Für neue Kunden:
Für bereits registrierte Kunden:
Doktorarbeit / Dissertation, 2014
List of Figures
List of Tables
1.2 Grid computing
1.4 Integration of mobile computing and grid services
1.5 Challenges of the integration
1.6 Security in mobile grid computing
1.7 Mobile grid security requirements
1.8 Route and resource discovery in mobile grid system
1.9 Research gap and objectives
REVIEW OF LITERATURE
2.1 Computational grid
2.3 Need of resource discovery
2.4 Outline of the survey
2.4.1 Components of resource discovery
2.4.2 Grid resource management model and algorithms
2.5.1 Grid Information Protocol (GRIP)
2.5.2 Grid Registration Protocol (GRRP)
2.5.3 Lightweight Directory Access Protocol
2.6.1 Decentralized approach
2.6.2 Agent based approach
2.6.3 Routing transferring model-based approach
2.6.4 Ontology description- based approach
2.6.5 Parameter based approach
2.6.6 Quality of service based approach
2.6.7 Request forwarding approach
2.6.8 Peer-to-peer approach
2.7 Discovery mechanism on different platform
2.7.2 Ninja service directory service
2.7.3 OPEN GRID SERVICE ARCHITECTURE (OGSA)
2.7.4 WEB SERVICE DISCOVERY ARCHITECTURE (WSDA)
2.7.6 UNIVERSAL DESCRIPTION, DISCOVERY AND INTEGRATION (UDDI)
2.7.11 A brokering protocol for agent based grid resource discovery
2.7.12 ANT COLONY OPTIMIZATION (ACO)
INTRODUCTION TO TRUST AND TRUST MANAGEMENT
3.2 Motivation for trust management
3.3 Definition of trust management
3.4 Trust context
a) Basis for Trust
3.5 Trust management: A survey
3.6 Discover and computations of trust
3.7 Trust propagation
3.8 Trust aggregation
3.9 Trust prediction
3.10 Existing research’s applications
ANT BASED RESOURCE DISCOVERY AND MOBILITY AWARE TRUST MANAGEMENT FOR MOBILE GRID SYSTEMS
4.1 Matrices estimation
4.1.1 Estimation of distance
4.1.2 Estimation of residual battery power
4.1.3 Estimation of available bandwidth
4.1.4 Estimation of local trust and global trust (TG)
4.1.5 Estimation of mobility
4.2 Proposed scheme
4.2.1 Phase 1 - Ant based resource discovery mechanism
4.2.2 Phase 2- - A mobility aware trust management technique
4.3 Simulation setting
a) Simulation on Load
b) Simulation on speed of mobile grid node
ENERGY CONSTRAINED HIERARCHICAL TASK SCHEDULING ALGORITHM FOR MOBILE GRIDS
5.1 Energy constrained Mobile Grids
5.2 Energy Constrained Hierarchical Task Scheduling Algorithm
5.2.1 Estimation of residual battery power and relevant mathematical representation
5.2.2 Hierarchical task scheduling
5.2.3 Proposed Algorithm
5.2.4 Merits of Proposed Scheme
5.3 Simulation setup
5.3.1 Based on Number of Requests
5.3.2 Simulation on Load
CONCLUSIONS AND FUTURE PROSPECTIVES
6.1 Main contributions
6.2 Future prospective research
TCL Script for Ant Based Resource Discovery and Mobility Aware Trust Management (ARDMTM)
File name: ARDMTM.tcl
Topology File: Topo30
Constant Bit Rate File: cbr2
Replication Based Job Scheduling TCL file: RBJS.tcl
Topology File for RBJS: topo30-e
Constant Bit Rate File for RBJS: cbr2-e
Implementation of ARDMTM Protocol
File Name: ARDMTM.cc
Brief Definition file for Agent ARDMTM
File Name: ARDMTM.h
Brief Implementation file for globally available methods
File Name: ARDMTM_Common.cc
Brief Implementation File for Routing Table
File Name: ARDMTM_rtable.cc
TCL Script for ENERGY CONSTRAINED HIERARCHICAL TASK SCHEDULING (ECHTC)
File Name: ECHTC.TCL
Implementation of ECHTS Protocol
File Name: ECHTC.cc
"I hereby declare that the work, which is beingpresented in the Thesis, entitled “An Experimental Study towards Realizing Ant Based Resource Discovery and Mobility Aware Trust Management for Mobile Grid Systems” is my own work and that, to the best of my knowledge and belief, it contains no material previously published or written by another person nor material which has been accepted for the award of any other degree or diploma of the university or other institute of higher learning, except where due acknowledgment has been made in the text”.
It is certified that the comments given by experts in DRC have been suitably incorporated (under the advice ofmy supervisor) in the draft thesis.
Abbildung in dieser Leseprobe nicht enthalten
Dr. Prasun Chakrabarti
Head and Associate Professor
Sir Padmapt Singhania University
This is to certify that the thesis entitled “An Experimental Study towards Realizing Ant Based Resource Discovery and Mobility Aware Trust Management for Mobile Grid Systems ” submitted by Mr. Arjun Singh to the Suresh Gyan Vihar University towards partial fulfillment of the requirements for the award of the Degree of Doctor of Philosophy in Faculty of Engineering (Computer Science and Engineering) is a bonafide record of the work carried out by him/her under my /our supervision and guidance.
It is certified that the comments given by experts in DRC have been suitably incorporated (under my supervision) in the draft thesis.
It is also certified that the candidate Mr. Arjun Singh have attended the course work of one semester at Sure Gyan Vihar University, Jaipur.
Dr. Prasun Chakrabarti
Place : Udaipur
We certify that the work presented by Mr. Ariun Singh is up to the standard ofSuresh Gyan Vihar University. The scholar has presented his/her research work before DRC on 11/07/2014. We are of the opinion that the candidate may be permitted to submit his thesis entitled “An Experimental Study towards Realizing Ant Based Resource Discovery and Mobility Aware Trust Management for Mobile Grid Systems ” in Computer Science and Engineering.
