Content-aware Trust Statement for semantic Grid
1
Li Wang 1, Wenli Wu2, YingJie Li 1,XueLi Yu1 College of computer and software, TaiYuan University of Technology,Taiyuan,Shanxi,030024 2 Taiyuan City vocational college, Taiyuan,Shanxi, 030027
[email protected] Abstract
Grid computing is a dynamic system and trust changes over time. A kind of flexible trust management is urgently required that could obtain new, not foreseen trust value by computing or reasoning. Semantic grid makes it easier for resources to be discovered and joined up automatically. But there are not explicit semantic content for trust management. By taking full advantage of the content of semantic grid service, trust content model and content-included trust statement model are designed in this paper. In order to obtaining new trust value by the content of trust semantic and grid service semantic information, semTrustStatement ontology, service content model, context model, situation model are defined and the influence relationship between them are shown. At last an example is taken verifying the method valid.
1. Introduction
Grid computing is a distinct distributed computing field, which focus on large-scale resource sharing and collaboration over enterprises, virtual organizations boundaries. Then trust management is a key issue and it is one way to realize grid security. Current Grid security is to establish trust relationships through X509-based digital certificates, security assertions and role-based access management [1]. However, such security mechanisms are too rigid. Trust needs to be established at all levels: Authentication, Policy based management, Business Rules [2]. Grid computing is a dynamic system and trust is a dynamic concept, i.e., it changes over time. So a kind of flexible trust management is urgently required that could obtain new, not foreseen trust worth by computing or reasoning. In [3] a structure of the trust assertion is proposed: PolicyName: trust(Trustor, Trustee, ActionSet, Level)? ConstraintSet (1) In this statement, ActionSet refers to the action undertaken by trustee .Level is an integer ranging from -100 to +100 showing trust rank. ConstraintSet is a Boolean value constraining the trust condition. All of this states that a Trustor trusts a Trustee to perform the actions listed in the ActionSet to a Level, given that the ConstraintSet hold true. There are three problems about this trust statement.
Proceedings of the Second International Conference on Semantics, Knowledge, and Grid (SKG'06) 0-7695-2673-X/06 $20.00 © 2006
1. There is some limitation in it using one level value because trust has rich and complex meaning. 2. It only points out one trust case between A and B. If information could be generalized, extracted and then some knowledge could be formed, more hidden unforeseen information, more trust instances should be gotten by reasoning or computing. 3. Semantic Grid utilizes service semantic information for resources to be discovered and joined up automatically, but there is no clear semantic support for trust management. In fact, some semantically description of semantic grid service is related to trust and to some extend it decides trust. Ontological technology will play a key role in semantic grid trust management at content-related level. This paper puts forward with content-included trust statement model to resolve the first problem, builds trust statement ontology for the second( in section 2) , establishes trust content model , some models about service and the influence relationship between them to deal with the last problem (in section 3). Finally, an example is taken verifying the method valid and some future work is given.
2. Trust content and content-included trust statement
Trust is a rich connotation concept and we build a trust content model is Tcontent={competence, honesty, security, reliability ,timeliness, recommendation} (2) Trust_coefficient= (3) (4) Trust_value= Level in the trust statement should be a result synthesizing trust value in all aspects. But in (1), it is obviously that ActionSet in trust statement only focus on service’s function. So it is necessary to point out the trustiness estimation of other aspects. We put forward a new component TrustType that just exhibits such meaning. In order to easy-use, we build a one-to-one mapping relationship between integer (ranging from 1 to 721)and different trust preference combination that the importance for each element is ranked by an integer ranging from 1 to 6 and the default value of TrustType is 721 that is every aspect makes equivalent effect ( its factor value is 3) on trustiness.
Trustlevel is a float between 0 and 1, which is a synthesized number from trust factors of trust content .The synthesized algorithm is as follow: (5) TrustLevel=(∑factoriⅹvaluei)/21 In social field, trusting others also implies some risk. So risk evaluation is necessary when asserting a trust statement. We take the same method dealing with risk value as trust level. Then the content-included trust model is semTrustStatement= ?constraintSet (6) We construct TrustContent ontology as fig 1.
Fig 1 the graph of TrustContent ontology
3. Semantic grid service based contentaware trust In order to making full use content of semantic grid service, from the trust point of view, we give three grid service definitions and models concerning trustiness. Definition1. Grid Service Context is the profile of grid service provider. Its model is {Organization_Profile,Precondition,Postcondition,Emo tion,theThirdRecommendation}.Organization_Profile means the organization profile information of semantic grid services and it influences the security and honesty in trust content. Precondition and Postcondition are complex entities represented by a logic expression, which includes begin time constraint, end time constraint, and other some control information and they influence the competence and timeliness of trust. ThirdRecommendation influences the whole trust content and Emotion influences honesty. Definition2. Grid Service Content is the functionrelated information of grid service and its model is {input, output, topic},which mainly influence the competence and reliability of trust content. Definition3. Grid Service Situation is the physical attributions of grid services such as network flow, bandwidth, and traffic and so on, which decide the trust content at the physical level.
Proceedings of the Second International Conference on Semantics, Knowledge, and Grid (SKG'06) 0-7695-2673-X/06 $20.00 © 2006
4. A case
Fig 2 the graph of SaleTrustStatement ontology
Here is a SaleTrustStatement ontology and an instance of buying car. Although there is not Salecartrust statement, but it is a subclass of SaleTrustStatement, it could be entailed all the attributes of SaleTrustStatement. For example Ming wants to purchase a car, but he did not touch with car seller Zhang. He knows by the third part that Zhang has been sold cars since 1981 and his selling plan has been done till 2008.He also knows that Zhang has a good habit that he agrees customers using goods free half month and then pay for it if satisfaction. But he could not get Zhang trust value and risk value in selling cars. He has a pleasure experience with a cars seller whom Zhang has the almost same experience as. Ming computes the trust value of Zhang by counting their context, content ontology semantic similarity, and then make his decision.
5. Conclusion
With development of semantic technology, new flexible trust management will become promising. After the work done in this paper, we would research the method to evaluate trust by counting and reasoning with grid semantic trust and services. Acknowledgement: Supported by ShanXi Science Foundation (No.2006021015,20051035), ShanXi Facilities and Information Infrastructure for Science and Technology Project ， ShanXi returnee foundation.
References [1]D. Olmedilla,O. F Rana,B. Matthews,W. Nejdl.Security and Trust Issues in Semantic Grids. Proc. Schloss Dagstuhl Seminar No. 05271: Semantic Grid: The Convergence of Technologies, 04-08 Jul 2005 [2] B. Matthews, J. Bicarregui, T. Dimitrakos. Building Trust on the GRID.Trust Issues Underpinning Scalable Virtual Organisations.http://epubs.cclrc.ac.uk/bitstream/643/trustedg ridERCIM.pdf [3] A. Arenas, B. Matthews, M. Wilson, J. Grant, ILRT. Vocabularies and Architecture for Implementing Trust in the Semantic web. http://www.w3.org/2001/sw/Europe/reports/pdf/11.2a.pdf
