Hinke, T. H. and Delugach, H. S. (1992). Aerie: An inference modeling and detection approach for databases. In Thuraisingham, B. and Landwehr, C., editors, Database Security, VI, Status and Prospects: Proceedings of the IFIP WG 11.3 Workshop on Database Security, pages 179--193, Vancouver, Canada. IFIP, Elsevier Science Publishers B.V. (North-Holland).  Home/Search   Document Not in Database   Summary   Related Articles   Check

....(e.g. it uses a great quantity of fuel) Data mining could findsuch a relationship; for example bases X, Y, and SB use an unusual quantity of fuel in relation to other supplies. If we know that bases X and Y support the SSA, we can make a good guess that SB does as well. Hinke and Delugach [Hinke and Delugach, 1992, Hinke et al. 1997] give a breakdown of inference into seven classes of problems. The firstsix rely on combining known rules (e.g. Part A is only used on an SSA) with non sensitive data (Part A was suppled to base SB) to infer sensitive facts. Class 7 is the inference of a sensitive rule. It is ....

Hinke, T. H. and Delugach, H. S. (1992). Aerie: An inference modeling and detection approach for databases. In Thuraisingham, B. and Landwehr, C., editors, Database Security, VI, Status and Prospects: Proceedings of the IFIP WG 11.3 Workshop on Database Security, pages 179--193, Vancouver, Canada. IFIP, Elsevier Science Publishers B.V. (North-Holland).

....SeaView [93, 90] ASD Views [52] and SWORD [112] projects are examples of this approach. Other works have also been done to provide tools which allow data designer to analyze a database schema for potential inference problems and remove those. DISSECT [53, 108] Multilevel Semantic Net [138] IAT [70], and Database Inference Controller [21] are examples. There are some non reference formal models which can be used to verify any good design against them as well. The works of [126] 130] 70] 18] and [86] are examples. In the second approach, the query processor is augmented with a ....

....potential inference problems and remove those. DISSECT [53, 108] Multilevel Semantic Net [138] IAT [70] and Database Inference Controller [21] are examples. There are some non reference formal models which can be used to verify any good design against them as well. The works of [126] 130] [70], 18] and [86] are examples. In the second approach, the query processor is augmented with a logic based inference engine to handle inferences during query processing. The inference engine will attempt to prevent users from the disclosure of the protected information. Some researchers argue that ....

T. H. Hinke and H. S. Delugach. AERIE: An Inference Modeling and Detection Approach For Databases. In B. M. Thuraisingham and Landwehr, editors, Database Security VI, pages 179--193. Elsevier Science Publishers B. V. (North-Holland) IFIP, 1993.

....and by providing a static security flaw detection mechanism. 1. 1 Related Work There are many researches on statically determining whether a user can infer sensitive information in the database especially in the context of relational database systems [Bur90, MJ88, LDS 90, JS91, S O91, HD92, QSK 93, Qia94] Those researches focus on whether users can know the existence of some entities or can make sensitive associations between entities or values in a database, while we focus on different aspect, i.e. whether a user can compute sensitive values from supplied values. Mor87, ....

Thomas Hinke and Harry S. Delugach. Aerie: An inference modeling and detection approach for databases. In Database Security VI: Status and Prospects, pages 179--194. IFIP WG 11.3, Aug. 1992.

....level is equal to the security level of the v. v:av level = v i :av level for some i, 1 i n. 5.2 Inference Inference problem occurs when a user can deduce (or infer) information from a collection of individual accesses against a database. Solutions to the inference problem were proposed in [11, 22, 17, 15, 27, 7] among others in the context of statistical and relational databases. Several approaches are used to handle the problem: 1) place restrictions on the set of allowable queries that can be issued by a user; 2) add noise to the data; and (3) augment a database with a logic based inference engine to ....

T. H. Hinke and H. S. Delugach. AERIE: An Inference Modeling and Detection Approach For Databases. In B. M. Thuraisingham and Landwehr, editors, Database Security VI, pages 179--193. Elsevier Science Publishers B. V. (North-Holland) IFIP, 1993.

....the attributes properly. However, redesigning the database schema results in data duplication which leads to update anomalies. It also requires modifications to the existing application programs. There is also work on incorporating external knowledge into the inference detection systems [18, 7, 16, 17, 3]. More recently, researchers suggest using data of the database to generate a richer set of functional dependencies for inference detection. Hinke et al. use cardinality associations to discover potential inference paths [8] Hale et al. incorporate imprecise and fuzzy database relations into ....

T. H. Hinke and H. S. Delugach. Aerie: An inference modeling and detection approach for databases. In B. M. Thuraisingham and C. E. Landwehr, editors, Database Security VI: Status and Prospects, pages 179--193. North-Holland, 1993.

....[5,15,18] have developed an interactive tool, DISSECT, for detecting and eliminating compositional inference channels due to foreign key FDs. The DISSECT model builds on earlier work on inference control, including tools and techniques developed by Buczkowski [1] Thuraisingham [23,24] and Hinke [9]. The current version of DISSECT [18] is limited to analyzing MLS database schemas (intensions) rather than actual MLS relations (extensions) Nevertheless, its success shows that it is possible to develop practical tools for dealing with the difficult problem of inference control. Effectively ....

T.H. Hinke and H. Delugach, Aerie: An inference modeling and detection approach for databases, Proceedings of the IFIP TC11/WG11.3 Sixth Working Conference on Database Security , 1992.

....for easier analysis; whereas the GEM was constructed from a collection of MKC s, the layers and facets are obtained by slicing the GEM into domains based on the nature of their relationships. The three analysis layers correspond roughly to the inference classes identified in our previous work [2] [8]. Figure 3 shows their general contents. The entitylayer contains knowledge of entities themselves and relationships between entities, e.g. entity part of, is a,orfunctionallydetermines. The activitylayer contains knowledge of activities themselves and relationships between activities , e.g. ....

....concepts they connect) within the entity layer. Fig. 3 shows relations in other facets; e.g. a funcdep facet (for functional dependencies) a temporal facet (with before and after relations) etc. Later in the paper we define each of the facets wehave so far identified 2 . Earlier papers [8], 10] called these facets layers and termed the MKC aLayered Knowledge Chunk (LKC) but the term was changed to facet to eliminate any connotation of a hierarchical relationship. 2 We said that our second subsystem constructs layers and facets. A facet is just a projection of the Global ....

[Article contains additional citation context not shown here]

Thomas H. Hinke and Harry S. Delugach. "AERIE: An Inference Modeling and Detection ApproachFor Databases". In B. W. Thuraisingham and C. E. Landwehr, editors, Database Security, VI: Status and Prospects,number A-21 in IFIP Transactions, Amsterdam, 1993. Elsevier Science Publ. (NorthHolland) .

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC