Cloud computing is a modern form of advanced information system that has developed with the proliferation of the Internet, broadband access networks and high-speed processors, and it is continuing to spread. Cloud computing enables users to use IT resources, such as processors and storage, through the network, by simply paying a fee and without needing to own servers. Running costs are reduced because resources can be used on-demand only as needed. Since hardware provisioning is not necessary, software development and launching new services can be done quickly. Cloud computing also drives innovation in information systems. For example, cloud computing has made it easy to build cluster systems using virtual machines, which has led to the development of scalable data stores such as object storage and key-value stores. Cloud computing also led to the development of software defined networks and software defined storage, to respond rapidly to the requirements of users and applications. In this paper, the benefits of cloud computing are reviewed and technologies supporting it and new technologies arising from it are outlined. Directions for cloud computing in the future are also discussed.

View full abstract

Recently, Internet of Things (IoT) has been attracting attention due to its economical impact and high expectations for drastically changing our modern societies. Worldwide by 2022, over 50 billion IoT devices including sensors and actuators are predicted to be installed in machines, humans, vehicles, buildings, and environments. Demand is also huge for the real-time utilization of IoT data streams instead of the current off-line analysis/utilization of stored big data. The real-time utilization of massive IoT data streams suggests a paradigm shift to new horizontal and distributed architecture because existing cloud-based centralized architecture will cause large delays for providing service and waste many resources on the cloud and on networks. Content curation, which is the intelligent compilation of valuable content from IoT data streams, is another key to fully utilize and penetrate IoT technologies. In this paper, we survey the emerging technologies toward the real-time utilization of IoT data streams in terms of networking, processing, and content curation and clarify the open issues. Then we propose a new framework for IoT data streams called the Information Flow of Things (IFoT) that processes, analyzes, and curates massive IoT streams in real-time based on distributed processing among IoT devices.

View full abstract

This paper presents a new smartphone game for collecting environmental sounds using a participatory sensing approach. In our game, a player becomes an owner of a virtual house and furnishes the house with in-game items that are obtained by collecting environmental sounds. The collected sounds are used to train an environmental sound recognition model for daily activity recognition. However, such participatory sensing systems have several issues related to, for example, motivation of users and reliability of collected data. To cope with the issues, we propose and implement gamified functions for controlling the quality and diversity of collected data. We conducted an experiment to evaluate the developed functions and confirmed the effectiveness of our developed functions.

View full abstract

The rapid growth of communication networks, such as the Internet and mobile networks, has spurred the development of numerous online trading systems. It is important for such a community that there be trust between the trading partners. A system that stores information about the reputation for trustworthiness of the individuals in the network is known as a reputation system. However, a malicious evaluator can provide an incorrect evaluation, and several such users may collude to create a false evaluation of a particular target; this is known as a pinpoint collusive attack and is a major problem. In order to resist such an attack, reputation systems have been proposed that consider the ability of a given evaluator to provide an accurate evaluation of a given target. However, these systems are unable to identify attacks from evaluators who have achieved a sufficiently high rating for their ability. In this paper, we propose a novel algorithm that, when calculating the ability of an evaluator, reduces the influence of values that are not in line with the daily trend. Simulation results show that the proposed method reduces the influence of an attack.

View full abstract

Brute force attacks are used to obtain pairs of user names and passwords illegally by using all existing pairs to login to network services. These are a major security threat faced by network service administrators. In general, to prevent brute force attacks, administrators can set limitations on the number of login trials and shut down the traffic of brute force attacks with an intrusion prevention system (IPS) at the entry point to their services. In recent years, stealthy brute force attacks that can avoid the security rules and IPS and intrusion detection system (IDS) detection have appeared. Attackers tend to arrange a large amount of hosts and allocate them fewer login trials than the limitations administrators set. In this paper, we report a kind of distributed brute force attack event (brute force attacks with disciplined IPs, or DBF) against the Remote Desktop Protocol (RDP) by analyzing IDS logs integrated from multiple sites. In DBF, a particular number of attacks is repeated automatically from a host to a service over a period. For this reason, existing countermeasures have no effect on DBF. We investigate the structure of DBF and improve the existing countermeasure system. We also present TOPASE, which is replaced at each step of the existing countermeasure system and is suitable for DBF countermeasures. TOPASE analyzes the regularity of login trials between a source host and a destination host. Furthermore, TOPASE intercepts the network traffic from the source host of the brute force attack for a specific period. As a result of the evaluation with our IDS log, we estimate the performance of TOPASE and clarify the factors that maximize TOPASE's effectiveness.

