With the increasing demand for high-performance computing, multicore architectures became appealing in various application domains. In order to exploit the parallelism of the multicore architectures, task scheduling has become more important than ever. Classical multicore task scheduling assumes that each task is executed on one of the cores. However, many tasks in modern applications have inherent parallelism and can be multi-threaded. A task is partitioned into threads which can be executed on multiple cores in a fork-join fashion. A multi-threaded task is called malleable if the number of threads is flexible and is determined at the same time as task scheduling. This paper proposes multicore scheduling methods for malleable tasks. Given a set of dependent tasks in the form of directed acyclic graph and homogeneous multiple cores, the proposed methods decide the number of threads for each task and schedule the threads on the multicores simultaneously, with the goal of minimizing the overall schedule length. The proposed scheduling methods are based on constraint programming. Experimental results show that the proposed methods outperform state-of-the-art work which is based on integer linear programming.

抄録全体を表示

This paper considers the global (l,k)-critical section problem which is the problem of controlling a system in such a way that at least l and at most k processes must be in the critical section at any time in the network, while each process alternates between in the critical section and out of the critical section. In this paper, first, a distributed solution for l-mutual inclusion is proposed in the asynchronous message-passing model. The proposed algorithm uses an ordinary quorum system and all processes play the same role, unlike existing algorithms for k-mutual exclusion. After that, using the proposed algorithm for l-mutual inclusion, we propose a distributed solution for the global (l, k)-critical section problem. The proposed approach is a versatile composition of algorithms for $l$-mutual inclusion and k-mutual exclusion. Its message complexity is typically O(√n), where n is the size of the network.

抄録全体を表示

In the high-associativity caches, the hardware overheads of the cache replacement policy become a problem. To avoid this problem, the Adaptive Demotion Policy (ADP) is proposed. The ADP focuses on the priority value demotion at a cache miss, and it can achieve a higher performance compared with conventional cache replacement policies. The ADP can be implemented with small hardware resources, and the priority value update logic can be implemented with a small hardware cost. The ADP can suit for various applications by the appropriate selection of its insertion, promotion and selection policies. If the dynamic selection of the suitable policies for the running application is possible, the performance of the cache replacement policy will be increased. In order to achieve the dynamic selection of the suitable policies, this paper focuses on the global fluctuations of the priority values. At first, the cache is partitioned into several partitions. At every cache access, the total of priority values in each partition is calculated. At every set interval, the fluctuations of total priority values in all the partitions are checked, and the information is used to detect the behavior of the application. This paper adapts this mechanism to the ADP, and the adapted cache replacement policy is called the ADP-G. The performance evaluation shows that the ADP-G achieves the MPKI reductions and the IPC improvements, compared to the LRU policy, the RRIP policy and the ADP.

抄録全体を表示

IoT technologies and cloud storages have been developed remarkably, and many types of data including private information have been gradually outsourced. For such a situation, new convenient functionalities such as arithmetic and a keyword search over ciphertexts are required to allow users to obtain information without leaking queried results, contents of ciphertext, and so on. Especially, searchable encryptions have been paid much attention to realize a keyword search on an encrypted domain. In addition, an architecture of searchable symmetric encryption (SSE) is a suitable and efficient solution for data outsourcing. In this paper, we focus on an SSE scheme which employs a secure index for searching a keyword. In conventional studies, it has been widely considered that the scheme searches whether a queried keyword is contained in encrypted documents or not. On the other hand, we additionally take into account the location of the queried keyword in the documents by targeting a matrix-type data format. The method enables an administrator to search personal information listed per line or column in CSV-like format data.

抄録全体を表示

Moving target defense (MTD) is a feasible idea for reducing the ratio of successful attacks by altering or diversifying the attributes or parameters of a protected system. As a result of applying MTD techniques to a system, an attacker would have more difficulties in launching attacks. Although several MTD techniques have been proposed for different types of attack, estimating the effectiveness of combining these MTDs remains a challenge. With the aim of setting up a method for evaluating MTDs, we first propose a model composed of two MTD diversification techniques to compare an attack success ratio between theoretical and experimental probability. To validate the proposed model, we conducted an experiment involving an actual attack and then analyzed how our MTD model can adequately estimate a binary-code injection attack. Results show that the rate of attack success is 100% when MTD diversification is not implemented, while the rate is reduced depending on how many variants can be diversified in a target system. Our method is an important first step toward establishing a method for evaluating MTDs, as well as predicting an MTD’s defensive abilities.

