this section, we propose a heuristic novel map service that our system could offer. Although different approaches have been considered in developing map services, unfortunately due to the limitations of mobile / PDA browsers capabilities, only a static .jpg file can be fetched and displayed efficiently. We divide a total geographic area into location map slices and according to the area a product or service is located, the corresponding map slice is displayed on mobile screen. In our approach, an area has to remember only four neighbor areas, since all map slices are a uniform rectangular shaped (figure 4., instance figure taken from (Dhaka Map 2010)). As for example, the surfer is interested to find out product or services in location A, and the neighbor area nodes are B, C, D, and E as depicted in figure 5.
Figure 4. Location Mapping, an example location Dhanmondi where adjacent nodes are Hazaribagh, Ramna, Motijheel and Lalbag Proximity Calculation Considering the area nodes (vertices) and distance (weighted edges) in the following scenario (when products and services are added to location in the system database) in figure. 4.
Figure 5. Example Proximity Calculation Graph of Location 'A', and its neighbour locations 'P', 'D', 'Q', 'B', 'C', 'R', 'E', and 'S'
http://www.bdresearchpublications.com/journal/
Nur and Rahman
122
We have locations (vertices) A, B, C, D, E, F, and G and approximate distance between them are shown as weighted edges which consequences the following neighbour table in the database (Table. 2) Where N is a neighbor and distance table the database (Table. 3). Table 2. Location Neighbour Table Location
N1
N2
N3
A
B
C
D
B
A
C
F
C
A
B
E
D
A
E
E
D
C
G
F
C
B
G
G
F
E
N4
F
Table 3. Location distance table between nodes Location Nodes
Distance
A↔B
2
A↔C
1
A↔D
1
B↔C
1
B↔F
1
C↔E
1
C↔F
2
F↔G
1
E↔D
1
E↔G
3
Case Study and Results In this section we present a typical case study where location is choosen “Dhanmondi” and search for “Hospital”. The output fetched from database is “JBF Hospital” and “Ibne Sinha” which was only instance entered in the database during prototype testing. The output also was able to show the location slice map. However, it did not show exactly where JBF Hospital located is or Ibne Sinha. This was major limitation in this prototype implementation. The precision can be achieved by slicing the location map into more pieces however exactly how many slices need depends on products and services entered in the database where location address has to be incorporated with detailed map. For this case study we summarize this as a tolerant search result which is far from precision. However this mapping scheme can be extended and enhanced easily.
http://www.bdresearchpublications.com/journal/
Proximity Based Mobile Local Searching
123
Figure 6. Product / Service Search Result at Dhanmondi Location Prospects in Business Arena Any form of Location Based Services (LBS) has important business impact especially in developing countries where access of Information Technology (IT) is difficult for mass population. Proximity based local searching using GPRS only enables users to enjoy IT services with the use of cheaper phone only. With the massive growing demand and popularity of Value Added Services (VAS), LBS shows potential business impact to all stakeholders including – mobile users, mobile operators, and business organizations. Business organizations would want expand their product and services location availability by including themselves in the database where mobile users would like to find product and services with ease by using GPRS enabled phone only. In the whole process mobile operators would gain profit out of data connectivity charges. Thus, our proximity based mobile searching also opens up opportunities for M-Commerce (Mobile Commerce).
Conclusion This paper attempts to offer a tolerant proximity based products and services search technique over GPRS /EDGE only by a mobile device without GPS. A major obstacle to the mobile web is the low Internet bandwidth, where 3G or 4G network is not present. In this research we first identify the limitations of WAP browsers and focus on mobile web site contents and technologies suitable for WAP enabled cheaper mobile phones. Then we propose a tolerant light weight client and server side processing LAMP implementation of product and service location search with location map which is low bandwidth savvy and ideal for GPRS / EDGE Internet connectivity. We demonstrate our heuristic proximity calculation by using GPRS / EDGE internet connectivity unlike GPS system services like google, yahoo maps. Then we demonstrate a mapping techniques which suitable and efficient on low bandwidth and high latency internet connectivity such as GPRS / EDGE. The case study of the prototype application can show product and service location map however it lacks precision of exact location. Although the product and service search in the case study lacks precision, the searching and mapping technique can be extended and enhanced in future work. http://www.bdresearchpublications.com/journal/
Nur and Rahman
124
References Steinfield, C. (2004). The development of location based services in mobile commerce. In Preissl, B., Bouwman, H. and Steinfield, C. (eds.), Elife after the dot.com bust. Berlin: Springer, p. 177-197. Horowitz, E., Sahni, S. and Rajasekaran, S. (2007). Computer Algorithms. Summit, NJ, USA: Silicon Press. Mobileisgood (2007). GPRS (General Packet Radio Service). Retrieved from http://www.mobileisgood.com/WhatIsGPRS.php Virrantaus, K., Tirri, H., Veijalainen, J., Markkula, J., Katanosov, A., Garmash, A., and Terziyan, V. (2001). "Developing GIS-Supported Location Based Services," Web Information Systems Engineering, International Conference on, vol.2, p.66. Wong, K.Y. and Choi, Y.M. (2007). A simple system providing location-based service on urban area. In ICCOM’07: Proceedings of the 11th Conference on 11th WSEAS International Conference on Communications. Stevens Point, Wisconsin, USA: World Scientific and Engineering Academy and Society (WSEAS), p.60-65. Pressman, R. S. and Lowe, D. (2008). Web Engineering: A Practitioner's Perspective, McGraw-Hill, New York, ISBN:9780073523293. Ginige, A. and Murugesan, S. (2001). Guest Editors' Introduction: Web Engineering—An Introduction, IEEE Multimedia, vol. 8, no. 1, pp. 14-18, doi:10.1109/93.923949 CartouCHe (2007). Foundations of LBS. Retrieved from http://www.ecartouche.ch/content_reg/cartouche/LBSbasics/en/html/index.html Rappaport, T. (2001). Wireless Communications Principles & Practice. Prentice Hall. ISBN:0130422320. Tang, Y (2009). Interactive Pervasive Wireless Services over GPRS and Bluetooth, Chengdu, China, Dec 12 - Dec14, DOI: http://doi.ieeecomputersociety.org/10.1109/DASC.2009.47 Dhaka
http://1.bp.blogspot.com/Map (2010). Retrieved from LFQZavWtiW4/TafrTbt2gvI/AAAAAAAAApY/8chX7dgzO3I/s1600/Picture+of+ Dhaka+City+map.gif
http://www.bdresearchpublications.com/journal/
