Recently, works on integrating HTM with STM, called hybrid transactional memory (HyTM), have intensively studied. However, the existing works consider only the prediction of a conflict between two transactions and provide a static HTM configuration for all workloads. To solve the problems, we proposes a hybrid transactional memory scheme based on both abort prediction and an adaptive retry policy, called HyTM-AP. First, our HyTM-AP can predict not only conflicts between concurrently running transactions, but also the capacity and other aborts of transactions by collecting the information of transactions previously executed. Second, our HyTM-AP can provide an adaptive retry policy based on machine learning algorithms, according to the characteristic of a given workload. Finally, through our experimental performance analysis using the STAMP benchmark, our HyTM-AP shows 12~13% better performance than the existing HyTM schemes.

Abstract Since studies on privacy-preserving database outsourcing have been spotlighted in a cloud computing, databases need to be encrypted before being outsourced to the cloud. Therefore, a couple of privacy-preserving kNN query processing algorithms have been proposed over the encrypted database. However, the existing algorithms are either insecure or inefficient. Therefore, in this paper we propose a privacy-preserving kNN query processing algorithm via secure two-party computation on the encrypted database. Our algorithm preserves both data privacy and query privacy while hiding data access patterns. For this, we propose efficient and secure protocols based on Yao’s garbled circuit. To achieve a high degree of efficiency in query processing, we also propose a parallel kNN query processing algorithm using encrypted random value pool. Through our performance analysis, we verify that our proposed algorithms outperform the existing ones in terms of a query processing cost.

Recently, privacy-preserving association rules mining algorithms have been proposed to support data privacy. However, the algorithms have an additional overhead to insert fake items (or fake transactions) and cannot hide data frequency. In this paper, we propose a privacy-preserving association rule mining algorithm for encrypted data in cloud computing. For association rule mining, we utilize Apriori algorithm by using the Elgamal cryptosystem, without additional fake transactions. Thus the proposed algorithm can guarantee both data privacy and query privacy, while concealing data frequency. We show that the proposed algorithm achieves about 3-5 times better performance than the existing algorithm, in terms of association rule mining time.

The shielded computation of hardware-based trusted execution environments such as Intel Software Guard Extensions (SGX) can provide secure cloud computing on remote systems under untrusted privileged system software. However, hardware overheads for securing protected memory restrict its capacity to a modest size of several tens of megabytes, and more demands for protected memory beyond the limit cause costly demand paging. Although one of the widely used applications benefiting from the enhanced security of SGX, is the in-memory key-value store, its memory requirements are far larger than the protected memory limit. Furthermore, the main data structures commonly use fine-grained data items such as pointers and keys, which do not match well with the coarse-grained paging of the SGX memory extension technique. To overcome the memory restriction, this paper proposes a new in-memory key-value store designed for SGX with application-specific data security management. The proposed key-value store, called ShieldStore, maintains the main data structures in unprotected memory with each key-value pair individually encrypted and integrity-protected by its secure component running inside an enclave. Based on the enclave protection by SGX, ShieldStore provides secure data operations much more efficiently than the baseline SGX key-value store, achieving 8--11 times higher throughput with 1 thread, and 24--30 times higher throughput with 4 threads.

Since studies on privacy-preserving database outsourcing have been spotlighted in a cloud computing, databases need to be encrypted before being outsourced to the cloud. Rao et al. proposed a k-Means clustering algorithm that supports the protection of sensitive data by using a paillier cryptosystem[1]. However, the existing algorithm is inefficient due to bit-array based comparison. To solve this problem, we propose an efficient privacy-preserving k-Means clustering algorithm. First, we provide a new secure comparison protocol that performs the fast comparison of encrypted data. Second, we select center points by considering the distribution of the entire data. Finally, we show from our performance analysis that our clustering algorithm achieves about 300% better performance on average than the existing algorithm.

No abstract available.

Hybrid Transactional Memory(HyTM) has recently gained much interest because it combines the advantages of Hardware Transactional Memory(HTM) and Software Transactional Memory(STM). However, the existing HyTM schemes show good performance only for a specific workload. To solve this problem, we propose an adaptive HyTM scheme as efficient synchronization mechanism for multi-core environment. Our HyTM scheme can select its optimal fallback path by using a comprehensive TM coordinator, based on the characteristics of transactions. Finally, we show from our performance analysis using STAMP benchmark that our adaptive HyTM scheme can achieve about 130~160% better performance than the state-of-the-art HyTM ones, i.e., Hybrid NOrec and RH-NOrec.