View full abstract

Users of smartphones and/or tablet terminals browse and download confidential document files routinely. Therefore, a higher security level is needed for smartphones and tablet terminals than conventional mobile phones. From this kind of background, Takahashi and Uchida proposed an image-based user authentication method for touch screen devices by using the latest image shot by the user as the pass-image. The proposed authentication method is resistant to smudge attacks, one of the most serious threats for touch screen devices. However, the security strength of the method is low. Therefore, in this paper, we propose SWIPASS, an image-based user authentication method for touch screen devices that has higher security strength, by improving on the method proposed by Takahashi and Uchida. This improves the security strength without any change in either the resistance to smudge attacks or the users' burden of memorizing. Although Takahashi and Uchida implemented their method only as a prototype system in their study, we implement SWIPASS as a real Android application in this study. Moreover, we also examined the usability and the resistance of SWIPASS against observation attacks by conducting several experiments in this paper.

View full abstract

This paper describes our experimental investigation of the end user physical characteristics (e.g., gender, height, weight, dominant hand, and skill at sport) that can be successfully estimated solely from sensor data obtained during daily activities (e.g., walking and dish washing) from body-worn accelerometers. For this purpose we use the huge quantities of data that we have collected, which include 14, 880 labeled activities obtained from 61 subjects. Our proposed method tries to estimate various kinds of characteristics based on our simple idea ‘When the activity sensor data of two users are similar, the physical characteristics of the two users may also be similar.’ We consider that estimating the end user's physical characteristics will enable us to realize new kinds of applications that automatically recommend information/services to an end user according to her estimated physical characteristics such as gender and weight.

View full abstract

The bandwidth of a mobile network is limited and exhausted very fast with the huge number of mobile devices and applications. In order to manage and utilize the limited bandwidth, precise mobile application identification is required. In this work, the combination of communication patterns extracted from graphlet and traffic patterns represented by packet size distribution is studied for enhancing the performance of identifying mobile traffic. There are no privacy concerns for identifying traffic with our technique; it is also effective against the complexities of mobile traffic. The real traffic of five famous mobile applications (Facebook, Line, Skype, YouTube, and Web) is used in our evaluation. The identification performance is high (0.96) of F-measure even considering only a random 50 packets of traffic in a 3-minute duration. While identifying applications, the effect of other mixed background traffic is also studied and mitigated by filtering out short lived flows with a flow duration condition. The high identification performance is still maintained after this filtering process.

View full abstract

Parsing Expression Grammars are a popular foundation for describing syntax. Unfortunately, several syntax of programming languages are still hard to recognize with pure PEGs. Notorious cases appears: typedef-defined names in C/C++, indentation-based code layout in Python, and HERE document in many scripting languages. To recognize such PEG-hard syntax, we have addressed a declarative extension to PEGs. The “declarative” extension means no programmed semantic actions, which are traditionally used to realize the extended parsing behavior. Nez is our extended PEG language, including symbol tables and conditional parsing. This paper demonstrates that the use of Nez Extensions can realize many practical programming languages, such as C, C#, Ruby, and Python, which involve PEG-hard syntax.

View full abstract

In order to realize highly-reliable volunteer computing (VC), this paper develops a VC system with credibility-based voting. Credibility-based voting is known as an efficient technique for eliminating incorrect calculation results returned by participant (worker) nodes. Although its theoretical performance has been studied in detail, its implementation and practical performance have not been fully studied yet. Our VC system consists of a management server and a number of worker nodes. In the management server, each process that includes credibility-based voting is multithreaded, and all information is managed in a database for high performance and stable operation. Through performance evaluations, we found that the most time-consuming processes are the credibility-based voting in a network environment with a small delay (e.g., Intranet), and the process for sending jobs in a network environment with a moderate delay (e.g., domestic Internet). Moreover, multithreading is shown to be effective for those performance bottlenecks, and feasible system configurations are also revealed for a variety of request rates.

