Research On Dynamic Inference Control Of XML Data

As one of the main means of information transmission and dataexchange on the Internet, it needs to have multiple mechanisms to ensurethe data security of XML. Inference control is one of security mechanisms.According to the difference of implementation mechanisms and runtimephases, the inference control of XML can be divided into static inferencecontrol and dynamic inference control. In this paper, the XML dynamicinference control detects whether the sensitive information is disclosed ornot, according to the information obtained at the query and the informationalready obtained before. This method ensures the security of the sensitivedata, and improves the availability of the XML data.Unlike in the database system, user uses XPath expressions to queryXML data, these query expressions itself contains a lot of XMLinformation. This paper starts from the semantics of XML, studies theinformation containment relationship of different fragments of XMLdocument. If a user got a result after querying the XML document using XPath expressions, it shows that the path of the expression is contained inthe XML document. The user can get the XML information of the nodecontained in the path, and the relationship between the nodes. According tothe semantics of XML, XML information obtained by different queries hasa containment relationship.This paper also proposes a method to combine the XML informationset based on the XML information containment relationship and XMLkeywords. The XML information obtained by the XPath query can bedescribed by the pattern tree. In order to save all the XML informationobtained by the user to the history file, we need save the return informationof each query to the XML information set in the history file. With the helpof XML information containment relationship and XML keywords, we caneliminate the redundant information, merge the different part of the sameelement, and simplify the structure of XML information set.In our research, we get a solution of XML dynamic inference controlfor single user access. We design the inference rules with variables, andget the instance of the inference rules by replace the variables with thestrings contained in user’s query. We construct the inference-closed XMLinformation set by these instances. The inference-closed XML information set surely contains all the information of the user. We achieve inferencecontrol by detecting whether the XML information set contains thesensitive information.