This is to certify that research work embodied in this entitled “An Experimental Study towards Realizing Ant Based Resource Discovery and Mobility Aware Trust Managementfor Mobile Grid Systems ” was carried out by Arjun Singh, Regd. No.
SGVU111805426 is approved for the award of the degree of Doctor of Philosophy in the faculty of Engineering (Computer Science and Engineering) by Suresh Gyan Vihar University.
Abbildung in dieser Leseprobe nicht enthalten
I would like to thanks my supervisor Dr. Prasun Chakrabarti for his guidance, assistance and unremitting support over the period of this research work. He has always been present to guide me and provide his insight on all the technical and non-technical aspects of research work. His intellectual, creativity and insightful suggestions have been invaluable for my thesis research. I could never thank him enough for being an excellent mentor and a wonderful person. Sometimes words arejust not enough to express one’s heartfelt gratitude.
My thanks also go to the Dr. Anu Poonia, Dean Research, Suresh Gyan Vihar University, for their invaluable comments and enlightening suggestions have helped improve the quality for this thesis.
I also want to acknowledge all members of Suresh Gyan Vihar University and Sir Padampat Singhania University for their continuous support. Their support and motivation make this research work possible.
I am very grateful to my wife Surbhi Chauhan, and daughter Charvi Singh, for their love, care and assistance during studies and research work.
I extend my thanks to my beloved sister-cum-daughter Reshu who has been my inspiration and honored. When I looked at her, it encourage me to do hard work.
I also express my special thanks to my beloved parents for their splendid cooperation & support all throughout my research work.
Last but not least, I would like to thank the Staffs of School of Engineering, Sir Padampat Singhania University for their comments and suggestions provided to me at every phase of the research work.
Figures Page No.
Fig. 1.1: A Grid Architecture and Its Relationship To Internet Protocol Architecture
Fig. 1.2: Dynamic and Fixed Wireless Grids
Fig. 3.1: Direct Trust
Fig.3.2: Indirect Trust
Fig.3.3: Hybrid Trust
Fig. 4.1: Selection of Super grid nodes
Fig. 4.2: Simulation Topology
Fig. 4.3: Rate Vs Delay
Fig. 4.4: Rate vs Delivery Ratio
Fig.4.5: Rate vs Packet Drop
Fig. 4.6: Rate Vs Throughput
Fig.4.7: Speed vs Delay
Fig. 4.8: Speed vs Delivery Ratio
Fig. 4.9 Speed vs Packet Drop
Fig. 4.10: Speed vs Throughput
Fig.5.1: Mobile Grid Network Architecture
Fig5.2: Simulation Topology
Figure 5.3 Number of Request vs Delay
Fig. 5.4 Number of request vs Delivery Ratio
Fig. 5.5 Number of Request vs Throughput
Fig. 5.6 Number of request vs Energy Consumption
Fig. 5.7: Loadvs Delay
Fig. 5.8: Load vs Delivery Ratio
Fig. 5.9: Load vs Throughput
Fig. 5.10: Loadvs Energy Consumption
Table 2.1: Comparison of resources Discovery Approaches
Table 4.1 Header of Ant Agent
Table 4.2 Resource Table
Table 4.3 Simulation Parameters
Table 4.4 Rate vs Delay
Table 4.5 Rate vs Delivery Ratio
Table 4.6 Rate vs Packet Drop
Table 4.7 Rate vs Throughput
Table 4.8 Speed vs Delay
Table 4.9 Speed vs Delivery Ratio
Table 4.10 Speed vs Packet Drop
Table 4.11 Speed vs Throughput
Table5.1. Simulation parameters
Table 5.2 Number of Request vs Delay
Table 5.3 Number of Request vs Delivery Ratio
Table 5.4 Number of Request vs Throughput
Table 5.5 Number of Request vs Energy Consumption
Table 5.6 Load vs Delay
Table 5.7 Load vs Delivery Ratio
Table 5.8Loadvs Throughput
Table 5.9 Load vs Energy Consumption
Grid technology is a new paradigm which has the potential to completely change the way of computing and data access. Generally speaking, we could consider the Grid as the new enabling technology to transparently access computing and storage resources anywhere, anytime and with guaranteed Quality of Service (QoS). Grid computing has emerged to cater the need of computing- on-demand due to the advent of distributed computing with sophisticated load balancing, distributed data and concurrent computing power using clustered servers. The Grid enables resource sharing and dynamic allocation of computational resources, thus increasing access to distributed data, promoting operational flexibility and collaboration, and allowing service providers to scale efficiently to meet variable demands.
The lack of adequate development methods for this kind of systems since the majority of existing Grid applications have been built without a systematic development process and are based on ad- hoc developments suggests the need for adapted development methodologies.
This thesis concern the resource discovery and trust management with security in large size of future grid. An automatic discovery mechanism is needed to find nodes willing to participate in the grid. For mobile grids, a decentralized discovery mechanism is vital to cope with the fluctuating topology and large number of participants.
The thesis implemented an Ant based discovery mechanism in which forward and backward ants are used to establish super-grid nodes. The criteria for selecting the super-grid nodes include distance, CPU speed, available bandwidth and residual battery power. After establishing the supergrid nodes among the grid nodes, they collect information about all the resources in a resource table. It consists of grid node id, resource availability, distance from super-grid etc. If any node wants a specific resource, it sends request to its nearest super-grid node from which the node ids matching the request, are returned.
The local and global trust values of each node can be estimated based on the factors Job Response time, percentage of correctly received data, Number of successfully finished jobs. These factors can be collected based on the feedback from the user.