View full abstract

Fail-stop signatures (FSS) provide the security for a signer against a computationally unbounded adversary by enabling the signer to provide a proof of forgery. Conventional FSS schemes are for a single-signer setting, but in the real world, there is a case where a countersignature of multiple signers (e.g., a signature between a bank, a user, and a consumer) is required. In this work, we propose a framework of FSS capturing a multi-signer setting and call the primitive fail-stop multisignatures (FSMS). We propose a generic construction of FSMS via the bundling homomorphisms proposed by Pfitzmann and then propose a provably secure instantiation of the FSMS scheme from the factoring assumption. Our proposed schemes can be also extended to fail-stop aggregate signatures (FSAS).

View full abstract

Nowadays, wireless LAN is equipped in most portable devices and widely deployed, and then we can communicate over wireless link with low cost. However, when the same content is simultaneously downloaded by several terminals with unicast in a wireless LAN, the bandwidth must be divided among the terminals and the download speed is reduced. With using the broadcast nature of wireless media, pseudo multicasting methods, which overhear the unicast flow to realize multicast like communication, have been proposed. In the pseudo multicasting methods, since parts of the content is obtained by overhearing the TCP/UDP flows of a specific content going to other terminals in a wireless LAN, and the server need not send the content for each client repeatedly. These methods enable to realize multicasting with a practical way. However, redundant unicast traffic is still generated because the receivers do not obtain contents via overhearing in an efficient way. Based on the pseudo multicasting, we propose an efficient, flexible and reliable content distribution method over single hop wireless ad hoc network in this paper. In the proposed method a content is divided into pieces and the server distributes it according to the download status of each receivers at the application level. To accelerate downloading, we propose a method for selecting the piece and the terminal for transmission taking the effect of network coding for efficient overhearing. We developed a testbed and confirmed that the proposed scheduling accelerates the download speed even when the communication speed of each terminal is different.

View full abstract

This paper presents a method for evaluating toothbrushing performance using audio data collected by a smartphone. This method first conducts activity recognition on the audio data to classify segments of the data into several classes based on the brushing location and type of brush stroke. These recognition results are then used to compute several independent variables which are used as input to an SVM regression model, with the dependent variables for the SVM model derived from evaluation scores assigned to each session of toothbrushing by a dentist who specializes in dental care instruction. Using this combination of audio-based activity recognition and SVM regression, our method is able to take smartphone audio data as input and output evaluation score estimates that closely correspond to the evaluation scores assigned by the dentist participating in our research.

View full abstract

In this paper, we introduce a new extension for bounded-SVD, i.e., a matrix factorization (MF) method with bound constraints for recommender system. In bounded-SVD, the bound constraints are included in the objective function so that not only the estimation errors are minimized but the constraints are also taken into account during the optimization process. Our previous results on major real-world recommender system datasets showed that bounded-SVD outperformed an existing MF method with bound constraints, BMF, and it is also faster and simpler to implement than BMF. However, an issue of bounded-SVD is that it does not take into account the bias effects in given data. In order to overcome this issue, we propose an extension of bounded-SVD: bounded-SVD bias. Bounded-SVD bias takes into account the rating biases of users and items - known to reside in recommender system data. The experiment results show that the bias extension can improve the performance of bounded-SVD in most cases.

