An Experience in Accessing Grid Computing Power from Mobile Device
with GridLab Mobile Services
, Abstract
In this paper review the notion of the use of mobile device in grid computing environment, We describe our experience in evaluating the GridLab Mobile service as a tool which supports the combination of two mainstreams, grid computing and mobile devices. We use GridLab Mobile services platform to conduct our experiment to enable the access of grid computing power from Poznan supercomputing and networking center in Poland through smart phones. In order to deploy GridLab Mobile services we need to have Mobile Client, Mobile Command Center and grid application. GridLab Mobile services functionality are GridLab Resource Management System (GRMS) Job submission and Control, message box browsing, visualization viewing, and auxiliary functionality such as login. From our experiment it was found that to perform Login to the services through smart phone, we need to send the average of 635 Bytes of data and received 467,6 of data in 14,6 seconds. For getting message box we need to send the average of 1389 Bytes of data and received 1638,8 Bytes of data in 26,6 Seconds. In addition to that, for visualization purposes we needed to send the average of 2701,8 Bytes of data and received the average of 46139,4 of data in 26,6 Seconds.
1. Introduction
In few last year mobile devices has emerged as a common communication tools which have been developed by academic and industry sectors. These devices benefit from the advancement in wireless connection, and become more and more powerful in the last few years, as well as growing rapidly in numbers.
Grid computing is a group of computer, server, storage that form a virtual computer system. With grid computing we can optimize performance and reliability. In this paper we will combine this two issues to benefit from accessing grid computing power from a mobile devices. Technology used in this
work is GridLab Mobile services, which is a workpackage from GridLab project. The discussion of combining grid computing and mobile devices with other technology in detail is beyond this work. 2. Grid Computing
Grid computing concept is not new. Grid computing has lot of common with the “metacomputing”, which is popular during 1990’s. One of project that develop metacomputing concept is Informational Wide Area Year (I-WAY) which has been strongly influenced by the development of Globus project, the present de facto standard of grid computing [3]. Ian Foster and Carl
Riri Fitri Sari and Rene Paulus ([email protected])
Kesselman in their book “The Grid, Blueprint for a new Computing Infrastructure” define grid as hardware and software infrastructure that provides dependable, consistent, pervasive, and inexpensive access to high-end computational capabilities. The other definition from Foster on Grid Computing is “A Computational Grid is a collection of computers, on-line instruments, data archives and networks all tied together by a shared set of services that, when taken together, provides users with transparent access through interface devices to the entire set of resources”.
According to IBM’s definition [6], grid is a collection of distributed computing resources available over a local or wide area network that appears to an end user or application as one large virtual computing system. The vision is to create virtual dynamic organizations through secure, coordinated resource-sharing among individuals, institutions, and resources. Grid computing is an approach to distributed computing that spans not only locations but also organizations, machine architectures, and software boundaries to provide unlimited power, collaboration, and information access to everyone connected to a grid. The Internet is about getting computers to talk together and is a network of data; while grid computing is about getting computers to work together and is a network of resources. Grid will help elevate the Internet to a true computing platform, combining the qualities of service of enterprise computing with the ability to share heterogeneous distributed resources—everything from applications, data, storage and servers.
Another definition from the Globus Alliance, is as follows: "The grid refers to
an infrastructure that enables the integrated, collaborative use of high-end computers, networks, databases, and scientific instruments owned and managed by multiple organizations. Grid applications often involve large amounts of data and/or computing and often require secure resource sharing across organizational boundaries, and are thus not easily handled by today’s Internet and Web infrastructures."
Figure 1 shows architecture of grid computing remote implementation where grid provides access to distributed resources. Computational grid enables the sharing, selection, and aggregation of a wide variety of geographically distributed computational resources (such as supercomputers, computing cluster, storage system, data source, instrument, and developers).
Figure 1. Grid Computing (remote implementation) [7]
Grid computing key component includes the following:
• Resource management: the grid must be aware of what resources are available for different tasks.
• Security management: the grid needs to take care that only authorized user can access and use the available resources. • Data management: data must be
transported, cleansed, parceled, and processed.
• Service management: users and application must be able to query the grid in an effective and efficient manner.
To coverage complex system of services and software required to integrate Grid resources Grid community [1] has develop a layered model show in Figure 2. The bottom horizontal layer consist of hardware resources such as computer, network, data archives and so on. The next horizontal layer (common infrastructure) consists of the software services and system which virtualize the Grid. The key concept at the common infrastructure layer is the community agreement on software, which will represent the grid as a unified virtual platform and provide the target for more focused software and application. The next horizontal layer (user and application-focused Grid middleware, tools and services) contains software packages built atop the common infrastructure. The topmost horizontal layer (Grid application) represents application and user.