The trust values can be updated based on the predictive residence time of each grid node, (ie) The node with least residence time (with high mobility) is penalized by reducing the trust value by a step value, Similarly, the node with high residence time (with low mobility) is rewarded by incrementing the trust value, These trust values are encrypted and signed by shared symmetric keys, Experiments and results shows that proposed system is efficient since it is not distributed. The encryption and signing ensures confidentiality and authentication of the system. The mobility aware trust management provides connectivity apart from providing security.
MOBILE GRID COMPUTING- A REVIEW
In Mobile Grid systems, the automatic service deployment initially requires the node discovery. Most of the present security mechanisms on Grid systems hardly reflect the mobility of the nodes which may affect the applied security mechanisms leading to inaccurate and insufficient security. In order to overcome from these issues, in this thesis, we propose an ant based resource discovery and mobility aware trust management for mobile grid systems. Super-grid nodes are selected in the network using ant colony optimization based on the parameters such as distance, CPU speed, available bandwidth and residual battery power. These selected nodes are utilized in the resource discovery mechanism. In order to maintain strong security with mobility management system, a proficient trust reputation collection method has been adopted. By simulation results, we show that the proposed approach is efficient and offers more security.
Due to increasing data computation requirements, a pool of ideal computers can very useful to compute the extremely complex and large technical problems. It is observe that most of the time individual computers processing and memory space power goes ideal and wasted. To make the maximum utilization of CPU power and Memory space, grid computing can be very beneficial. Grid computational architecture is the interconnection of ideal computers, meant to compute a large problem and all machines works seamlessly. As competence and swiftness become significance criteria, computational grid have surface as possible substitute to maximizing processing resources. Grid computing has emerged to supply the requirement of computation on demand due to the emerging technologies like, urbane load balancing, distributed data and parallel
computing power using clustered servers. The main objective of grid computing to allow resource sharing, distributed on wide geographical area and coordinate with them in dynamic multi- institutional virtual organizations. It enables these heterogeneous resources to be aggregate and harvest the power to accomplish new functionalities and capabilities. Basically grid is service oriented architecture and many organizations are adopting it minimize the cost by utilizing maximum power of ideal resources. Grid is proven technology and currently using in many areas such as high energy, aerospace, health care, bio-medical, learning and is continuing to evolve and expand. For most application grid has become very cost effective and efficient computing platform for solving complex scientific problems and used a high performance distributed applications.
In recent years with the advancement in the mobile technology, mobile computing has been successful in utilizing academic and industry study efforts to bring product to the consumers. Mobile devices are taking the advantages of wireless technologies to enrich our daily life and helped in increasing the productivity.
Mobile computing is general term describing the applications of portable, small and wireless computing and communication devices. It includes devices like PDA, laptop, smart phones, Phablets and Tablets. Mobile computing focuses on providing the service “Anytime Anywhere” on the go.
Main challenges in the grid environment are resource discovery, route discovery and security. With new technologies it is possible to discover the node and route on the fly and access them dynamically. A mobile device is an instrument which becomes necessity of everyone’s life. Everyone carrying it is everywhere. Everyday technology is changing at rapid speed in mobile device arena. Mobile devices are getting more and more powerful in the terms of processing power, memory power and battery power. Mobile computing getting popular is devices are always with the user on mobility. Keeping this thing in mind, mobile grid computing architecture was designed. Mobile grid, in relevance to both, Grid and mobile computing, is full inheritor of grid with additional feature of supporting mobile users and resources in a secure, seamless transparent and effective way.
The grid computing mainly focused on sharing computer power and data storage capacity over Internet. A well-known example of grid computing is SETI (Search for Extra Terrestrial Intelligence) project (@home) in which many users share the unused processing cycle of their PC. As earlier said, that our desktop or laptops spend vast majority of their time sitting ideal. Networking has improved from a few kilobytes per second to thousands of megabytes per second; at the same time it is unusual to find a machine not connected to a corporate network or Internet. Finally in the last 20 years disk capacity has increased from a few megabytes to hundreds of gigabytes on the desktop and many terabytes in the servers.
While these developments have occurred, business pressures have consistently forced companies to seek new and innovative ways to enhance the flow of their operations. A car manufactures wants to conduct simulations with maximal precision to come up with a competitive design of a new auto model. A bank may desire to process ever large sets of historical data to better forecast the coming months. A pharmaceutical researcher who has access to remote scientific data can make faster progress in developing a new medicine. Finally, dislocated teams of software developers want to access collaborative environments enabling them to cooperate on a common source code.
Grid computing, which deals with resource sharing, provides abstraction and technology addressing such issues. These are problems characterized by distribution, resource heterogeneity, large scale and the need for collaboration, at the same time requiring guaranteed levels of reliability, safety, security and quality of service.
Implementing Grid solutions presents challenges. Permitting the use of resources that were otherwise tightly controlled demands that proper security measures exist. Large-scale resource sharing demands scalability and interoperability of the participating systems. Methods for service discovery, metering, accounting and billing need to be in place to enable effective sharing even within the same organization. Policies for access control need to be carefully defined to ensure that resource usage does not exceed the availability limits for CPU, storage, licenses, bandwidth and a number of other technology and contract related metrics.
The analogy used for the grid is Power grids where consumers or electric devices get access to electricity through the power sockets fit on the home wall without caring that how and from where this electricity is coming. In grid computing environment computing becomes omnipresent and each person gets access to computing resources. User gain access to these resources as needed with no or less knowledge of where those resources located on internet.
Grid Computing require a middleware that can divide and distribute the program (Job) into small pieces and deliver the distributed nodes for the execution. A Grid Architecture and its relationship to Internet Protocol Architecture shown in figure 1.1.
Abbildung in dieser Leseprobe nicht enthalten
Fig. 1.1: A Grid Architecture and Its Relationship To Internet Protocol Architecture
Fig. 1.1 illustrates the categorization of Grid architecture layers. At the Fabric layer, all sharable resources such as computers, systems, storage resources, data and catalogs etc. are available. The Connectivity layer provides a means of communication and authentication, needed to communicate with these resources. Above connectivity layer resource layers is available, which provides, the protocols that allow user to obtain data from an individual resource and to manage them, controlling the access, starting of processes, management, monitoring and audit. Protocols and services available at resource and collective layer are not associated with any specific resource. On dissimilar, they are global in nature and apprehension relations across collections of resources. At the top the application layer have user applications that function within a Virtual Organization.
