Wenzhong Guo

A Cloud-based monitoring framework for Smart Home

Today, Smart Home monitoring services have attracted much attention from both academia and industry. However, in the conventional monitoring mechanism the remote camera can not be accessed for remote monitoring anywhere and anytime. Besides, traditional approaches might have the limitation in local storage due to lack of device elasticity. In this paper, we proposed a Cloud-based monitoring framework to implement the remote monitoring services of Smart Home. The main technical issues considered include Data-Cloud storage, Local-Cache mechanism, Media device control, NAT traversal, etc. The implementation shows three use scenarios: (a) operating and controlling video cameras for remote monitoring through mobile devices or sound sensors; (b) streaming live video from cameras and sending captured image to mobile devices; (c) recording videos and images on a cloud computing platform for future playback. This system framework could be extended to other applications of Smart Home.

Mobile Cloud Based Framework for Remote-Resident Multimedia Discovery and Access

Due to realistic barriers (e.g., network heterogeneity, NAT traversal, limited computing/storage, etc.), a majority of legacy devices (especially mobile devices) are difficult to discover and access residential services from remote environment. In this paper, we propose a mobile cloud based architecture for enabling remote-resident multimedia service discovery and access. The main considered issues in the paper includes: security, cloud-based storage, cloud-based remote service discovery and control, REST-based lightweight negotiation, etc. After that, AHP and fuzzy TOPSIS algorithms are combined for cloud candidate selection optimization. Prototype implementation and theoretical analysis show availability and efficiency of proposed approaches.

Architecture-based integrated management of diverse cloud resources

Cloud management faces with great challenges, due to the diversity of Cloud resources and ever-changing management requirements. For constructing a management system to satisfy a specific management requirement, a redevelopment solution based on existing management systems is usually more practicable than developing the system from scratch. However, the difficulty and workload of redevelopment are also very high. As the architecture-based runtime model is causally connected with the corresponding running system automatically, constructing an integrated Cloud management system based on the architecture-based runtime models of Cloud resources can benefit from the model-specific natures, and thus reduce the development workload. In this paper, we present an architecture-based approach to managing diverse Cloud resources. First, manageability of Cloud resources is abstracted as runtime models, which could automatically and immediately propagate any observable runtime changes of target resources to corresponding architecture models, and vice versa. Second, a customized model is constructed according to the personalized management requirement and the synchronization between the customized model and Cloud resource runtime models is ensured through model transformation. Thus, all the management tasks could be carried out through executing programs on the customized model. The experiment on a real-world cloud demonstrates the feasibility, effectiveness and benefits of the new approach to integrated management of Cloud resources

Online optimization scheduling for scientific workflows with deadline constraint on hybrid clouds

The tremendous parallel computing ability of cloud computing encourages investigators to research its drawbacks and advantages on processing large‐scale scientific applications such as workflows. The current cloud market is composed of numerous diverse public clouds and a local private cloud, and workflow scheduling is one of the biggest challenges on hybrid clouds due to the highly fragmented cloud market with respect to service provisions, pricing models, and bandwidths. In this paper, we propose an online‐scheduling strategy for continuous submitted scientific workflows on hybrid clouds, which aims to complete the deadline‐constrained applications as more as possible at a lower price. Firstly, a hierarchical iterative application partition (HIAP) algorithm is proposed to partition the application into a set of dependent tasks. Moreover, many online‐scheduling algorithms cooperated with HIAP are presented to finish the workflows with a low average payment. Our strategy takes into account the basic characteristics on hybrid clouds such as bandwidth constraints, data transfer cost and computational cost. Various well‐known workflows are used for evaluating the multiple scheduling algorithms and the results show that the MLF_ID approach can achieve a promising performance. Copyright

Interest prediction in social networks based on Markov chain modeling on clustered users

Effective user interest prediction is significant for service providers in a set of application scenarios such as user behavior analysis and resource recommendation. However, existing approaches are either incomplete or proprietary. In this paper, user interest prediction based on the Markov chain modeling on clustered users is proposed with the following procedure: collect dataset from 4613 users and more than 16 million messages from Sina Weibo; obtain each users interest eigenvalue sequence and establish single‐Markov chain model; and implement user clustering algorithm for the multi‐Markov chain construction in order to divide users into a set of predefined interest categories. The proposed solution is capable of predicting both long‐term and short‐term user interests based on a suitable selection of the initial state distribution, λ. The proposed solution also proves that short‐term interests are consistent with long‐term interests if the influences of social or user‐related events that cause interruptions (e.g., earthquake and birthday) are not considered. Furthermore, experiments show that the proposed solution is feasible and efficient and can achieve a higher accuracy of prediction than that of the other approaches such as Support Vector Machine (SVM) and K‐means. Copyright