抄録全体を表示

The problem of efficiently identifying critical nodes that substantially degrade network performance if they do not function is crucial and essential in analyzing a large complex network such as social networks on the Web and road network in the real world, and it is still challenging. In this paper, we tackle this problem under a realistic situation where each link is probabilistically disconnected as assumed in studies in uncertain graphs. This reflects that in case of a social network an information path between two persons is not always open and may not pass on any information from one to the other and in case of a road network a road between two intersections is not always travelable and may be blocked by a traffic accident, a road repair, a nearby construction, etc. To solve this problem, we focus on the articulation point and utilize the bridge detection technique in graph theory to efficiently identify critical nodes when the node reachability is taken as the performance measure. In case of a social network disfunction of a node causes loss of the total number of people receiving information and in case of a road network it causes loss of the total number of people movable to other places. Using two real-world social networks and one road network, we empirically show that the proposed method has a good scalability with respect to the network size and the nodes our method identified possesses unique properties and they are difficult to be identified by using conventional centrality measures.

抄録全体を表示

Essential services, such as security software or logging software, are considered important because of an increase in attacks on computers. These essential services are provided by processes that sometimes involve file manipulation and communication. Moreover, these essential services can be a target of attacks and become disabled, as they can be an obstacle to attackers. Attackers can speculate essential services by monitoring the behavior of the processes. To avoid such attacks on essential services, methods for hiding their behavior are proposed. The methods use a virtual machine (VM) monitor to make it difficult for attackers to identify essential services by hiding process information and file manipulation. However, the communication information remains visible to attackers. To address this problem, this study proposes a method for hiding the communication of essential services by using a system call proxy. We assume that a process providing essential services (essential process) runs on a protection target VM and a proxy process runs on a proxy VM. In the proposed method, the system calls in the communication invoked by the essential process are executed by the proxy process. The system calls invoked by the proxy process are not executed on the protection target VM; therefore, attackers cannot identify the communication of essential services by monitoring their communication. This paper presents the design, implementation, and evaluation of the proposed method.

抄録全体を表示

Wireless energy transfer is a technology to transmit electricity without wire, and it is a promising technology for charging battery of mobile devices. In battery powered sensor networks, it is important to balance electric energy of batteries of nodes in order to maximize the life time of networks. In this paper, we propose three distributed protocols to balance electric energy of batteries of nodes. The proposed algorithms are based on the population protocol model which is a computational model for networked nodes with very limited resources. The goals of the algorithms are twofold: minimizing the loss of electric energy caused by wireless transmission, and minimizing the time to balance. The proposed algorithms are evaluated by computer simulation.

抄録全体を表示

A cellular automaton (CA) is a well-studied non-linear computational model of complex systems in which an infinite one-dimensional array of finite state machines (cells) updates itself in a synchronous manner according to a uniform local rule. A sequence generation problem on the CA model has been studied for a long time and a lot of generation algorithms has been proposed for a variety of non-regular sequences such as {2ⁿ | n = 1, 2, 3,...}, prime, and Fibonacci sequences etc. In this paper, we study a real-time sequence generator for {n³ | n=1, 2, 3, ...}. In the previous studies, Kamikawa and Umeo(2018) showed that sequence {n³ | n=1, 2, 3, ...} can be generated in real-time by an eight-state CA. We present a new six-state implementation of real-time sequence generator for {n³ | n=1, 2, 3, ...} rather than reducing the internal state of the Kamikawa and Umeo's sequence generator and give a formal proof of the correctness of the generator. In addition, we show the number of state-changes and number of cells of sequence generators, and compare sequence generators.

抄録全体を表示

Agriculture in Japan is facing a crisis due to the aging of farmers and a shortage of successors. In addition, much of agriculture depends on know-how that new farmers cannot easily obtain; therefore, new farmers often give up farming after a few years. Consequently, it is necessary to share agricultural know-how with new farmers. However, the difficulty in verbalizing this know-how poses a challenge in sharing it. Therefore, it is necessary to provide incentives to the providers of know-how. In this paper, we propose an incentive decision method using auto-negotiation while considering the intentions of the know-how providers. This method consists of agents that have utility functions, which express the intentions of farmers, and a service. The performance of our method is evaluated by simulation experiments. We verify that our method collects a large amount of high-quality know-how with low incentives. The results indicate that our proposed method performs better than the existing methods considered in the experiments.