Internet was the path breaking technology in the 90’s. Internet brings the rapid and dramatic transformation in the business world and changed the common people life in various aspects. Extensive research and innovation in wireless technologies (mobile computing) has profound impact on our daily lives. India felt the benefit of these technologies in between orlate 20’s and 21’s, on the other hand, developed countries started the use of these technologies in early days of their development. Mobile computing and Internet technologies helped developed countries in digitizing the economy. Mobile and wireless technologies helping the developing countries to bring e-governance system for more transparency in the system. India’s business and economy going online with its vast pool of IT experts. Countries like India has bright future in the mobile computing as daily new technologies getting introduced in Indian market.
With the convergence of communication and computing industries, mobile phones, tablets, phablet and PDA are quickly sprouting into powerful adaptable devices. The role of these devices has grown from organizing contact address and to-do list, sending faxes and emailing to stock market, agriculture, e-business, e-ticketing and e-banking. The development of wireless sensor technologies changing the agriculture, the parameters of soil and water will be transmitted to farmer for necessary action. Doctors will be able to help and assist the patients from remote location. Limiting the wireless technology to communication alone is bigoted. Wireless sensor and Mobile computing is in a position to play a vital role in the area of business, agriculture, and healthcare, even more so for developing countries like India.
Mobile computing can be defined as a computing environment of physical mobility with small, portable and nomadic devices. The user of mobile computing environment will be able to information, access data, or other rational objects from any devices in any network while on the move. Mobile Computing main emphases on the requirement of providing access to communication, information, and services everywhere, anytime and any available means. The solutions for achieving is not always easy to implement. Mobile computing requires the communication infrastructure, modified computer networks, application program and specific operating systems. The mobility issue implies some constraints that must be addressed since they limit the capability of a moving resource in contrast to a fixed one. Besides, the rise of computing devices such as laptops or tablets and their increase of computational power, memory power and usability have given the users the freedom to move around the world without any problems. The requirement to not only to carry the devices, but also to use the services during the mobility of the user has been a natural concern. Mobile Computing can be defined as a paradigm of computing in which users who carry portable devices have access to information services through a shared infrastructure, irrespective of their physical movement or location behavior. Below are the few scenarios of mobile computing:
- Carrying the laptop/tablet during travel with ability to send the documents and e-ticketing from train/aircraft.
- Getting correct and timely information related to different stocks are very important. Also, online trading of stocks while on the move is quite critical for certain lifestyles. Stocks tickers, stock alerts, stock quotes and stock trading can be made ubiquitous so that users can check their portfolio and play an active role in the market.
- The possibility to access Internet services and information while on the move. For instance, to be able to read e-mail at any time or to obtain the address of a restaurant in a foreign town (and maybe the description of the fastest way to reach the place).
With the introduction of mobile and pervasive computing, new models of distributed communication and computation are being introduced, leading to systems that are open in that they do not pre identify a set of known participants and dynamic in that the participants change regularly and not just due to sporadic failures. It is exciting to note that this development is occurring at many levels, infrastructure, communication, and application. At the communication level, for example, mobile ad hoc networks treat nodes as independent routers, requiring new techniques to guard against faculty or malicious nodes that topple or black hole packets. Correspondingly, as applications become more intelligent and sophisticated, they need greater degrees of decision making so that they can interact spontaneously with other peers that happen to be in their vicinity.
The core challenges of mobile computing can be categorized into:
Communication challenges. Mobile computers and devices entail a wireless network to access the network on the move, and wireless communication faces more hindrances than wired communication because the contiguous environment interacts with the radio signal, introducing echoes, noises or even blocking it. As an outcome, wireless communication is described by and frequent disconnections, higher error rates and lower bandwidths.
Mobility challenges. The capability to change locations while connected to a network increase the volatility of some information. Data considered static for stationary systems becomes dynamic in Mobile Computing. Mobility presents several problems: A mobile computer’s network address changes dynamically, its current location affects configuration parameters, and its communication bandwidth varies according to its position.
Devices challenges. Classical systems and applications are usually designed for fixed devices. Mobile devices need to be light, small, operational under wide environmental conditions and durable. They require minimal power usage for a long battery life.
Security challenges. Mobile Computing imposes special security requirements, particularly with regard to identification and certification. Device is not the person, technology can hide the identities of the users. Privacy is at risk when utilizing in secure channels, as it is possible in wireless communication.
Performance challenges. These are twofold. Firstly, mobile devices are usually less powerful compared to their static counterparts. This is mostly due to weight, size, and ergonomics for mobile devices. Secondly, mobility or, in general, all changes to a given system, prevent traditional optimization.
The availability of portable computing devices and advances in wireless networking technologies have contributed to the growing acceptance of mobile computing applications and opened the door for the possibility of seamless and pervasive services in mobile environments. Nevertheless, due to the limited network connectivity, device capabilities, transmission range, and repeated changes instigated by the user or device mobility, a substantial load is placed on applications to be deployed in an environment where mobile devices must connect to each other through automatic configuration and communicate with each other over wireless links.
In the purview of Grid and Mobile Computing, Mobile Grid is an heir of Grid that addresses mobility issues with the added elements of supporting mobile users and resources in a seamless, transparent, secure and effective way. It has the capability to organize primary ad-hoc networks and offer a self-configuring Grid system of mobile resources (hosts and users) connected by wireless links and forming random and changeable topologies.
