Abstract:The revolutionary progress of data collecting techniques, dramatic decrease of the price of storage devices, as well as the desirability of people to extract information from the data have given birth to the so-called big data and data management technologies usher in the age of big data. RDBMS (relational database management system) undergoes a development of 40 years since the 1970s and now encounters some difficulties such as limited system scalability and limited data variety support. In recent years, noSQL technologies has risen suddenly as a new force. The technologies can manage, process, and analyze various types of data, achieve rather high performance with the help of parallel computing, can handle even bigger volume of data with the nice property of highly scalability. The paper follows the path of database technology progress and unfolds the new landscape of data management technologies from the angle of applications (operational as well as analytic applications). The paper also identifies some chanllenging and important issues that deserve further investigation, with the authors' recent research work introduced at the end.

Abstract:Data gathering in wireless sensor networks by employing mobile data collectors (MDCs) can greatly reduce the relay hops when the sensors transmit data to static base station. This prolongs the lifetime of whole network for the energy saving at sensors. To avoid the energy holes incurred by multi-hop relaying among sensors, MDCs gather data from the sensors directly or some nodes buffered data of other sensors. Furthermore, MDCs relay data from sensors to the base station when there are no complete links between them due the fact that some sensors were invalid. However, new challenges have arose for the mobility of MDCs in data gathering, which is different from static wireless sensor networks. This paper focuses on the mobility-assisted data of gathering strategies in wireless sensor networks. Some current novel theories and algorithms for data gathering based on MDCs are reviewed, and the taxonomy is described. More specifically, several typical algorithms and protocols are discussed in detail. In the end, advantages and disadvantages of the algorithms are summarized. The open research issues in this field are also pointed out.

Abstract:Low-Duty-Cycle (LDC) is the most important technique in increasing the lifecycle of wireless sensor networks. This paper imports low duty cycle technology into opportunistic networks to extend the lifecycle of opportunistic networks effectively. However, the existing routing algorithms for opportunistic networks perform poorly in LDC environment. To solve this problem, this paper proposes an erasure-coding algorithm based on energy-aware (E-EC) for LDC opportunistic networks. Simulation results show that, compared with existing classical algorithms, E-EC algorithm achieves significant performance improvement in terms of life cycle and data success rate.

Abstract:The problem of aggregated multicast in optical transmission networks has been known to be a complete NP-hard problem. A double neighborhood search algorithm (DNSA) for solving the multicast aggregation problem is presented. The objective of this algorithm is to minimize the bandwidth wastage ratio that is subject to the constraints that the number of multicast aggregation tree is affected by the amount of wavelength. In this paper, a priority aggregate rule based on the greedy strategy is proposed to generate initial solution: two kind neighborhood structures are proposed to search effectively, and some off-trap strategies are proposed to jump from local optimal solution and carry the search to the feasible areas in promising directions. Simulation experiments show the double neighborhood search algorithm can aggregate multicast trees effectively. The multicast group blocking ratio is 0 at light load. Compared with other algorithms at heavy load, the average bandwidth wastage ratio has decreased more than 25%. The results indicate that improvements may be obtained in different network condition.

Abstract:Recording flow statistics for each network packet is resource-intensive. Various sampling techniques are used to estimate flow statistics. However, the estimation accuracy based on the sampling remains a significant challenge. This paper introduces both sampling techniques denoted as Integral and Iteration algorithms, which can accurately infer the number of original flows from the sampled flow records. The Integral algorithm uses only the number of sampled flows with one sampled packet to approximately deduce the number of unsampled flows. The Iteration algorithm can estimate the number of unsampled flows using an iteration method. The number of original flows can be precisely estimated according to both the number of sampled flows and unsampled flows. Both the algorithms are compared to the EM (expectation maximization) algorithm using multiple traffic traces collected from CERNET (China education and research network) backbone. The result shows that the Iteration algorithm is superior to the EM algorithm and can provide highly accurate estimation on the number of original flows.