View full abstract

Random walks-based (RW) segmentation methods have been proven to have a potential application in segmenting the medical image with minimal interactive guidance. However, the approach leads to large-scale graphs due to number of nodes equal to voxel number. Also, segmentation is inaccurate because of the unavailability of appropriate initial seed points. It is a challenge to use the RW-based segmentation algorithm to segment organ regions from 3D medical images interactively. In this paper, a knowledge-based segmentation framework for multiple organs is proposed based on random walks. This method employs the previous segmented slice as prior knowledge (the shape and intensity constraints) for automatic segmentation of other slices, which can reduce the graph scale and significantly speed up the optimization procedure of the graph. To assess the efficiency of our proposed method, experiments were performed on liver tissues, spleen tissues and hepatic cancer and it was extensively evaluated both quantitatively and qualitatively. Comparing our method with conventional RW and state-of-the-art interactive segmentation methods, our results show an improvement in the accuracy for multi-organ segmentation (p<0.001).

View full abstract

Text in video compressed by lossy compression at a low bitrate will easily be deteriorated, resulting in blurred text and a lower readability. In this paper, we propose a novel image coding method to preserve the readability of text in the video at a very low bitrate. During the encoding process, we estimate the parameters for each character of the text. Then, an image without text is generated and compressed. During the decoding process, we reconstruct video sequences with text from images without text and character images generated by the estimated parameters. The experimental results show the effectiveness of the proposed method in terms of the readability at a very low bitrate.

View full abstract

Open Source Software (OSS) is vital to both end users and enterprises. As OSS systems are becoming a type of infrastructure, long-term OSS projects are desired. For the survival of OSS projects, the projects need to not only retain existing developers, but also attract new developers to grow. To better understand how projects retain and attract contributors, our preliminary study aimed to measure the personnel attraction and retention of OSS projects using a pair of population migration metrics, called Magnet (personnel attraction) and Sticky (retention) metrics. Because the preliminary study analyzed only 90 projects and the 90 projects are not representative of GitHub, this paper extend the preliminary study to better understand the generalizability of the results by analyzing 16, 552 projects of GitHub. Furthermore, we also add a pilot study to investigate the typical duration between releases to find more appropriate release duration. The study results show that (1) approximately 23% of developers remain in the same projects that the developers contribute to, (2) the larger projects are likely to attract and retain more developers, (3) 53% of terminal projects eventually decay to a state of fewer than ten developers and (4) 55% of attractive projects remain in an attractive category.

View full abstract

One of the major causes of traffic accidents according to the statistical report on traffic accidents in Japan is the disregard of traffic lights by drivers. It would be useful if driving support systems could detect and recognize traffic lights and give appropriate information to drivers. Although many studies on intelligent transportation systems have been conducted, the detection of traffic lights using images remains a difficult problem. This is because traffic lights are very small as compared to other objects and there are many objects similar to traffic lights in the road environment. In addition, the pixel colors of traffic lights are easily over-saturated, which renders traffic light detection using color information difficult. The rapid deployment of the new LED traffic lights has led to a new problem. Since LED lights blink at high frequency, if they are captured by a digital video camera, there are frames in which all the traffic lights appear to be turned off. It is impossible to detect traffic lights in these frames by searching the ordinary color of traffic lights. In this paper, we focus on the stable detection of traffic lights, even when they are blinking or when their colors are over-saturated. A method for detecting candidate traffic lights utilizing intensity information together with color information is proposed for handling over-saturated pixels. To exclude candidates that are not traffic lights efficiently, the sizes of the detected candidates are calculated using a stereo image. In addition, we introduce tracking with a Kalman filter to avoid incorrect detection and achieve stable detection of blinking lights. The experimental results using video sequences taken by an in-vehicle stereo camera verify the efficacy of the proposed approaches.

View full abstract

Currently, lecturers have been offered various LMS's. The Learning Management System (LMS) could provide high quality, rich and complete materials in Higher Education Institution (HEI) that use LMS as part of academic activities. In those activities, lecturers are the key persons responsible for enriching learning materials. Therefore, it was a common sense if lecturers need to be supported for enriching learning materials in any condition, whether it is online or offline. There are several options for a lecturer when he needs to use a LMS: use the LMS directly (access to the LMS server) or by implemented on a local machine as a local server. The use of LMS as a server application to give services for a virtual classroom brings many issues like installation, configuration, learning materials authoring, LMS operation, etc. In this particular case, we focus on the Indonesian educational environment, where a highly motivated lecturer is eager to use LMS to deliver lectures but restricted to use a limited bandwidth. Therefore, tools that can be used online as well as offline and are able to share the contents over diverse LMS are needed. The present paper discusses a supportive tool to support offline authoring, and delivery methodology development in existing LMS or even HEI with no LMS installed, considering all the issues related to LMS and the limited bandwidth environment where it will be used. As a result, we developed a Lecture Based Supportive Tool (LBST) as an approach to enable a lecturer to create learning materials in offline conditions and/or limited bandwidth, and then upload it on a remote LMS as an activity to share and enrich learning materials.