Mobile Grid computing is about making Grid Services available and accessible anytime anywhere from mobile devices. Wireless mobile devices are generally characterized by several limitations and constraints. The processing power is low in most of these devices (maybe with the exception of laptops), the built-in memory is low and the storage capacity is very limited compared to desktops despite the use of external memory cards by few devices. The battery life is very short and display panels are also very restricted in quality and size. Due to mobility nature the network is not stable (poor bandwidth and intermittent). These limitations make mobile devices a platform that would benefit from the Grid. The main advantages of Mobile Grid computing include mobile- to-mobile and mobile-to-desktop collaboration for resource sharing, improving user experience, convenience and contextual relevance and novel application scenarios. A Grid-based mobile environment would allow mobile devices to become more efficient by off-loading resourcedemanding work to more powerful devices or computers.
The characteristics of mobile devices such as resource poor nature and low bandwidth connectivity (wireless) must be addressed if these devices are to be integrated as computational resources into a Grid. The system must be capable to operate on constrained hardware. The periodic and dynamic network environment must be handled elegantly. The Grid should be flexible and reflective, to allow users to make tradeoffs and select the combination of services that is best suited for their purpose. This has led to the requirement of dynamic query and adjustment, at runtime. The system should also be capable to support computation and service migration, replication, application recovery and run-time monitoring. The disconnection of a given node in the Grid should not disrupt service accessibility. In addition, the system must be able to divest computations if the device is likely to turn out to be unavailable soon.
Mobile Grid enables both the mobility of the users requesting access to a fixed Grid and the resources that are themselves part of the Grid (see Figure 1.2). Both cases have their own limitations and constraints that should be handled (. In the first case the devices of the mobile users act as interfaces to the Grid enabling monitoringjob submission, and management of the activities in an anytime, anywhere mode, while the Grid provides them with a cost-efficiency, high reliability and performance. Physical limitations of the mobile devices make necessary the adaptation of the services that the Grid can provide the users’ mobile devices. In those cases Mobile Grid has the meaning of ‘gridifying’ the mobile resources. In the second case of having Mobile Grid node/resources, we should emphasize that the performances of current mobile devices are ominously increased. Laptops and PDAs can provide aggregated computational capability when assembled in hotspots, creating a Grid on site. This ability can improve the usage of Grid applications in places where this would be imaginary for most of the people.
Abbildung in dieser Leseprobe nicht enthalten
Fig. 1.2: Dynamic and Fixed Wireless Grids
However, there will always be the question on why a Grid solution should be adopted in comparison to any other non-Grid Information Technology (IT) solution. Grid is not intended to be the ‘panacea’ to the problems related to IT domain. It is a promising emerging technology that has the ambition to provide more efficient and more beneficial solutions than its ‘competitors’ by enabling the simple ‘connect and share’ approach in the same manner as the current Internet search engines apply the acquire information concept and connect. By this, Grids and Mobile Grids can be the ideal answer for many large scale applications that are of dynamic nature and require transparency for users.
A possible solution to the shortfall in required processing power is that mobile devices make use of Grid services to enable users to access distributed computational resources automatically on demand. Many Grid services can improve the capabilities of ubiquitous mobile devices so that complicated tasks can be completed through user handheld devices. For example, the traveler’s smart phone with a built-in camera can produce a large volume of data which will need to be processed if any special tasks are required. Large volumes of data demand significant computational power (e.g. image processing, location recognition) which can be best supplied by Grid services. The traveler can use his smart phone to discover local available Grid services and submit a complex request. The Grid will assist the traveler to achieve the complex task through distributed resources available anywhere in the world and return the results to his smart phone. In addition to providing new application opportunities for ubiquitous users, offloading complex tasks from resource-limited devices to the Grid has the potential to save energy, storage space, processing cycles, and memory capability hence possibly further reducing the size, weight, and cost of mobile and pervasive devices.
The concept of Grid service on mobile devices can benefit from another significant movement in computing, the move toward machine-process able explicit knowledge as exemplified by the Semantic Web. Semantic Web technologies are already being used within mobile computing, for such tasks as signifying context information which requires the description of suitable ontologies. An integration between Semantic Web technologies and Grid computing is also recognized, and several “Semantic Grid” projects have demonstrated a high degree of easy-to-use and seamless automation to facilitate flexible collaborations and computations on a global scale. Semantic Web technologies have the potential to provide a very considerable degree of automatic processing, integration and interoperation, which is a essential requirement of the necessary system infrastructure to allow mobile devices to use Grid services effectively.
Mobile devices form the intersection between the physical world and the digital world. In this view, the digital world of the Grid meets the physical world through a variety of instruments, sensors, and interfaces. As a result of several case studies from various “e-Science” projects, we can conclude that mobile devices need the Grid for integration and computation, whereas the Grid wishes mobile devices to interface with the physical world. The Semantic Web delivers the necessary automation and interoperability required to build an ambient intelligence infrastructure. Actually, it is the interaction of the Semantic Grid and the physical world (interacting through pervasive and mobile devices) that will enable us to realize this new concept of Ambient Intelligence, the idea of intelligence in the nearby environment supporting the activities and interactions of users.
The combination of these two computing models has the potential to realize a very significant development in the adoption of high performance Grid accessed through mobile devices. At first glance, this combination does not seem either efficient or appropriate. Clients that need to interact with Grid resources before they can accomplish a task will be required to install and utilize Grid client end libraries. At present, the existing Grid client libraries are relatively resource-intensive when compared with the limitations of mobile nodes. Furthermore, most of the current Grid applications have been developed with the conjecture that the end-systems possess sufficient resources for the task at hand and the communication between clients and resource providers will be reliable.
In a mobile computing environment, users are able to come in the range covered by mobile services and access ubiquitous resources with their mobile devices suitably and effortlessly. It will be the responsibility of the computing environment to detect user presence or absence and configure services automatically based on various context information. If Grid services are included in the mobile computing environment, they are also required to be made context aware. For instance, Grid services will need to be tailored and provided in different ways under different application cases. One of the key features of a mobile Grid environment is that it is highly dynamic: mobile users and Grid services can be integrated on-the-move, and Grid services are essential to be located, composed, acquired, and coordinated depending on various context information such as the mobile device capability and the user position. Self-management, self-configuration, and self-adaptation must also be taken into consideration for the mobile Grid environment.
