In this paper, we study the problem of mining frequent diamond episodes efficientlyfrom an input event sequence with sliding a window. Here, a diamond episode is of the form a → E → b, which means that every event of E follows an event a and is followed by an event b. Then, we design a polynomial-delay and polynomial-space algorithm PolyFreqDmd that finds all of the frequent diamond episodes without duplicates from an event sequence in O(|Σ|²l) time per an episode and in O(|Σ|+l) space, where Σ and l are an alphabet and the length of the event sequence, respectively. Finally, we give experimental results on artificial and real-world event sequences with varying several mining parameters to evaluate the efficiency of the algorithm.

View full abstract

Single IP Address cluster offers a transparent view of a cluster of machines as if they were a single computer on the network. In such an environment, process migration can play a significant role for providing services seamlessly and for increasing sustainability. In this paper we propose a live migration mechanism which is capable of moving processes that maintain a massive amount of network connections, supporting both TCP and UDP sockets. Incoming packet loss during socket migration is prevented by exploiting the broadcast property of the Single IP Address cluster, while process live migration minimizes the execution freeze time during the actual migration of the process context. Performance evaluation on machines equipped with a 2.4GHz CPU and Gigabit Ethernet interconnect shows that migrating a process of 1GB image size and over 1000 established network connections results in less than 200ms process freeze time, rendering the transition fully transparent and responsive from the clients' point of view. The implementation is comprised entirely of a kernel module for Linux 2.6, without any changes to the existing kernel code.

View full abstract

The prevalence of multi-tuners, high-definition digital video recorder systems and home networking is increasing the number of simultaneous streams that must be processed by recorder storage devices. Whereas recent hard-disk drives provide enough performance to theoretically handle such workloads, general purpose file systems and I/O schedulers used by operating systems such as Linux do not satisfy the quality of service (QoS) requirements necessary for efficient processing of real-time video streams. In this paper, we introduce the Audio/Video File System (AVFS) composed of a file system and a disk I/O scheduler optimized for handling simultaneous high bit-rate real-time streams. Using a precise QoS measurement method, evaluation results of a Linux implementation of AVFS show that, compared to traditional file systems such as ext3 and JFS, AVFS provides QoS guarantees for real-time streams and more stable performance.

View full abstract

GridRPC is known as an effective programming model to develop Grid applications. However, it is still difficult for non-expert users to apply it efficiently. For example, a GridRPC application user needs to select computational resources, monitor the resources and estimate the application performance on the resources. In this paper, we propose InterS, an interactive scheduling system for GridRPC applications. First, the automatic scheduling mechanism provides resource allocation plans, from which the user can choose the most suitable one. Second, the execution advice mechanism helps the user to improve the performance of the application at run time while overload or failure on the resource(s) is(are) detected. Third, the scheduling policy mechanism provides the user with an interface in ClassAd format to define the scheduling policy applied in InterS. This paper also presents experimental results to show the advantage of interactive scheduling and how they can be performed at run time.

View full abstract

Superinstruction is well-known techniques of improving performance of interpreters. Superinstructions eliminate jumps between VM operations (interpreter dispatch) and enable more optimizations in merged code. In past, processors with simple BTB-based branch predictors had high misprediction rate when executing interpreted code, resulting in high overhead of interpreter dispatch, so superinstructions were used to reduce it. However, this assumption is incorrect for Ruby on current hardware. Accordingly, using superinstructions for eliminating jump instructions only marginally improves performance. In this paper, we consider applying superinstructions differently to improve performance of floating point computation. We note that high percentage of objects allocated during numeric computation are boxed floating point values, meanwhile garbage collection takes significant part of the execution time. Using superinstructions composed from pairs of arithmetic operations we were able to reduce allocation of boxed floats by up to 36%, and obtain improvement in performance of up to 22%.

View full abstract

In this paper, we theoretically analyze a certain extension of a finite automaton, called a linear separation automaton (LSA). An LSA accepts a sequence of real vectors, and has a weight function and a threshold sequence at every state, which determine the transition from some state to another at each step. Transitions of LSAs are just corresponding to the behavior of perceptrons. We develop the theory of minimizing LSAs by using Myhill-Nerode theorem for LSAs. Its proof is performed as in the proof of the theorem for finite automata. Therefore we find that the extension to an LSA from the original finite automaton is theoretically natural.

View full abstract

This paper presents a novel labeling scheme for dynamic XML trees. The scheme employs history-offset encoding method for multidimensional datasets and takes advantage of this method by embedding an XML tree into a multidimensional extendible array. Even if structural updates are made on the XML tree, no relabeling of nodes is required under the support of extra data structure for preserving the document order. The most significant advantage of our scheme over other existing labeling schemes is that the storage cost for generated labels is very small irrespective of the order and the position of node insertions; in most of our competing schemes, the generated label size would become very large if the insertions occur around the same position. After describing our labeling scheme, label size, total label storage cost and node access performance are examined compared with other sophisticated schemes, such as ORDPATH, QED, DLN and Prime Numbering, and proves that our scheme outperforms these schemes in some criteria.