With the prevalence of cloud computing, privacy-preserving database outsourcing has been spotlighted. To preserve both data privacy and query privacy from adversaries, databases need to be encrypted before being outsourced to the cloud. However, there exists the only kNN classification scheme over the encrypted databases in the cloud. Because the existing scheme suffers from high computation overhead, we proposed a secure and efficient kNN classification algorithm that conceals the resulting class label and data access patterns. In addition, our algorithm can support efficient kNN classification by using our encrypted index scheme and the Yao's garbled circuit. We show from our performance analysis that the proposed algorithm achieves about 17 times better performance than the existing scheme, in terms of classification time.

This paper proposes a novel hybrid transactional memory scheme based on both abort prediction and an adaptive retry policy. First, the proposed scheme can predict not only conflicts between transactions running concurrently, but also the capacity and other aborts of transactions by collecting the information of previously executed transactions. Second, the proposed scheme can provide an adaptive retry policy based on machine learning algorithms, according to the characteristic of a given workload. Finally, through our experimental performance analysis using STAMP, the proposed scheme shows about 20% better performance than Hybrid NOrec, a hybrid version of the efficient NOrec STM.

This paper proposed a novel hybrid transactional memory(HyTM) that exploits the benefits of both Haswell's RTM(restricted transactional memory) and software transactional memory(STM). Unlike the existing HyTMs, the proposed HyTM can predict and resolve conflicts between transaction running concurrently by using a prediction matrix and transaction metadata. Also the proposed HyTM can provide the optimal HTM configuration for a given workload by computing the optimal retry threshold based on deep learning algorithms.

ABSTRACT Outsourcing database to a third‐party data provider is becoming a common practice for data owners to avoid the cost of managing and maintaining the database. Meanwhile, because of the popularity of location‐based services, the need for spatial data is increasing dramatically. However, the most important challenge in database outsourcing is how to meet privacy requirements and guarantee the integrity of the query result as well. Unfortunately, most of the existing techniques support either data privacy or integrity on spatial databases. To carry on both privacy and integrity for outsourced spatial data, we propose a spatial transformation scheme that makes use of shearing transformation with rotation shifting. We describe attack models measuring the data privacy of our transformation scheme. Finally, our extensive experiments have demonstrated that our scheme has adequate strength for data privacy by showing outstanding performance against different kinds of attack models and efficiently handles the query integrity of the query result sets. Copyright © 2013 John Wiley & Sons, Ltd.

Because much interest in spatial database in cloud computing has been attracted, studies on preserving location data privacy in cloud computing have been actively done. However, since the existing spatial transformation schemes are weak to proximity attack, they cannot preserve the privacy of users who enjoy location-based services from the cloud computing. Therefore, a transformation scheme for providing a safe service to users is required. So, we, in this paper, propose a new transformation scheme based on a line symmetric transformation (LST). The proposed scheme performs both LST-based data distribution and error injection transformation for preventing proximity attack effectively. Finally, we show from our performance analysis that the proposed scheme greatly reduces the success rate of the proximity attack while performing the spatial transformation in an efficient way.

SUMMARY Location‐based services (LBSs) are currently some of the fastest growing information technology industries. User privacy in LBSs has attracted great interest in the research community. However, the proliferation of position identifying devices has become an increasing privacy threat for users in LBSs. It is very difficult to avoid the privacy threat of a user in processing his/her request because the user has to submit his/her exact location with a query to the LBS. To protect privacy in road networks, the existing method employs an X‐Star framework to hide the query issuer and provide protection from attack resilience. However, it incurs low anonymization success rate and high computation cost. To solve the problems, we propose a Hilbert‐order‐based star network expansion cloaking algorithm (H‐Star). Our H‐Star guarantees k‐anonymity under the strict reciprocity condition and increases anonymization success rate by reducing computation overhead. We also propose k‐nearest neighbor and range query processing algorithms based on the anonymized region. Through comprehensive experimental analysis, we show the effectiveness of our algorithms in the field of spatial cloaking.Copyright © 2012 John Wiley & Sons, Ltd.