Abstract:In traditional public-key encryption schemes, security guarantees will be fully lost once decryption secret keys are exposed. With the ever-increase in encryption systems used in mobile and low secuirity devices, key exposure seems unavoidable. An intrusionresilient public-key encryption is proposed to mitigate the damage for the encryption systems brought by key exposure, which provides more security than the forward-secure encryption and key-insulated encryption. In its primitive, the whole lifetime is divided into discrete periods where the public key is fixed. Secret keys are shared in a decrypter and a base. The former performs the decrypting operations on his own while the latter provides an updated message to help evolve secret keys in each period. Furthermore, multiple operations of refresh secret keys are performed to refresh decrypter and base secrets periodically. The security can be preserved when both the user and base are compromised, as long as they are not compromised simultaneously. In addition, the simultaneous compromise doesn't affect the security of the ciphertext generated in previous periods. This paper proposes an intrusion-resilient public-key encryption scheme. All the parameters in this scheme have at most a log-squared complexity in terms of the total number of time periods. The proposed scheme is proven to be secure in the standard model and is a provably secure intrusion-resilient public-key encryption scheme without random oracles.

Abstract:MIMO (multi-input multi-output), a promising technology that hopes to improve reliability and bandwidth of wireless networks, has been widely deployed. However there has been little literature that describes how to provide high-quality video service for multi-user over MIMO multi-hop wireless networks. The co-channel interference among multiple wireless links is one of the critical issues. This paper proposes a scheme for multi-user video transmission over multi-hop wireless networks and combats with interference by employing link selection, spatial multiplexing and spatial diversity gain of MIMO. Furthermore, the study formulates this scheme as an optimization problem, which is a NP-hard problem. To make it suitable for practical implementation, the study employs a modified genetic algorithm to solve the link selection problem. Moreover, according to the characteristics of scalable video and MIMO, the antenna grouping problem is transformed into one 0/1 knapsack problem. The depth first search and branch bound techniques are used to further reduce the searching range. The study demonstrate the effectiveness of the proposed scheme by extensive experiments.

Abstract:Under the new application mode, the traditional hierarchy data centers face several limitations in size, bandwidth, scalability, and cost. In order to meet the needs of new applications, data center network should fulfill the requirements with low-cost, such as high scalability, low configuration overhead, robustness and energy-saving. First, the shortcomings of the traditional data center network architecture are summarized, and new requirements are pointed out. Secondly, the existing proposals are divided into two categories, i.e. server-centric and network-centric. Then, several representative architectures of these two categories are overviewed and compared in detail. Finally, the future directions of data center network are discussed.

Abstract:Time-Space consistency (TSC) is a crucial and fundamental problem in service-oriented distributed simulation application (SODSA), but current researches have been inclined to consider the concepts of special time and space and have been unable to achieve the global TSC in whole modeling and simulation. By analyzing requirements of generalized TSC and its inconsistent factors, the study presents the global axis of TSC. The notion generalized time-space hierarchical consistency is proposed, consisting of formal definition, of the consistency-resource-model-service-perceive model (C-RMSP), consistency-time-space maturity (C-TSM), and consistencyhierarchical framework (C-HF). The research achievements in this paper reinforce the understanding of global time-space demands of consistency and support global quantitative evaluation on TSC of distributed simulation applications in the future, which are beneficial to resolving inconsistent issues in the whole simulation system.

Abstract:Object-Based storage is a good choice for large scale storage systems. Load balancing of metadata is important to improve the performance of I/O. The existing load balancing schemas cannot evenly distribute the accesses of metadata in a dynamic way. Moreover, the adaptability and fault-tolerance ability need to be improved. This paper presents an adaptable distributed load balancing of metadata (ADMLB) which is composed of basic load balancing algorithm (BBLA) and distributed incremental load balancing algorithm (IBLA). Specially, ADMLB first uses BBLA to distribute metadata loads according to the performances of the metadata servers and then uses IBLA to incrementally reorganize loads on each metadata server. ADMLB can evenly distribute loads between metadata servers and adapts well to the changes of loads. It also has good fault-tolerance ability, and locates metadata servers very quickly.

Abstract:This research focuses on the problem of operating system (OS) scheduling on asymmetric multi-core processors (AMP). A task scheduling model based on linear programming is proposed. Several attributes of AMP factors are taken into account in this model. Scheduling principles of behavior matching, migration avoiding, and load balancing are adhered as well. A comprehensive scheduling algorithm is also proposed based on the model. The algorithm has two parts: an integrated workload characterization, which proposes integrated behavior to measure the global and local behaviors of tasks comprehensively, and an integrated behavior- based scheduling algorithm, which efficiently utilizes the asymmetric multi-core processors without frequent task migration. This guarantees the load balance between cores. In addition, the algorithm achieves universality with a flexible parameter adjustment mechanism. It is an algorithm to achieve universality as well as the first to handle the global and local behaviors of tasks comprehensively. The evaluation on real platform demonstrates that the algorithm is universal for different conditions, and it always outperforms other scheduling algorithms on asymmetric multi-core processors (by 6%～22%).