抄録全体を表示

To reduce power consumption, approximate computing is an efficient approach for error-tolerant applications such as image processing. Approximate arithmetic adders can be used for the approximate computing, and can trade off accuracy for power. CMA, a dynamically accuracy-configurable approximate adder, had been proposed. CMA can sharply reduce power consumption compared with other accuracy-configurable approximate adders, while allowing it to change accuracy-setting at run-time. In this paper, we evaluate CMA with error corrector that needs only two gates for each digit in actual image processing circuit. By increasing slight extra power, the proposed value corrector can improve PSNR quality of output images by up to 73.71%.

抄録全体を表示

Currently, devices with wireless technologies often communicate each other ad hoc. For example, a presenter wirelessly distributes ad-hoc meeting materials from a PC to mobile device in a meeting room. However, there is a problem of spoofing by an impersonator outside the room. Hence, devices must conduct secure pairing, which is exchange of key necessary for encrypting communication contents, before the communication. As a pairing method between devices, there are pairing methods using RSS from access point as features. However, RSS changes significantly due to environmental factors. On the other hand, there are the pairing methods which compare acceleration data from devices with displacement data of devices from camera of a server. However, these methods have problems that it is necessary to use infrared camera and difficult to recognize inclination of devices. Thus, these methods cannot perform accurate device pairing. Therefore, in this paper, we propose a method that perform pairing using devices' accelerometers and markers displayed on devices, and a camera of authentication server. This method performs pairing by calculating similarity between velocity data from acceleration data from devices and displacement data of the marker from camera after comparing marker sequence displayed on the device. This method has advantage that can detect devices' inclination by recognizing markers' inclination. We performed three types of experiments to confirm the similarity of displacement data and acceleration data, whether an impersonator outside camera range can perform pairing, and possibility of several devices pairing together. As a result, we founded that the larger the device's display is, the higher the similarity, the proposed method can distinguish legitimate user from impersonator outside camera range by average similarity, and three devices succeeded pairing at rate of 71.8%.

抄録全体を表示

A transactional system consists of a concurrency control system and a recovery system. TicToc is one of the state of the art concurrency control protocols today, but it lacks recovery system. We studied the ways to integrate TicToc and recovery system. For efficiency, we adopted a parallel write ahead logging scheme for the recovery system. There are two methods to optimize the logging. First method is early lock release which executes lock release early on data objects. Second method is group commit which executes batched logs transfer to storage from memory. We integrated a transactional system consisted by TicToc and P-WAL logging system assuming non-volatile memory. We found that the two optimization methods incur performance degradation when storage access latency is equivalent to that of NVRAM.

抄録全体を表示

In this study, we examine the relation between the spatio-temporal patterns generated by two-dimensional symmetrical elementary cellular automata and a singular function. In a previous paper, we proved that a specific cellular automaton admits a "limit set" (a limit on the series of spatio-temporal patterns contracted with time), and we calculated the fractal dimension of the boundary of this limit set. In this paper, we provide an overview of the previous results and a more precise analysis. Numerical simulations demonstrate that the essential fractal-like patterns created by two-dimensional cellular automata are also related to a singular function.

抄録全体を表示

Nowadays, a hash function is used for password management. The hash function is desired to possess the following three characteristics: Pre-Image Resistance, Second Pre-Image Resistance, and Collision Resistance. They are set on the assumption that it is computationally difficult to find the original message from a given hash value. However, the security level of the password management will be further reduced by implementing a high speed hash function on GPU. In this paper, the implementation of high speed hash function Keccak-512 using the integrated development environment CUDA for GPU is proposed. The following four techniques are used in order to speed up its implementation. The first one is reforming lookup tables from 2 dimensional arrays to 1 dimensional arrays at step ρ and π. The second is an investigation into the effect of using constant memory and shared memory for constant values. The third is the finding out the optimal configuration of blocks-threads, then evaluate the implementation according to the occupancy. And the last one is using CUDA streams with overlapping to hide the overhead of data transfer and GPU processing.

抄録全体を表示