The propagation of position identifying devices, such as GPS (Global Positioning System), becomes increasingly a privacy threat in location-based services (LBSs). However, in order to enjoy such services, the user must precisely disclose his/her exact location to the LBS. So, it is a key challenge to efficiently preserve user's privacy while accessing LBS. For this, the existing method employs a 2PASS cloaking framework that not only hides the actual user location but also reduces bandwidth consumption. However, it suffers from privacy attack. Therefore, we aim to provide the solutions which can preserve user privacy by utilizing k-anonymity mechanism. In this paper, we propose a weighted adjacency graph based k-anonymous cloaking technique that can provide protection to user and also reduce bandwidth usages. Our cloaking approach efficiently supports k-nearest neighbor queries without revealing private information of the query initiator. We demonstrate via experimental results that our algorithm yields much better performance than the existing one.

Wireless sensor networks (WSN) are widely used in the various monitoring systems. When implementing WSN-based monitoring systems, there are three important issues to be considered. At first, we should consider a node failure detection method to provide continuous monitoring. Secondly, because sensor nodes use limited battery power, we need an efficient data filtering method to reduce energy consumption. At last, we should consider a data filtering method for reducing processing overhead. The existing Kalman filtering scheme has good performance on data filtering, but it causes too much processing overhead for estimating sensed data. To solve this problem, we, in this paper, propose a sampling-based data filtering scheme based on statistical data analysis. First, our scheme periodically aggregates nodes' survival massages to support node failure detection. Secondly, to reduce energy consumption, our scheme sends the sampled data including node survival massage and perform data filtering based on the messages. Finally, it analyzes the sampled data to estimate filtering range at a server. As a result, each sensor node can use only a simple compare operation for filtering data. Through performance analysis, we show that our scheme outperforms the Kalman filtering scheme in terms of the number of data transmissions.

Recently, wireless sensor networks (WSN) are actively used for various monitoring systems. While implementing WSN-based monitoring systems, there are three important issues to be considered. At first, we should consider a node failure detection method to provide continuous monitoring. Secondly, because sensor nodes use limited battery power, we need an efficient data filtering method to reduce energy consumption. At last, we should consider a data filtering method for reducing processing overhead. The existing Kalman filtering scheme has good performance on data filtering, but it causes too much processing overhead for estimating sensed data. To solve this problem, we, in this paper, propose a new data filtering scheme based on statistical data analysis. First, the proposed scheme periodically aggregates nodes' survival massages to support node failure detection. Secondly, to reduce energy consumption, the proposed scheme sends the sample data including node survival massage and perform data filtering based on the messages. Finally, it analyzes the sample data to estimate filtering range at a server. As a result, each sensor node can use only a simple compare operation for filtering data. Through performance analysis, we show that the proposed scheme outperforms the Kalman filtering scheme in terms of the number of messages transmission.

Recently, Wireless Sensor Network (WSN) has been broadly studied in ubiquitous computing environments. In WSN, it is important to reduce communication overhead by using an energy-efficient routing protocol because the resources of the sensor node are limited. Although there exist some cluster-based routing protocols, they have some problems. First, the random selection of a cluster head incurs a node concentration problem. Secondly, they have a low reliability for data communication due to the less consideration of node communication range. Finally, data communication overhead is greatly increased while constructing clusters. To solve these problems, we, in this paper, propose a new cluster-based routing protocol using message success rate. To resolve the node concentration problem, we design a new cluster head selection algorithm based on node connectivity and devise cluster maintenance algorithms. Moreover, to guarantee data communication reliability, we use message success rate, which is one of popular measures for data communication reliability, in order to select a routing path. Finally, to reduce data communication overhead, we use only the information of neighboring nodes during both cluster construction and cluster head selection phases. Through our performance analysis, we show that our protocol outperforms existing schemes in terms of communication reliability and energy efficiency.