Abstract:Virtualized resources management for service level objectives (SLOs) of applications has been one of the key problems of system management in current data centers. To solve this problem one needs to: 1) dynamically and distributed allocating resources to virtual machines (VMs) of applications on demand; 2) efficiently control interference among VMs consolidated on a single physical server, due to their contention on non-virtualized resources. Many existing methods, however, are not suitable for this virtualized data center scenario. This paper presents a method for the virtualized resource management for SLOs of applications. First, based on the feedback control theory, this method can achieve SLOs of applications through dynamically resourced allocation. Second, a two-layer self-adaptive mechanism is devised and used to dynamically capture the non-linear relationship between the performance of applications and the resources allocation. Third, this method can control the performance interference among VMs on non-virtualized resources through virtualized resources allocation. The study has evaluated this method on the RUBiS system and TPC-W benchmark in a Xen-based virtualized cluster. The experimental results show that the average rate of SLOs achieved by this method is 29.2% higher than ones by two existing methods. Along with the average deviation from SLOs, this method is 50.1% lower than ones of the existing methods. Furthermore, when resource contention occurs on non-virtualized disk I/O between RUBiS and TPC-W, this method can almost entirely satisfy the disk I/O requirement of RUBiS of high priority through restraining TPC-W requests, e.g. 28.7%, on virtualized CPU.

Abstract:This paper suggests that the periodic scheduling of identical independent tasks on a single-level tree grid and identical independent tasks are independent tasks in the same size in computation. First, an integer linear programming model for scheduling identical independent tasks is presented. By analyzing the procedure of solving the linear programming model, and obtaining the optimal number of tasks assigned to each computing node, the study finds that a single-level tree grid composed of different nodes will be in one of three states, respectively: unsaturated, critical, and redundant states. Furthermore, the optimal number of tasks assigned to each node increases linearly as the number of tasks arriving to the master grows, and the periodic feature appears in the task scheduling. Next, some features and theories that determine the states of grid systems among unsaturated, critical, and redundant states are obtained, periodic scheduling of identical independent tasks on single-level grid is proposed, the length of period is derived. Their computational complexity is also analyzed, and a periodic scheduling algorithm, Periodic-Sched, is given. Both the theory and experiments have proved that the periodic scheduling is feasible. In this way, the problem of large-scale independent task scheduling is simplified and resolved in one period, which reduces the computational complexity dramatically. The periodic scheduling of independent tasks is suitable for Map tasks scheduling on Hadoop platform.

Abstract:As an unstructured-grid high efficient solver, the multigrid algorithm, with its serial and parallel application, can achieve the optimal properties of being on time and having space complexity. To illustrate the numerical simulation process of an unstructured grid, this paper begins with the discretization of governing equations and points out that the multigrid algorithm is mainly used for solving large scale algebraic equation, which is derived from the time marching scheme. For the multigrid algorithm, the study briefly describes its basic structure and efficient principle. Secondly, the paper reviews research that trend about the geometric multigrid and algebraic multigrid and discusses the basic design principles and hot topics on parallelization. At the same time, for the practical application of unstructured grid, the paper summarizes and classifies many smoothers, followed by examples of open source software about unstructured grid industrial application. Finally, some applications and key problems in this field are highlighted, as well as the future progress of parallel multigrid solver on unstructured grid.

Abstract:To evade the detection of security monitoring systems, malware often hides its behavior. Current monitoring systems usually reside in the operating system (OS). Thus, it is hard to detect the existence of malware, especially the kernel rootkits. In this paper, a hidden OS objects detection and correlation approach based on VMM (virtual machine monitor) is proposed, and the corresponding detection system, vDetector, is designed and implemented. Both implicit and explicit information are used to create multiple views of OS objects, and a multi-view comparison mechanism are designed to identify three kinds of hidden OS objects: process, file and connections. The relations among hidden objects are established based on OS semantic information to trace the complete attack path. vDetector is implemented based on KVM virtualization platform and the effectiveness and performance overhead of vDetector are evaluated by comprehensive experiments. The results show that vDetector can successfully detect the existence of hidden OS objects with reasonable performance overhead.