Energy-Efficient VM Placement Algorithms for Cloud Data Center

Cloud is the computing paradigm which provides computing resource as a service through network. The client can use computing resource in a convenient and on-demand way, just like the water and the electricity we use daily. The mapping between virtual machine and physical machine is the key of the VM scheduling problem. Nowadays we advocate low-carbon life. It calls for the green cloud computing solutions whether protecting the environment or saving the cost of cloud suppliers. The proposed VM placement algorithm is energy-efficient, and considers the multi-dimentional resource constrains, such as CPU, memory, network bandwidth, and so on. The experimental results show that the proposed algorithms not only contribute a lot to energy saving, but also try best to meet the quality of service QoS. Therefore, we make significant savings in operating cost and make full use of various resources in the cloud data center. The algorithm has promising prospect in application.

Real-Time Task Scheduling Algorithm for Cloud Computing Based on Particle Swarm Optimization

As a new computing paradigm, cloud computing is receiving considerable attention in both industry and academia. Task scheduling plays an important role in large-scale distributed systems. However, most previous work only consider cost or makespan as optimized objective for cloud computing. In this paper, we propose a soft real-time task scheduling algorithm based on particle swarm optimization approach for cloud computing. The optimized objectives include not only cost and makespan, but also deadline missing ratio and load balancing degree. In addition, to improve resource utilization and maximize the profit of cloud service provider, a utility function is employed to allocate tasks to machines with high performance. Simulation results show the proposed algorithm can effectively minimize deadline missing ratio, maximize the profit of cloud service provider and achieve better load balancing compared with baseline algorithms.

Data Replication Placement Strategy Based On Bidding Mode for Cloud Storage Cluster

The data availability in large-scale cloud storage has been increasing by means of data replica. To provide cost-effective availability, minimize the response time of applications and make load balancing for cloud storage, a new replica placement policy with bidding thought is proposed. The policy combines the own characteristics of replica and factors of bidding mode(e.g. bidding time, bidding standard, bidding price etc.) and starts replica bidding activity when the file data availability cannot meet the given requirement. Replica placement is based on capacity and accessing probability of data nodes. The experimental results show that our policy has a better performance in both load balance and response time comparing to the static replica policy and CDRM scheme.

A Dynamic-alliance-based Adaptive Task Allocation Algorithm in Wireless Sensor Networks

In distributed sensor networks, which have limited resources, such as energy and storage, and work in a dynamic environment, the networks should effectively allocate some real-time tasks on those limited resources. Additionally, we should do our best to maximize the lifetime of wireless sensor networks (WSNs) and the accuracy of the results. Due to most of previous works focusing on static task allocation for WSNs and only a few works having paid attention to dynamic resource allocation for sensor networks, this paper present an effectively adaptive task allocation (EATA) in WSNs which applied dynamic alliance. Instead of being aid of manual adjustment, each node can autonomously adjust its parameters and state by means of EATA according to tracking the change of environment, such as energy depletion. By comparing with static task allocation, experiment results show that our scheme can save a great deal of energy and prolong the lifetime of the network.

A Runtime Architecture Based Framework Managing Hybrid Clouds

Cloud management becomes increasingly complex and brings high costs, especially with the advent of hybrid cloud. In a hybrid cloud, numerous resources like Virtual Machines (VMs) and Physical Machines in different clouds have to be managed together to make the whole hybrid cloud work cost-effectively. For controlling the management cost, in particular the manual management cost, many programs have been developed to take over manual management tasks or reduce their complexity and difficulty. These programs are usually hard-coded by languages like Java and C++, which bring enough capability and flexibility but also cause high programming effort and cost. As the architecture-based runtime model is causally connected with the corresponding running system automatically, constructing a hybrid cloud management system based on the architecture-based runtime models of clouds can benefit from the model-specific natures, and thus reduce the development workload. This paper proposes a runtime architecture based approach to developing the management programs in a simple but powerful enough manner. First of all, the manageability (such as APIs, configuration files and scripts) of different kinds of clouds, is abstracted as a runtime architecture based model of cloud software architecture, which can automatically and immediately propagate any observable runtime changes of the target platforms to the corresponding architecture models, and vice versa. Second, we provide a unified model of cloud software architecture, according to the domain knowledge of current cloud platforms, such as Cloud Stack, Open Stack and Eucalyptus. Third, the synchronization between the unified model and cloud runtime models is ensured through model transformation, thus, all the management tasks of the hybrid cloud, could be carried out through executing programs on the unified model, which decreases the complexity of use and management. The experiment on a real-world hybrid cloud demonstrates the feasibility, effectiveness and benefits of the new approach to managing hybrid clouds.