Jinbao Wang

Indexing multi-dimensional data in a cloud system

Providing scalable database services is an essential requirement for extending many existing applications of the Cloud platform. Due to the diversity of applications, database services on the Cloud must support large-scale data analytical jobs and high concurrent OLTP queries. Most existing work focuses on some specific type of applications. To provide an integrated framework, we are designing a new system, epiC, as our solution to next-generation database systems. In epiC, indexes play an important role in improving overall performance. Different types of indexes are built to provide efficient query processing for different applications.n In this paper, we propose RT-CAN, a multi-dimensional indexing scheme in epiC. RT-CAN integrates CAN [23] based routing protocol and the R-tree based indexing scheme to support efficient multi-dimensional query processing in a Cloud system. RT-CAN organizes storage and compute nodes into an overlay structure based on an extended CAN protocol. In our proposal, we make a simple assumption that each compute node uses an R-tree like indexing structure to index the data that are locally stored. We propose a query-conscious cost model that selects beneficial local R-tree nodes for publishing. By keeping the number of persistently connected nodes small and maintaining a global multi-dimensional search index, we can locate the compute nodes that may contain the answer with a few hops, making the scheme scalable in terms of data volume and number of compute nodes. Experiments on Amazons EC2 show that our proposed routing protocol and indexing scheme are robust, efficient and scalable.

Efficient Skyline Computation on Big Data

Skyline is an important operation in many applications to return a set of interesting points from a potentially huge data space. Given a table, the operation finds all tuples that are not dominated by any other tuples. It is found that the existing algorithms cannot process skyline on big data efficiently. This paper presents a novel skyline algorithm SSPL on big data. SSPL utilizes sorted positional index lists which require low space overhead to reduce I/O cost significantly. The sorted positional index list Lj is constructed for each attribute Aj and is arranged in ascending order of Aj. SSPL consists of two phases. In phase 1, SSPL computes scan depth of the involved sorted positional index lists. During retrieving the lists in a round-robin fashion, SSPL performs pruning on any candidate positional index to discard the candidate whose corresponding tuple is not skyline result. Phase 1 ends when there is a candidate positional index seen in all of the involved lists. In phase 2, SSPL exploits the obtained candidate positional indexes to get skyline results by a selective and sequential scan on the table. The experimental results on synthetic and real data sets show that SSPL has a significant advantage over the existing skyline algorithms.

The Roles of Social Network Mavens

This paper studies social influence from the perspective of users characteristics. The importance of users characteristics in word-of-mouth applications has been emphasized in economics and marketing fields. We model a category of users called mavens where their unique characteristics nominate them to be the preferable seeds in viral marketing applications. In addition, we developed and verified methods to learn their characteristics from a real dataset. Also, we illustrated ways to maximize information flow through mavens in social networks. Our experiments show that our model successfully detected mavens as well as fulfilled significant roles in maximizing the information flow in a social network comparing to the spread that was a result of traditional influencer users in influence maximization problem. These results showed the compatibility of our model with real marketing approaches.

Ad-hoc aggregate query processing algorithms based on bit-store for query intensive applications in cloud computing

Ad-hoc Aggregate query is extremely important for query intensive applications in cloud computing which extracts valuable summary information on massive datasets to help the decision-maker make right decisions. Current data storage schemes (row-store and column-store) cannot efficiently answer ad-hoc aggregate query on massive data sets in cloud computing. A new data storage structure (bit vector storage structure, bit-store for short) is proposed in this paper. The paper focuses on proposing ad-hoc aggregate query algorithms based on bit-store. Firstly, the storage model of bit-store including its attribute encoding schemes and bit file organization is introduced. Secondly, different aggregate operations for query processing are presented based on different encoding schemes. Thirdly, cost analysis for different aggregate operations is presented. Finally, the effectiveness and efficiency of the proposed algorithms is showed by the analytical and experimental results.

An Efficient Social Event Invitation Framework Based on Historical Data of Smart Devices

A lot of people prefer to attend activities or do daily workout with partners, friends, or even strangers instead of solo, however, it is not easy to have partners or friends as companies all the time since everyone has different schedule. With the rapid growth of using smart devices and social network applications in our daily life, more and more people start establishing connections with people through the social network. Therefore, designing a framework for organizing social events and activities based on collected historical data of smart devices is an efficient way to encourage people to enjoy activities they like with people they can get along well. In this paper, we proposed a smart social event invitation framework to select invitation receivers for an event organizer based on smart device historical data of owners. Habits and abilities of smart device owners can be summarized from the historical data. The selection algorithms would choose invitation receivers who have time and ability to attend the event. Moreover, in order to guarantee each participant can enjoy participating in the event, we proposed two algorithms, greedy searching and k-core algorithm, to pick invitation receivers. k-core algorithm uses k-core graph theory to ensure that each receiver has at least k friends are invited as well. Also, the proposed framework encourages less active or isolated people via issuing higher priority to them thus improving overall active levels of all members. The simulation results show that our proposed framework has good performance.