View full abstract

In this paper, we propose a simple but flexible virtual machine consolidation method for power saving. This method is specifically designed for datacenters where heterogeneous high-density blade servers host dozens or even hundreds of virtual machines. This method utilizes an extended First-Fit Decreasing (FFD) algorithm. It selects a migration destination server on the basis of server rank. The rank represents server selection priority and is uniquely assigned to each physical server. Our simulation results show that this method reduces power consumption by 34.5% under a typical workload and 33.8% under a random workload.

View full abstract

Skyline objects in a database are objects that are not dominated by any other object in the database. Skyline queries retrieve a set of skyline objects so that the user can choose promising objects from them and make further inquiries. Therefore, such skyline queries are important for several database applications. However, a skyline query often retrieves too many objects to analyze intensively especially for high-dimensional dataset. Recently, k-dominant skyline queries have been introduced, which can reduce the number of retrieved objects by relaxing the definition of the dominance. On the other hand, the maintenance of k-dominant skyline objects under continuous updates is much more difficult compared to conventional skyline objects. This paper addresses the problem of efficient maintenance of k-dominant skyline objects of frequently updated database. We propose an algorithm based on divide and conquer strategy for maintaining k-dominant skyline objects. Intensive experiments using real and synthetic datasets demonstrated that our method is efficient and scalable.

View full abstract

As demand for high fidelity multimedia content has soared, content distribution has emerged as a critical application. Large multimedia files require effective content distribution services such as content distribution networks (CDNs). A recent trend in CDN development is the use of peer-to-peer (P2P) techniques. P2P-based CDNs have several advantages over conventional non-P2P-based CDNs in scalability, fault resilience, and cost-effectiveness. Unfortunately, P2P-based content distribution poses a crucial problem in that update propagation is quite difficult to accomplish. This is because peers cannot obtain a global view of replica locations on the network. There are still several issues in conventional approaches to update propagation. They degrade the scalability, the fault resilience, and the cost-effectiveness of P2P-based content distribution, they also consume the network bandwidth, or take a long time. In this paper, we propose the speculative update, which quickly propagates an update to replicas with less bandwidth consumption in a pure P2P fashion. The speculative update enables a fast update propagation on structured P2P-based CDNs. Each server attempts to determine the directions in which there will be replicas with a high probability and speculatively relays update messages in those directions. Simulation results demonstrate that our mechanism quickly propagates an update to replicas with less bandwidth consumption. The The speculative update completes update propagation as fast as the simple gossip-based update propagation even with up to 69% fewer messages per second. Compared to the convergence-guaranteed random walk, the speculative update completes an update propagation faster by up to 92%.

View full abstract

We propose a type inference algorithm for a polymorphic type system which provides improved error messages. While the standard type inference algorithms often produce unnecessarily long or incomplete error messages, our algorithm provides relevant and complete information for type errors. It is relevant in the sense that all the program points and types in the output of our algorithm contribute to some type error, and is complete in the sense that, for each type error, our algorithm identifies not only two conflicting types, but also all types which conflict with each other. The latter property is particularly useful for debugging programs with lists or case branches. Our algorithm keeps track of the set of program points that are relevant to each type. To achieve completeness, we introduce a new type variable which represents a conflict among two or more incompatible types, and extend the unification algorithm to handle the special type variable appropriately. Finally, we argue that our algorithm is more efficient than those in the literature when there are more than two conflicting types in the given expression.

View full abstract

For fast ε-similarity search, various index structures have been proposed. Yi, et al. proposed a concept multi-modality support and suggested inequalities by which ε-similarity search by L₁, L₂ and L_∞ norm can be realized. We proposed an extended inequality which allows us to realize ε-similarity search by arbitrary L_p norm using an index based on L_q norm. In these investigations a search radius of a norm is converted into that of other norm. In this paper, we propose an index structure which allows search by arbitrary L_p norm, called mm-GNAT (multi-modality support GNAT), with the extention of ranges of GNAT, instead of extending the search radius. The index structure is based on GNAT (Geometric Near-neighbor Access Tree). We show that ε-similarity search by arbitrary L_p norm is realized on mm-GNAT. In addition, we performed search experiments on mm-GNAT with artificial data and music data. The results show that the search by arbitrary L_p norm is realized and the index structure has better search performance than Yi's method except for search by L₂ norm.