Figure 2. Layered architecture of the Community Grid Model [1]
The vertical layers represent the next step for grid development. The left vertical layer represent the influence of new device – sensors, PDA, wireless, and mobile devices. Over the next 10 years, these devices will need to be integrated with the Grid. At the same time, the increasing globalization of the Grid will
require serious policies of sharing, global-area networking and the development of Grid economies (the right vertical layer). 3. Mobile Devices
Mobile devices have become more and more powerful as their size got smaller. It can run heavy application as with earlier desktop computer. By their high mobility mobile devices gaining popularity and their number are growing rapidly.
Although mobile devices are powerful they have many limitations like limited screen size, input capability, processing power, small amount of memory, and short battery life. Current wireless network connections of mobile devices are slow, the latency is high and the bandwidth is limited. The connection to the network is not permanent. Mobile client normally open connection only when it is needed, the network connection may not always be available either.
Mobile services are developed for multiple platforms. The most common Operating System (OS) are Palm and Symbian. In addition, there are millions of mobile phones running some other proprietary OS. All of these platforms are different and not compatible with each other. Luckily, most of these platforms are capable of running application written in Java. Java 2 Platform Micro Edition (J2ME) [8] is an application environment for small and limited devices. Program running in this environment are called middlet.
J2ME define two configuration which is Connected Limited Device Configuration (CLDC) the most lightweight configuration and Connected Device Configuration (CDC) used in device with larger memory and performance. On top configuration layers
there is a profile layer specifies set of API fo a particular category of devices and minimum hardware, software, networking, and application requirements. Figure 3 show this configuration.
Combining these two key trends (grid computing and mobile devices) could mean the ability to command the power of supercomputers with mobile devices. GridLab Mobile Service will be used to combining these two key trends. GridLab Mobile services using J2ME technology at the mobile client which is the mobile devices.
Figure 3. Java 2 Platform, Micro Edition (J2ME) [8]
4. GridLab Project
GridLab project researched undertaking development of application and middleware for grid environments. GridLab project organize by Institute of the Bioorganic Chemistry of the Polish Academy of Sciences, Poznan Supercomputing and Networking Center (PSNC), Poland and co-funded by the European Commission under the Fifth Framework Programme.
GridLab project consist of 12 core workpackage (WP), with additional workpackages covering exploitation, dissemination and project management.
All GridLab technology fit into GridLab architecture shown in Figure 4.
WP1 Grid Application Toolkit (GAT) objectives are to design and build a Grid Application Toolkit from the services developed in other GridLab workpackages. GAT is a set of coordinated, generic managements systems, together with working implementations provided by the tools developed in GridLab project. WP2 Cactus
Figure 4. GridLab Architecture [4] Grid Application Toolkit (CGAT) will provide a very general toolkit framework supporting different Grid applications, from astrophysics to chemical engineering. CGAT extended GAT interface to Cactus [2]. WP3 Work-Flow Application Toolkit (TGAT) based on Triana [5] packages developed for gravitational wave data analysis.
WP4 Grid Portal goal is to design, build and support the GridLab Portal. WP5 Testbed Management takes care of all the activities necessary for deploying and running a large-scale grid testbed and to establishment of an actual prototype. WP6 Security aims at building solid and consistent testbed security management
tools and policies, as well as inter-testbed interoperability security policies.
WP7 Adaptive Grid Component goal is to provide adaptive components to other modules of the grid application toolkit (GAT). WP8 Data Handling and Visualization creates a framework that enables application to stream data for visualization purpose. WP9 Resource Management aim is to describe GridLab Resource Management System (GRMS).
WP10 Information Services goal is to design and build iGrid, the GridLab Information Services to supplies information about a specific resource. WP11 Monitoring objectives are to design and develop a general grid monitoring infrastructure, compliant to the Grid Monitoring Architecture proposal of the Global Grid Forum Performance Working Group. WP12 Access to Mobile Users objective is to develop implementation for mobile device so user can used it to accessing grid computing environment.
5. GridLab Mobile Services
WP12 GridLab Mobile Services give the Grid user possibility to access grid application and resource from any place using their mobile devices. Mobile device act as a mobile client and tightly coupled with the gateway. Minimal grid interface for GridLab Mobile Service are GridLab Mobile Client, Mobile Command Center as gateway, and grid service. Figure 5 present that minimal grid interface.
Mobile Command Center (MCC) is a portlet that serves as gateway between mobile devices and grid services. MCC required Java 1.4.2, Jakarta Ant >=1.6+, Jakarta Tomcat 4 or 5, GridSphere [5] portal (WP4) installed and deployed, GridPortlets installed and deployed in the same GridSphere container. Portlets are
reusable Web components that display relevant information to portal users. GridLab Mobile Client (GMC) run on mobile device which support J2ME technology.