View full abstract

Cloud computing provides many advantages for both the cloud service provider and the clients. It is also infamous for being highly dynamic and for having numerous security issues. The dynamicity of cloud computing implies that dynamic security mechanisms are being employed to enforce its security, especially in regards to access decisions. However, this is surprisingly not the case. Static traditional authorization mechanisms are being used in cloud environments, leading to legitimate doubts on their ability to fulfill the security needs of the cloud. We propose a risk adaptive authorization mechanism (RAdAM) for a simple cloud deployment, collaboration in cloud computing and federation in cloud computing. We use a fuzzy inference system to demonstrate the practicability of RAdAM. We complement RAdAM with a Vulnerability Based Authorization Mechanism (VBAM) which is a real-time authorization model based on the average vulnerability scores of the objects present in the cloud. We demonstrated the usefulness of VBAM in a use case featuring OpenStack.

View full abstract

Risk-aware data replication (RDR), which replicates data at primary sites to safe backup sites, has been proposed to mitigate a service disruption in a disaster area even after a widespread disaster that damages a network and a primary site. RDR assigns a safe backup site to a primary site while considering a damage risk for both the primary site and the backup candidate site. When the backup candidate sites are widely distributed in an urban and suburban area, RDR sometimes assigns a backup site too far from the primary site. However the backup site is desired to be reachable from the primary site by physical transfer such as walking, bicycle, car, or drone in case that a severe disaster damages network among the sites. Therefore, limiting the distance between the primary site and the backup site is required. To approach this challenge, we propose two possible methods: the average distance limiting (ADL) method and the maximum distance limiting (MDL) method. In this paper, we compare the distance distributions, the data availability, and the computation time of two methods. Then, we conclude that the MDL method is the most practicable from a comprehensive perspective.

View full abstract

Ford and Fulkerson's labeling method is a classic algorithm for maximum network flows. The labeling method always terminates for networks whose edge capacities are integral (or, equivalently, rational). On the other hand, it might fail to terminate if networks have an edge with an irrational capacity. Ford and Fulkerson also gave an example of such networks on which the labeling method might fail to terminate. However, their example has 10 vertices and 48 edges and the flow augmentation is a little bit complicated. Simpler examples have been published in the past. In 1995, Zwick gave two networks with 6 vertices and 9 edges and one network with 6 vertices and 8 edges. The latter is the smallest, however, the calculation of the irrational capacity requires some effort. Thus, he called the former the simplest. In this paper, we show the simplest and smallest network in Zwick's context. Moreover, the irrational edge capacity of our example can be arbitrarily assigned while those in the all previous examples are not. This suggests that many real-valued networks might fail to terminate.

View full abstract

We propose a multicast tree management method in an OpenFlow controller that handles both fast failure recovery and dynamic multicast group membership changes. Multicast communication is an efficient tool to distribute data to many hosts in various services such as live video streaming. To use multicast in such services, multicast communication must be reliable, which means multicast communication should be restored quickly after failures, and multicast tree management mechanism should support frequent group membership changes. A conventional approach, Point to Multipoint (P2MP) MPLS, only supports fast failure recovery for reliability, and is not very effective in terms of group membership changes. A new approach using OpenFlow supports dynamic group membership changes, but does not consider fast failure recovery in physical switches whose flow entry modification is slow. Our proposed method is to control multicast trees centrally, and it uses a precomputation and pre-installation approach for tree management. A controller calculates and keeps multiple trees that cover all switches where receivers are potentially connected and that have less common nodes and edges, and installs their sub-trees covering switches where receivers are actually connected. The controller calculates the difference per tree between sub-trees before and after membership changes, and reflects them into the network. At the time of failure, the controller checks and finds a pre-installed tree that is unaffected by the failure, and installs a new rule only to a root switch to send packets through the pre-installed alternate tree. Our experiments using switches and our prototype controller show that our proposed method can restore packet delivery quickly after a failure, as well as that our proposed method can handle tree modifications faster than a method of recalculating or reinstalling a tree every time that group memberships are changed.