Considering the assumptions of conventional Grid computing and the highly dynamic requirement of the mobile computing environment, it is quite apparent that a number of challenges must be resolved to comprehend the vision of building the bridge between Grid services and mobile devices. Here are the following key challenge points:
- A framework is required to convey mobile devices into the service-oriented Grid environment in an open, flexible and interoperable way.
- A task implementation mechanism is essential to be built so that complex tasks can be accomplished by invoking suitable Grid services.
- Grid and mobile computing have similar challenges in terms of service description, discovery and composition at the appropriate level of abstraction. The description mechanisms have not been identical, and the methods for service discovery and composition are not enough mature at this time.
- Because of the dynamic nature of mobile computing environment, context awareness is important for accomplishing the goal of providing Grid services that are appropriate for mobile users at the right time, in the right place, at the right device, in the right format.
A very important aspect in Grid computing is security that has been a central issue from the beginning, and has been observed as the most substantial challenge for Grid computing. Now the growing size and profile of the Grid require comprehensive security solutions as they are critical to the success of the endeavor. A comprehensive security method, proficient of replying to any attack on Grid resources, is indispensable to guarantee its anticipated adoption by both the users and the resource providers. Some studies have been carried out by the GGF (The Global Grid Forum) proposing a strategy for addressing security with OGSA. According to the group, the security challenges faced in a Grid environment can be grouped into three categories:
- Integration solutions where existing services needed to be used, and interfaces should be abstracted to provide an extensible architecture;
- Interoperability mechanism so that services hosted in different virtual organizations that have different security mechanisms and policies will be able to invoke each other;
- Solutions to define, manage and enforce trust policies within a dynamic Grid environment.
A Grid system presents unique security problems that are not addressed by traditional client-server/distributed computing environments. In addition to providing basic security requirements like authorization, authentication, confidentiality and integrity, a Grid security infrastructure must be able to support more advanced security features like dynamic delegation of access rights, single sign-on/sign-off, dynamic establishment of trust relationships among multiple domains, privacy and policy related security issues in federated environments.
The first, most notable security architecture for Grids was proposed by Foster I. et al in (Foster, I. et al., 1998). This architecture addresses several unique Grid security requirements. Today the world is witnessing the convergence of Grid and web services. The security requirements of Grid services overlap deeply with the security requirements of web services. So the architecture proposed in (Foster, I. et al., 1998) can be examined further for its applicability in today’s Grid and web services world.
Moreover, we must add all the features and particularities of security offered by mobile computing, that due to the limited devices, to unstable and open networks, to mobility, etc. play an important role in the development of mobile environments. In a mobile environment, the user roams through different networks with heterogeneous security infrastructure. In such an environment where device mobility and network mobility is an essential, offering homogenous facility over heterogeneous devices and networks is the key. In such environment weak security connection from a wireless network could become a socket of vulnerability for the complete system. Hence, in a mobile computing environment, it is compulsory to have a robust security and trust infrastructure.
Therefore, security in a Mobile Grid system should cover all aspects of security, both of Grid computing and of mobile computing, integrating both proposals and providing a security infrastructure for Mobile Grid systems solid, comprehensive, reliable, scalable and interoperable, that is able to provide security solutions to the problems, weaknesses, attacks and vulnerabilities, of any magnitude, found in this kind of systems.
As reflected in recent research, theprinciplewhich establishesthat the building of security into the early stages of the development process is cost-effective and also brings about more robust designs is widely-accepted. The biggest problem, however, is that in the majority of software projects security is dealt with when the system has already been designed and put into operation. Added to this, the actual security requirements themselves are often not well understood. This being so, even when there is an attempt to define security requirements, many developers tend to describe design solutions in terms of protection mechanisms, instead of making declarative propositions regarding the level of protection required.
The special security requirements of Grid applications derive mainly from the dynamic nature of Grid applications and the notion of virtual organization (VO), which requires the establishment of trust across organizational boundaries. In this kind of environment, security relationships can be dynamically established among hundreds of processes spanning several administrative domains, each one with its own security policies. As a result, the Grid security requirements are complex and pose significant new challenges.
There are most common general security requirements and challenges associated with Grids and Mobile computing are presented below:
- Authentication: Authentication mechanisms and policies are supposed to constitute the basis on which local security policies can be integrated within a VO. Difficult issues with respect to authentication in Grids are scalability, trust across different certification authorities, revocation, key management, and delegation. Since processes with delegated authority act on behalf of their owner, there is a question of authentication in delegation, which becomes even more complex when delegation is chained. Authentication processes using asymmetric cryptography are computationally very expensive, which in our case becomes a problem with regard to battery power. Special care has to be taken when choosing algorithms for both symmetric and asymmetric cryptography.
- Confidentiality: The Nature of Grids forces data to be stored in accessible online databases. Confidential code may be requested to execute on a remote host, and confidential data may need to be used at remote locations. Data may also need to be replicated at multiple sites, and thus should be stored in an encrypted form and remain consistent throughout. This is accentuated if the access to data is made over wireless networks. Finally, laws regarding privacy rights and encryption vary among countries and must be taken into account when deploying Grid technologies across international borders.
- Integrity: Many applications have strong code or data integrity concerns. The trust status of remote resources is important when data arises from remote processing as the accuracy of results can be trusted only to the extent that the remote host generating the data is trusted. Integrity is also an issue with regard to delegation, since the set of rights that have been delegated must not be modified maliciously. In the wireless networks, messages cannot be modified in transit between the wireless clients and the access point in an active attack.