In this paper, we propose a new privacy preserving data aggregation scheme for WSNs. Our scheme applies additive property of complex numbers in order to combine sensor data and preserve data privacy during transmission to the sink node. In addition, for supporting scalability, we propose a novel mechanism in which a special set of real numbers are assigned to sensor nodes as their IDs so that a single bit is sufficient to hold ID of a sensor node during transmission of aggregated data to the sink node. For this, we, first, generate fixed size signatures for the IDs of all sensor nodes and then superimpose the signatures during data aggregation phase. By analytical evaluations, we show that our scheme is more scalable and energy efficient than the existing methods to achieve data privacy and transmit IDs of sensor nodes along with the aggregated data to the sink node.

Many applications require the privacy of the sampled data while they travel from the source sensor nodes to data collecting device, say data sink. Providing an efficient data aggregation scheme with preserving data privacy is a challenging problem in the research of wireless sensor networks (WSNs). Although the secure data aggregation in WSNs has been well studied in the recent years, there exists a little work, for instance PDA (Privacy-preserving Data Aggregation), which focuses on protecting sensor data not only from adversaries but also from the participating trusted sensor nodes. However, PDA suffers from one main problem which is the high communication cost due to unnecessary traffics in the network during data transmissions. To resolve the problem, we, in this paper, propose a secure aggregation scheme of private data for WSNs. The proposed scheme applies the additive property of complex number to unite sensor data and for their privacy during their transmissions to a sink node. With our analytical performance evaluations, we show that our scheme is more efficient than the PDA scheme in terms of communicational cost.

In Location-Based Services (LBSs), users send location-based queries to LBS servers along with their exact locations, but the location information of the users can be misused by adversaries. For this, a mechanism to deal with the userspsila privacy protection is required. In this paper, we propose a new cloaking algorithm for privacy protection in LBSs. Our cloaking algorithm can support both k-anonymity and l-diversity. That is, it first creates a minimum cloaking region by finding l buildings (l-diversity) and then finds k users (k-anonymity). To generate the minimum cloaking region efficiently, we make use of a grid structure for storing buildings and users as well as a pruning technique for reducing unnecessary computation. Finally, we show from our performance analysis that our cloaking algorithm outperforms the existing grid-based cloaking algorithm, in terms of the size of cloaking regions, their creation time and query processing time based on them.

For wireless sensor networks (WSNs), how to design routing protocols is very important to gather data efficiently so that the life of the network can be prolonged. A usual concept to collect data by a sink node is to transfer data from sensor nodes to the sink node by multihop. However, it gives rise to two problems. The first one is that core sensor nodes being more frequently used to gather data run out of energy sooner than other sensor nodes. As a result, the network loses its service ability, regardless of a large amount of residual energy of the other nodes. The second one is that the network generates unnecessary traffic during data transmission for choosing a proper data sending path. To resolves the problems, we, in this paper, propose a designated path scheme for energy balanced data aggregation in WSNs. The proposed scheme predetermines a set of paths and runs them in round robin fashion so that all the nodes can participate in the workload of gathering data and transferring them to the sink node. With our analytical performance analysis, we show that our designated path scheme is more energy efficient and can prolong the life time of the network, compared with the existing two data centric routing schemes, i.e., directed diffusion and hierarchical data aggregation (HDA).

In Location-Based Services (LBSs), users send location-based queries to LBS servers along with their exact locations, but the location information of the users can be misused by adversaries. In this regard, there must be a mechanism which can deal with the privacy protection of the users. In this paper, we propose a cloaking method considering both K-anonymity and L-diversity. Our cloaking method creates a minimum cloaking region by finding L number of buildings (L-diversity) and then finds K number of users (K-anonymity). To support it, we use R*-tree based index structures as well as efficient filtering techniques to generate a minimum cloaking region. Finally, we show from our performance analysis that our cloaking method outperforms the existing grid-based cloaking method in terms of the size of cloaking regions and cloaking region creation time.

GML is a markup language presented as exchange standard for geographic information by the OGC (Open GIS Consortium). In spatial network databases, researches for supporting GML (Geographic Markup Language) can be divided into parsing, storing and retrieval of GML documents. Among them, the study on the storage of GML documents is essential for their efficient retrieval. However there is little research on the storing of GML documents whereas there have been a lot of researches on the storing of XML documents. Because the storage schema designed for XML documents are not appropriate for geographic information, we, in this paper, propose three storage schemas for efficiently storing GML documents including geographic information in order to solve the problem that the XML storage schema store duplicate data and need to search many tables for obtaining elements. In addition, we design and implement a storage manager which can store GML documents using proposed GML storage schema.