View full abstract

Research on multimodal data analysis such as annotated image analysis is becoming more important than ever due to the increase in the amount of data. One of the approaches to this problem is multimodal topic models as an extension of Latent Dirichlet allocation (LDA). Symmetric correspondence topic models (SymCorrLDA) are state-of-the-art multimodal topic models that can appropriately model multimodal data considering inter-modal dependencies. Incidentally, hierarchically structured categories can help users find relevant data from a large amount of data collection. Hierarchical topic models such as Hierarchical latent Dirichlet allocation (hLDA) can discover a tree-structured hierarchy of latent topics from a given unimodal data collection; however, no hierarchical topic models can appropriately handle multimodal data considering inter-modal mutual dependencies. In this paper, we propose h-SymCorrLDA to discover latent topic hierarchies from multimodal data by combining the ideas of the two previously mentioned models: multimodal topic models and hierarchical topic models. We demonstrate the effectiveness of our model compared with several baseline models through experiments with three datasets of annotated images.

View full abstract

Load balancing is a major concern in massively parallel computing. X10 is a partitioned global address space language for scale-out computing and provides a global load balancing (GLB) library that shows high scalability over ten thousand CPU cores. This study proposes a multistage mechanism for GLB to assign execution stages to tasks and introduces a multithread design into GLB to allow efficient data sharing between CPU cores. The system gives high priority to tasks that are assigned to earlier stages and then proceeds with subsequent stage tasks. When a computing node runs out of tasks at the earliest stage, it requests tasks at the earliest stage from other nodes and awaits responses by processing subsequent stage tasks. When the system identifies the task termination at a certain stage, it executes a reduction operation over nodes. Programmers can define their reduction operations to gather or exchange results of completed tasks. This study provides the implementation method of the extended library and evaluates its runtime overhead using the K computer to a maximum of 256 nodes.

View full abstract

XML is a popular data definition language and is widely used for representation of arbitrary data structures. For queries on XML documents, XPath has commonly been used in many applications. The complexity of applying queries increases as the number of nodes in an XML document increases. Querying very large XML documents becomes really difficult when there is not enough computer memory to store and manipulate the whole tree data. The objective of this study is to develop an algorithm for querying very large XML trees in a distributed-memory environment. We split a large XML document into small chunks and parse the chunks to create special trees called partial trees. Then the query is executed in parallel on the partial trees. The results from the partial trees are concatenated to form the final query results for output. The algorithms were tested on a 16-node PC cluster, and the experiment results showed a speedup of a factor of 6 on 16 nodes.

View full abstract

In this paper, an implementation of exception handling for task-parallel languages is proposed such that all running parallel tasks in a try block with an exception are automatically aborted as soon as possible. In parallel tree search, exception handling that allows such a collateral task abortion is useful when the objective is to complete the search as soon as one solution is found or to allow a worker to abort the traversal of a subtree that is found to be redundant by another worker, even when it has been initiated. However, few existing task-parallel languages, such as Cilk Plus and X10, have this capability. In this study, we enhanced a task-parallel language, Tascell, with this capability. Since the Tascell compiler is implemented as a translator to C code, techniques are required for implementing the non-local exit mechanism with cleanup code execution in the “finally” clauses. We achieved this implementation by exploiting nested functions, which are already used in the temporary backtracking mechanism of Tascell. We also modified the task scheduler provided by Tascell such that a worker can abort a task after it is started. When aborting a task, the scheduler also aborts all its descendant tasks. We evaluated our implementation in terms of overheads and time taken to abort tasks.

View full abstract