- Authorization and Access control: Authorization refers to the ability to control the level of access that individuals or entities have to a wireless network or resource and how much information they can receive. In Grids local access mechanisms should be applied whenever possible, and the owner of a resource should be able to enforce local user authorization. A resource provider in a Grid environment must have reached some form of agreement with the VO to allow the use of the resource. The VO may wish to specify a portion of the resource usage policies, to manage jobs running on VO resources, or to give some group of users the ability to manage those jobs. The authorization policy system must thus be able to combine policies from the resource owner and the VO, express policies about resource usage, manage VO-wide jobs and resource allocations, and dynamically enforce fine-grained policies about resource usage.
- Revocation: Revocation is crucial for authentication in case of a compromised key and for authorization when a VO is terminated or a user or mobile user proves untrustworthy.
- Distributed trust: Trust is a complex theoretical issue. A Grid must be constructed in a dynamic fashion from components whose trust status is hard to determine. Determining trust relations between participant entities in the presence of delegation is significant, and delegation mechanisms need to be rely upon stringent trust requirements.
- Freshness: Freshness is related to authentication and authorization and is important in many Grid applications. Validity of a user’s proof of authentication and authorization is an issue when user rights are delegated and the duration of ajob may span several weeks.
- Scalability: A Grid must be easy to extend and capable of progressive replacement in mobile environments. Fault recovery and dynamic optimization should be usually possible, and degradation should happen gracefully.
- Trust: Trust refers to the assured reliance on someone or something. Since VOs can span multiple security realms, trust associations between domains are of principal importance. Sites in a Grid must be able to enter into trust relationships with Grid users, mobile users and maybe other Grid sites as well. In a Grid environment trust is usually established through exchange of identifications, either on a session or a request basis. Due to the dynamic nature of Grid environments, trust can scarcely be established prior to session execution. Mobile users will use resources at various locations, provided by various service providers. It is important to understand the trust issues involved when mobile clients are allowed to use resources of different servers at different locations.
- Single sign-on: A user should be able to authenticate only once, whereupon he may acquire, use and release resources without further authentication in different domains of the Grid. Users may want to initiate computations running for long periods of time without needing to remain logged on all the time.
- Delegation: Privilege delegation for operations executed by a proxy is a basic requirement for Grid environments, among other reasons in order to satisfy the single sign-on requirement. Delegation of user rights depends upon the security requirements of the application. Delegation is hard to achieve securely in practice, since enabling the delegation of a user’s rights gives rise to many unresolved subtle issues and has a great impact on the overall security of a system.
- Privacy: Privacy is the ability to keep information from being disclosed to determined actors. Privacy can be important in many Grid applications, for instance in medical and health Grids. It is also very important in mobile devices with limited memory and whose access is through wireless networks.
- Non-repudiation: Non-repudiation refers to the inability to falsely deny the performance of some action. It is especially important in e-commerce involving money transactions and mobile environments. With the advent of Enterprise Grid this requirement becomes very important.
- Credentials: Inter-domain access requires a uniform way of expressing the identities of users or resources, and must thus employ a standard for the encoding of credentials. In many scenarios, a job initiated by a user may take longer than the life span of the user’s initially delegated credential. In those cases, the user needs the ability to be notified prior to expiration of the credentials, or the ability to refresh those credentials such that the job can be completed.
- Exportability: Code is required to be exportable and executable in multinational test-beds. As a result, bulk encryption cannot be required.
- Secure group communication: Authenticated communications for dynamic groups is required since the composition of a process group may change dynamically during execution.
- Multiple implementations: It should be possible to enforce security requirements with distinct security technologies and mechanisms.
- Interoperability: In the context of mobile Grids, interoperability means that services within a single VO must be able to communicate across heterogeneous domains. Interoperability guarantees that services located in different administrative domains are able to interact at multiple levels.
- Interoperability with local security solutions: Access to local resources is normally enforced by local security policies and mechanisms. Interoperability between sites and domains with different local policies is necessary in a mobile Grid environment. In order to accommodate inter-domain access, one or several entities in a domain may act as agents of external entities for local resources.
- Integration: In order to allow the use of existing services and resources, integration requirements call for the establishment of an extensible architecture with standard interfaces. Security integration is facilitated by the use of existing security mechanisms. The latter is also in part a consequence of the requirement for site autonomy with regard to security policies, and also of the fact that no single security technology would be able to address the inherent complexity of Grid computing.
- Uniform credentials and certification infrastructure: A common way of expressing
identity, e.g. by a standard such as X.509, is necessary for inter-domain access.
- Secure Logging: Provide all services, including security services themselves, with
facilities for time-stamping and securely logging any kind of operational information or event in the course of time - securely meaning here reliably and accurately.
- Assurance: Provide methods to qualify the security assurance level that can be expected of a hosting environment.
- Manageability: Explicitly recognize the need for manageability of security functionality within the OGSA security model. For example key management, policy management, identity management and so forth.
- Firewall traversal: A major obstacle to dynamic, cross-domain Grid computing today is the existence of firewalls. As mentioned above, firewalls offer limited value within a dynamic Grid environment. Though, it is also the case that firewalls are not likely to disappear anytime soon. Consequently, the OGSA security model must take them into account and provide mechanisms for cleanly traversing them—without compromising local control of firewall policy.
A wide set of these security requirements are shared by mobile environments but considering that we are working with wireless networks and mobile devices with limited capacities. For that reason, issues as to encrypt data transmissions to and from the device and on the devices themselves, to protect the identity (fMEf) of the device from being changed (i.e., cloning the device for illegal resale and use), to measure the trustworthiness of the hardware, OS and applications to detect an unauthorized configuration, to allow IT staffers to deactivate, lock and/or wipe devices which have been stolen or lost, to provide strong user authentication both to activate the device and to access the network, management functions on the device and on the back-end which allow IT staffers to rollout, centrally create, change and enforce their security policies, or password protection on all devices at power-on must be taken into account in each security requirement. Besides, we have to add an important requirement for mobile environments:
- Anonymity: Anonymity is the state of being not identifiable within a set of principles (Pitzmann, A. & Köhntopp, M., 2001). Preserving anonymity is of greater concern in mobile systems for several reasons. Mobile systems yield more easily to eavesdropping and tapping, compared to fixed networks, making it easier to tap into communication channels and obtain user information. Different degrees of anonymity can be provided, such as hiding user identity from eavesdroppers or from certain administrative authorities.