In this paper, we design an intelligent system architecture for dealing with context-aware application services in pervasive computing environment. The context-aware intelligent system architecture is composed of middleware, context server, and client. The middleware component of our intelligent system architecture plays an important role in recognizing a moving node with mobility by using a Bluetooth wireless communication technology as well as in executing an appropriate execution module according to the context acquired from a context server. The context server functions as a manager that efficiently stores into the database server context information, such as user's current status, physical environment, and resources of a computing system. To verify the usefulness of our context-aware intelligent system architecture, we finally develop a context-aware application to provide users with a music playing service in pervasive computing environment.

In this paper, we develop an XML document retrieval system supporting a multimedia Web service for a digital museum. It can support unified retrieval on XML documents based on both document structure and image content. To achieve it, we perform the indexing of XML documents describing Korean porcelains used for a digital museum, based on not only their basic unit of element but also their image color and shape features. In addition, we provide a similarity measure for a unified retrieval to a composite query, based on both document structure and image content. Finally, we implement our XML document retrieval system designed for a digital museum Web service and analyze its performance in terms of retrieval time, insertion time, storage overhead, as well as recall and precision measure.

In this paper, we propose a new spatio-temporal representation scheme for modeling multiple moving objects' trajectories in video databases. For content based video retrieval on moving objects efficiently, our representation scheme considers the moving distance of an object during a given time interval as well as its temporal and spatial relationships. We present a similarity measure algorithm for the multiple moving objects' trajectories

Because content-based image retrieval is essential to retrieve relevant multimedia documents, we represent images as a set of recognizable symbols, i.e., icon objects, and do indexing by regarding the icon object as a representative of a given document. When users request content-based image retrieval, we convert a query image into icon objects and retrieve relevant images in the database. In this paper, we propose a new spatial-match representation scheme, called SRR(Spatial-match Representation supporting Ranking) scheme, which combine directional operators with positional operators. Therefore, our SRR scheme can represent spatial relationships between icon objects precisely and can provide ranking for the retrieved images. In addition, we compare our scheme with the conventional 9DLT and SMR schemes in terms of retrieval effectiveness. Finally, we show from our experiment that our SRR scheme holds about 25% higher recall and about 10% higher precision, compared with the 9DLT and the SMR.

SGML has become very popular as the source language of structured documents because its usefulness has already been verified by the CALS (Commerce At Light Speed) project. Traditional information retrieval systems (IRSs), however, cannot support retrieval based on the logical structure of SGML documents. In this paper, we design a structured IRS (SIRS) supporting both content-based retrieval and structure-based retrieval, and implement it using the O2 Store storage system with the standard C language under the UNIX operating system environment. In order to make it easy to write a user query for SIRS and to obtain SGML documents that are relevant to the query, we also implement a WWW-based user interface by using CGI. Finally, we evaluate the performance of SIRS in terms of the record insertion time, deletion time, retrieval time and storage overhead.

In order to support content based iconic image retrieval efficiently, we propose a new spatial match representation scheme, called SR scheme, which combines directional operators with positional operators. Therefore, our scheme can represent spatial relationships between icon objects precisely, and can provide ranking for the retrieved images. In addition, we compare our scheme with the conventional 9DLT and SMR schemes in terms of retrieval effectiveness. Finally, we show from our experiments that our SR scheme holds about 7% higher recall and about 20% higher precision, compared with the 9DLT and the SMR.