TJJE: An efficient algorithm for top-k join on massive data

In many applications, top-k join is an important operation to return the k most important join tuples among the potentially huge answer space according to a given ranking function. PBRJ is an algorithm template that generalizes previous top-k join algorithms. In this paper, our analysis shows that PBRJ needs to maintain a large quantity of candidate tuples on massive data. Based on the analysis, this paper proposes a novel top-k join algorithm TJJE which is suitable for handling massive data. By some pre-computed information, TJJE first estimates an upper-bound on scan depth of each joined table. Then it determines the file that contains the join positional index pairs of the top-k join results. A novel algorithm is proposed to retrieve the required join tuples by a single sequential and selective scan on the joined tables. Finally, the top-k join results are obtained by a single scan on the retrieved join tuples. The correctness proof and cost analysis of TJJE are presented in this paper. Extensive experiments show that TJJE maintains up to three orders of magnitude fewer candidate tuples and obtains up to one order of magnitude speedup compared to PBRJ.

Dirty Data Management in Cloud Database

Data quality problem is caused by dirty data. Massive data sets contain dirty data in higher probability. As an important platform for massive data management, it is necessary to manage dirty data in cloud databases. Since traditional data-cleaning-based methods cannot clean dirty data entirely and are costly for massive datasets, a massive dirty data management method is presented in this chapter to obtain query result with quality assurance. To achieve this goal, a dirty database storage structure for cloud databases as well as a multi-level index structure for query processing is presented. Exploiting this index for a query on dirty data, candidates nodes in the cloud are selected to run and process the query efficiently. This chapter discusses the index structure and index-based query processing techniques. Experimental results show the efficiency and effectiveness of the presented techniques.

iKey: An Intelligent Key System Based on Efficient Inclination Angle Sensing Techniques.

The elderly may have different aspects of inconvenience in their daily life. Among them, many old people have trouble remembering things even just happened hours ago. They often forget whether they have locked the door while leaving so that they may have to return and check. Such situation also happens to many younger people that do not concentrate their mind while locking the door. In this paper, an intelligent key system, iKey, is proposed to solve such problem. It can be deployed on an existing key to detect user’s locking actions and store locking status in the form of time. Related hardware architecture and working process are proposed. The sensing module based on inclination angle sensors is designed to reduce the amount of data generated. Furthermore, efficient locking detection algorithms are proposed accordingly. Such system and techniques can also be applied in knobs or rotating handles of machines and facilities to detect illegal operations and to avoid user’s forgetting to operate them.

Near-Complete Privacy Protection: Cognitive Optimal Strategy in Location-Based Services

Abstract While enjoying the amenities and convenience provided by the location-based service (LBS) in our daily life, wireless device users may surrender their location together with social activity privacy to the LBS provider. The untrusted LBS server has all the information about users in the LBS and it may track them in various ways or release their privacy data to third parties. To address the privacy protection issue, many location obfuscation algorithms behind a location-privacy preserving mechanisms (LPPMs) were proposed but only approached the trade-off between utility and privacy. Unlike the existing approaches, we propose the first methodology and application, to the best of our knowledge, that enables a near-complete location privacy protection for LBS users to achieve an cognitive optimal resolution by employing the social network associated with the LBS to separate the utility and privacy. Evaluation of both simulation and practical smartphone application shows that the proposed methodology could achieve a near-complete privacy protection in terms of entropy and significantly improve the quality of service utility.

Protecting query privacy with differentially private k -anonymity in location-based services

Nowadays, location-based services (LBS) are facilitating people in daily life through answering LBS queries. However, privacy issues including locationprivacy and queryprivacy arise at the same time. Existing works for protecting queryprivacy either work on trusted servers or fail to provide sufficient privacy guarantee. This paper combines the concepts of differential privacy and k-anonymity to propose the notion of differentially private k-anonymity (DPkA) for queryprivacy in LBS. We recognize the sufficient and necessary condition for the availability of 0-DPkA and present how to achieve it. For cases where 0-DPkA is not achievable, we propose an algorithm to achieve 𝜖-DPkA with minimized 𝜖. Extensive simulations are conducted to validate the proposed mechanisms based on real-life datasets and synthetic data distributions.