- Mobility: Because mobile devices come with many capabilities, mobile applications must run on a wide variety of devices, including the devices embedded in various environments and devices carried by users. Applications must also support varying levels of network connectivity. Ideally, an application is hosted on the network and is able to execute on any device with multiple levels of connectivity.
- Self-organization: The wireless networks topology must be adapted in case of node or system compromise and failure. If a malicious node discloses the network topology, routing establishment paths may be affected as well. One critical factor is that many wireless systems are mobile, and this mobility affects self-organization.
In a wireless environment, battery power and frequency channels limit the message routing among mobile units. Traditional wired networks normally do not have such limitations, and their units also have more powerful network computing capabilities than that of mobile unit.
Besides these limited resources, the mobility characteristics of the mobile units are quite different from the computing units of the wired networks. In the wired networks, the movement of the units and the network topology changes are rare. But it often happens in a distributed mobile wireless computer network, especially for the wireless mobile ad-hoc network comes from its inherent characteristics of high uncertainty. Therefore, to deal with the routing problems within a distributed wireless mobile computer networks, one should resort to the concepts that are different from that of the traditional wired data computing network architectures [1-5].
Most of the wireless data computing networks can be classified into two types of architectures: 1) base station supported, such as the wireless Ether LAN and 2) no base station supported, such as the wireless mobile ad-hoc networks. For the wireless LAN, the base stations generally support any network. For the wireless LAN, the base station generally support any network computing needed of its subordinate unit for their local network access and services. The wireless mobile Ad- hoc network must be depend on its subordinate units for their local network access and services. The wireless mobile Ad-hoc network must depend on its subordinate unit to do their own network computing and all these units in this network act as dynamic distributed routers [6-8].
The main purpose of the various routing algorithm is to guarantee an error free message transmission from source to destination and maintain its correctness. There are three major algorithms utilized in today’s traditional wired networks, namely the distance vector routing, the link state routing and the source routing. These three algorithms try to take the shortest path approached to route data in a network. However, there are several inherent weakness with these algorithms.
The link state routing algorithm uses a centralized approach where heavy route computations and the broadcast of massive maintenance information increase the load on the units and the consumptions of extra power. Also, the maintenance information should be broadcast in relatively short duration to respond to the dynamic environments. The massive broadcast may degrade the data transmission traffic. The route computation time could be worse since it is propositional to the cube of the number of the units in the network.
The distance vector routing is based on the distributed Bellman-Ford algorithm. The inherent weakness of this algorithm are two-fold: the periodical broadcasting may cause network congestion and consume extra power, and the other drawback is the oscillation problem. The oscillation problem occurs when selected route increases its routing distance, which makes the selection of another route next time. However, the second route increases the routing distance that makes the first route more desirable at the subsequent routing. Eventually, the routing will be oscillatory between the two routes. This problem is caused by the feedback effect between link lengths and routing updates. For a highly dynamic environment, the previous broadcasting of link information may be unable to reflect the newest changes of links in time.
The source routing algorithm has been used in today’s bridged local area networks. This routing lets a source unit determine a complete sequence of units through which to forward packets to the destinations, and explicitly lists this route in the header of the forwarding packets. This routing algorithm tries to discover a route from the source to the destination by broadcasting an exploratory message though the extension of the network. The destination sends a route-confirmed message back to the source unit through each possible loop-free route. From the many possible routes, the one that satisfies the shortest path criterion is selected. The address of all the units in the route are included in the header of each message communicated between the source-destination pairs, all the units have to periodically broadcast message throughout the network and these actions may causes the problems of link congestion and power consumption.
In 1996, Johnson and Maltz proposed the dynamic source routing (DSR) algorithm. This routing algorithm is based on the source routing algorithm with improvement for the dynamic environment. It is explicitly designed for use in the wireless environment of an ad-hoc network. Because it does not periodically broadcast the routing advertisement, it greatly reduces the network bandwidth overhead and the battery power consumption. In 1997, a new signal stability-based adaptive routing (SSA) algorithm, tried to find the most stable route by analyzing the signal strength. This SSA algorithm has further improved the communication quality for the message routing in a wireless ad-hoc mobile network. Besides these above two algorithms, there have been many attempts from different researchers trying to solve some basic problems for such networks.
Large number and the diverse nature of grid computing resources make the resource management an expressively challenging task. Resource management is the process of managing available resources and system workloads accordingly. It is the way in which resources are allocated, assigned, authorized, assured, accounted and authenticated. The discovery of node/resource that user can access through grid information servers, exchanges with resource or their agents using middleware services, scheduling and deployment is done by broker. Discovery in the grid environment is a complex task as resources are geographically distributed, heterogeneous in nature and owned by different organizations each having their own resource management policies and access and cost models. The basic service in the grid computing is resource discovery. There are several methods are proposed for resource discovery. Deniz Cokuslu, Abdelkader Hameurlain and, Kayhan Erciyes, proposed node discovery based on Centralized and Hierarchical Architectures. Author synthesized and analyzed some recent grid resource discovery methods which are based on centralized and hierarchical systems. Systems are evaluated by defining some qualitative criteria, and compared different classes of methods with each other. Ali Sarhadi, Ali Yousefi and Ali Broumandnia proposed resource discovery with Dynamic Structure of Peer-To-Peer Model Based on Learning Automata. This system uses learning automata to generate dynamic grid information. Learning automata based peer-to-peer approach used to increase the efficiency of resource discovery service in the grid computing. This system has two points 1) Highly dynamic information system is required, 2) Organization can be created using learning automata.