GridLab Mobile Service functionalities are GridLab Resource Management System (GRMS) Job submission and control, Visualization viewing, Message Box browsing, and Auxiliary functionality.
Figure 5. Minimal Grid Interface [4] The GridLab Resource Management System (GRMS) is an open source meta-scheduling system, based on dynamic resource selection, mapping and advanced scheduling methodology. The main goal is to directly manage the whole process of remote job submissions to various batch queuing system, cluster or resources. The most important part is ability to submit job to the best resources according to job resource requirement. GRMS functionality can be used from mobile
device are submitJob, migrateJob, cancelJob.
Visualization viewing (Figure 6) working together with WP8 so the WP12 user can display scaled-down visualization (mainly picture) on mobile device many times. The picture send to mobile device according to the limitation of that device like low resolution screen and color depth, small bandwidth network, etc.
Figure 6. Visualization Viewing Message Box Service (Figure 7) main purpose is to store users notification messages in the folder structured repositories. GMC can access user folders and messages stored in Message Box.
Figure 7. Message Box browsing Auxiliary functionality is a part of GridSphere framework for user login and profile which is user login, user logout, user credential usage (for authorization), dan user data checking. Figure 8 represents the login service.
Figure 8. Login Form
From experimental we get some data for each service and calculate the average of those data represent in following tables. The specification used in this testing are:
• Sony Ericsson K508i device • 128x160 pixel screen • 65,536 color depth • Indosat M3 GPRS • Connect to http://rage1.man.poznan. pl:8089/mccp/mccp Table 1. Login n Sent (Bytes) Received (Bytes) Time (s) 1 635 478 17 2 635 478 9 3 635 478 9 4 635 426 29 5 635 478 9
Table 2. Message Box n Sent (Bytes) Received (Bytes) Time (s) 1 1293 1442 24 2 1233 1394 17 3 1233 1394 31 4 1303 1436 20 5 1883 2528 41 Table 3. Visualization n Sent (Bytes) Received (Bytes) Time (s) 1 2773 45829 53 2 2685 47249 44 3 2709 45873 42 4 2697 45873 43 5 2645 45873 35
Table 4. Average Service Send (Bytes) Received (Bytes) Time (s) Login 635 467,6 14,6 Message box 1389 1638,8 26,6 Visualization 2701,8 46139,4 43,4 6. Future Work
This work is part of our on going effort to review the current state of the art and available tools in the area of mobile devices and Grid Computing Platform. Some work such as the ones in [9, 10] show the need of a mechanism to ensure the effectiveness and the role of mobile devices to interact with grid. Therefore plan to extend the work to implement our mobile command center and merge that with other Grid Computing Applications being developed in our university at the moment.
7. Conclusion
The development of mobile device in last few years enabled us to access grid computing environment from this device. GridLab Mobile Service can be used to accessing grid application and resources from mobile devices. GridLab Mobile Service need some minimal grid interface which are the GridLab mobile client running on J2ME enabled mobile devices, Mobile Command Center as gateway, and the grid services it self.
Login to Mobile Command Center from GridLab Mobile Client need 635 Bytes send data and 467,6 received data in 14,6 second. Message box viewing need 1389 Bytes send data and 1638,8 received data in 26,6 second. Visualization service from mobile client need 2701,8 Bytes send data and 46139,4 received data in 43,4 second.
8. Acknowledgement
This work is supported by the support of Ministry of National Education of The Republic of Indonesia through Hibah Bersaing XIII project. We would like to thank Prof. B. Budiardjo, Dr. K. Ramli, Chaeru, Fauri, Syafar, Rizky, Fudhianto for their support and discussion.
Reference:
[1] Berman, F., Hey, A.J.G., Fox G.C., “Grid Computing Making the Global Infrastructure a Reality”, John Wiley & Sons, Ltd, England, 2003.
[2] Cactus,http://www.cactuscode.org, last accessed 24 June 2005
[3] Globus Alliance, http://www. globus.org. last accessed 20 June 2005 [4] GridLab, http://gridlab.org.
[5] GridSphere,http://www.gridsphere.org, last accessed 10 June 2005
[6] IBM, http://www.ibm.com, last accessed 15 May 2005
[7] Minoli, Daniel, “A Networking Approach to Grid Computing”, John Wiley & Sons, Inc, 2005.
[8] Sun Microsystem, http://java.sun. com, last accessed 15 May 2005
[9] S. Kurkovsky, Bhagyavati, A. Ray, Mei Yang, “Modelling a Grid based Problem Solving Environment from Mobile Devices, Proceedings of the International Conference on IT: Coding and Computing (ITCC’04).
[10] David C. Chu, M. Humprey, “Mobile OGSI.NET: Grid Computingon Mobile Devices”, Fifth IEEE/ACM International Workshop on Grid Computing (Grid’04), PP 182-191.