View full abstract

The purpose of this paper is to realize an authentication system which satisfies four requirements for security, privacy protection, and usability, that is, impersonation resistance against insiders, personalization, weak-unlinkability, and memory efficiency. The proposed system is the first system which satisfies all the properties. In the proposed system, transactions of a user within a single service can be linked (personalization), while transactions of a user among distinct services can not be linked (weak-unlinkability). The proposed system can be used with smart cards since the amount of memory required by the system does not depend on the number of services. First, this paper formalizes the property of weak-unlinkability, which has not been formalized in the literatures. Next, this paper extends an identification scheme with a pseudorandom function in order to realize an authentication system which satisfies all the requirements. This extension can be done with any identification scheme and any pseudorandom function. Finally, this paper proposes an implementation with the Schnorr identification scheme and a collision-free hash function as an example of the proposed system.

View full abstract

In this paper, we present a minimization algorithm of the number of states of a linear separation automaton (LSA). An LSA is an extended model of a finite automaton. It accepts a sequence of real vectors, and has a weight and a threshold sequence at every state, which determine the transition from the current state to the next at each step. In our previous paper, we characterized an LSA and the minimum state LSA. The minimum state version for a given LSA M is obtained by the algorithm presented in this paper. Its time complexity is O(( K + k )n²), where K is the maximum number of threshold values assigned to each weight, k is the maximum number of edges going out from a state of M, and n is the number of states in M. Moreover, we discuss the minimization of a threshold sequence at each state.

View full abstract

This paper presents mathematical and general models of electronic money systems. The goal of the paper is to propose a first framework in which various kinds of e-money systems can be uniformly represented and their security properties can be evaluated and compared. We introduce two kinds of e-money system models; a note-type e-money system model and a balance-type e-money system model. We show that any balance-type e-money system with efficient data transmission cannot be simulated by any note-type e-money system. This implies that balance-type e-money systems are strictly faster in data communication. Then, we show that a forged monetary value can be detected in some note-type e-money systems, while it cannot be detected in any balance-type e-money systems with efficient data communication. This implies that note-type e-money systems seem to be more secure.

View full abstract

Inverter power supplies have been developed as an analog circuit. Over the past decade, digital control has emerged as a viable candidate for the inverter power supplies due to the benefits which can be added by applying the digital control. However the complexity of digital control has been dramatically increasing. One of the leading methods in the digital control design is the model based development technique. Modeling and simulation act as the first step of model based development technique and they are very useful for the development and implementation in both hardware and software. They can lead to improve the performance, reduce development and production cost and time. This paper constructs a Model In the Loop Simulation (MILS) environment for designing the embedded software for the inverter power supply. The principle of the operation and the key features of using the digital PWM controllers are presented. Moreover, the important software parameters are optimized using the simulator. To verify the validity of the developed model, comparisons between the circuit simulation and experimental results are shown. The illustrations show that the results obtained by the developed model are in good agreement with the experimental results.

View full abstract

In this paper, the effectiveness of deterministic Multi-step Crossover Fusion (dMSXF) and deterministic Multi-step Mutation Fusion (dMSMF), which are types of genetic multistep searches based on a neighborhood search mechanism, in solving an unsupervised design problem of suitable structuring elements (SEs) of a morphological filter is shown. In our previous work, it was shown that dMSXF and dMSMF are very effective for solving combinatorial optimization problems, particularly on problems for which the landscape is an AR(1) landscape observed in the NK model. In addition, their effectiveness for reproduction mechanisms to obtain the offspring was shown to be retained with increasing level of epistasis. In this paper, we show that a characteristic of the AR(1) landscape is observed in an objective function for the unsupervised design of SEs, and superior search performances of both dMSXF and dMSMF for conventional crossover are shown. The processing results of the obtained SEs are also compared with those of conventional filters used for impulse noise removal.

View full abstract

Traditional XML query processing methods, such as XPath and XQuery, are fragile to changes in the underlying XML structure because path expressions cannot accommodate structural variations that may occur in designing or updating XML data. In this paper, we discuss the problem of processing path-free XML queries in a pure RDBMS. It is ideal to list all possible structural variations of a given path-free XML query, though it is non-trivial to devise an efficient implementation due to the combinatorial explosion of potential structural variations. In addition the problems that RDBMS cannot offer an ideal query plan and efficient XML structural join algorithms also pose a big challenge. We teach RDBMS aware of tree structures by adding XML-specific information, and supply FDs among attributes in an amoeba join to eliminate unfavorable results and achieve a marked reduction in the query space. Dealing with XML query in a pure RDBMS efficiently bridges the gap between XML and relational database. Experiments carried out on SQL Server have proven orders of magnitude improvement over the naïve implementation, demonstrating the feasibility of path-free XML query processing using a pure RDBMS kernel.

View full abstract