In multimedia information retrieval applications, contentbased image retrieval is essential for retrieving relevant mnltimedia documents.The purpose of our paper is to provide both effective representation and efficient retrieval of images when a pixel-level original image is antomaticahy or mannally transformed into its iconic image containing meaningful graphic descriptions, called icon objects.For this, we propose a new spatial match representation scheme to describe spatial relationships between icon objects preciseIy by using accurate positional operators as well as by expressing objects as rectangles, rather than pointer.In order to accelerate image searching, we also design an efficient retrieval method using a signature file technique.Finally, we show from our experiments that our representation scheme achieves better retrieval effectiveness than the 9-DLT scheme.Pm((ision 10 111nkl: di&l/lmd q&s OI';III or pa11 o(*(liis m;l(rri;d (i,r ~WSOI~~ W ChSSrtWlll USC IS g~nillllcd wi(hou( kc pvided (],a( (llc mpia .lrC11~1 mndc w dislrilwkul IiF pmli( or cwnmerci;tl ;)d\~l)(~g~e (lie cc,l,yrlf$l a(Niw.IIIC lilk ol'lht puhhca(ittn antI i(a d;t(c ;t(,p~:~r.:#lid a&y is #@II 1ll:ll LVp>Tlglll IS lly pmllision ol'llrc .\CB(.ltjc '('0 q+9 ~(l,cn,~ww (0 r~~Idlkh.10 IlOS1 WI .WlTCls or lo rcdis(rihu(c 10 lisls rq"irF s+yi(ic painiswln mid~or Lx,

Due to the wide acceptance of object-oriented database systems, some implementation issues such as query processing and indexing become a crucial factor in the success of object-oriented database systems.To tackle the issues, we propose an s-signature scheme, which is an efficient access method for query processing, especially, when the path dictionary is used.We also compare the s-signature scheme with other indexing techniques such as path index, path signature, and path dictionary.Finally, we demonstrate that the proposed method shows significant improvements in the retrieval operation, while at the same time producing lower storage overhead than the path index.Permission to make digitallhafd copies of all or part of thk material for personal or classroom use is granted without fee provided that the copies are not msde or dLmibuted for pro~t or commeffiial advantage, tie.Gopyright notice, the titte of the pubhcatlon and its date appear, and nottce is given that copyright ia by permission of the ACM, Inc.To copy othetwiae, to republish, to post on servers or to rrdstribute to lists, requires specific

In this paper, we design a new hypermedia markup language using SGML and implement an object-oriented hypermedia system on top of the Postgres. Compared with the conventional systems, our hypermedia system has some advantages. First, since our hypermedia markup language is designed using SGML, the language can interchange documents in a system-independent manner and can support content-based and structure-based retrieval. Second, since we apply an object-oriented paradigm to modeling hypermedia data and links, we can inherit the properties and methods of the object-oriented model. Finally, our hypermedia system can provide database management system functions such as query processing, transaction management, storage management, security, crash recovery, and version control of structured document components as separate objects.

As a result of the wide acceptance of object oriented database systems, some implementation issues such as query processing and indexing become a crucial factor in the success of object oriented database systems. To tackle the issues, we propose an s-signature scheme, which is an efficient access method to nested objects for query processing, especially, when the path dictionary is used. We also compare the s-signature scheme with other indexing techniques such as path index, path signature, and path dictionary. Finally, we present that the proposed method shows significant improvements in the retrieval operation, while at the same time producing lower storage overhead than the path index.

The signature file method has been widely advocated as an efficient index scheme to handle new abdications demand-. .ing a large volume of textual databases.Moreover, it has recently been extended to support multimedia data.In order to achieve good performance when handling multimedia data, the signature file approach has been required to support parallel database processing.Therefore, in this paper we propose a horizontally-divided parallel signature file (HPSF) method using extendible hashing and frame-slicing techniques.In addition, we also propose a heuristic processor allocation method so that we may assign a set of signatures to a given number of processors in a uniform way.To show the efficiency of HPSF, we evaluate the performance of HPSF in terms of retrieval time, insertion time, and storage overhead.Finally, we show from the performance results that HPSF outperforms the conventional parallel signature file methods regarding retrieval performance and dynamic operating measures used to combine both retrieval and insertion time.

In order to improve the two-level signature file method designed by Sacks-Davis et al. [20], we propose new multikey access methods based on term discrimination and signature clustering. By term discrimination, we create separate, efficient access methods for the terms frequently used in user queries. We in addition cluster similar signatures by means of these terms so that we may achieve good performance on retrieval. Meanwhile we provide the space-time analysis of the proposed methods and compare them with the two-level signature file method. We show that the proposed methods achieve 15-30% savings in retrieval time and require 3-9 % more storage overhead.