Based on the characteristics of land resources information management, this research firstly designs a land resources cloud computing platform. Four main strategies, namely job scheduling, cloud cache pool, Map Reduce and resources-oriented techniques are then discussed in details. A prototype system named Land Cloud is then introduced, followed by two experiments which are conducted to test the performance of the system. The results show enough evidence of the usefulness of cloud computing services in managing land resources data.
Volunteer computing has recently received the strong attraction for executing high throughput applications as CPU, storage and network capacities improve and become cheaper. Volunteer resource can provide computing power for cloud computing in many respects, but there is no general survey for appraisal of the suitability of the organization's structure and resources. Therefore, we propose a new comprehensive taxonomy and survey of VCS in order to characterize and introduce the deployment process of VCS.
In web service computing, a business transaction is integrated by web services provided by multiple enterprises, therefore the transaction is deployed in a loose-coupled, heterogeneous and autonomous environment. When failures or exceptions occur, the traditional rolling back technique is inappropriate for recovery since locking is inapplicable. Current solutions, including forward recovery and backward recovery, didn't take into consideration enough business logic of the business transaction and thus were lack of flexibility and efficiency. In this paper, we introduce the conception-constraint rule, which express the business logic of a business transaction. A better control of recovery is achieved based on the constraint rules. A detailed description of our algorithms for recovery is given and a travel agency case study is illustrated. In addition, we compare our recovery mechanism with some existing solutions and the simulation results show that our method has both better success rate and efficiency.
Business-to-Business (B2B) workflow/service interoperation across Virtual Organisations (VOs) brings about novel business scenarios. In these scenarios, parts of workflows (or services) corresponding to different partners can be aggregated in a producer-consumer manner, leading to hierarchical structures of added value interpreted as service value chains. Service Level Agreements (SLAs), which are contracts between service providers and service consumers, guarantee the expected quality of service (QoS) to different stakeholders at various levels of a service value chain. The essential requirements for such an SLA-based choreography of services include agile component-based infrastructure to support the corresponding choreography of their SLAs; proactive validation of SLAs to prevent violations; reactive validation of SLAs for penalty enforcement and breach management; and business enabling requirements such as trust, privacy, security and automation. In this paper we highlight the significance of these issues and propose their solutions. We then proceed to weave these solution components together into a comprehensive SLA validation framework that blends together two major systems: a rule based distributed system for SLA validation and a service monitoring system for the prevention of SLA violations.
Traditional business process modeling approaches focus on the activities, data flows are an afterthought. Concerning cross-organizational collaboration on the Internet, data flow is still far from optimal, causing resource contention and unnecessary data exchanges. In this paper, a data-driven approach to Internet based business collaboration is proposed, which treats data as a centerpiece as data objects and associates activities with the data objects. The data objects hold information structures pertinent to global context to support logic data flow, both the resource contention and data exchanges can be alleviated. A case study shows that the approach works efficiently.
VM (virtual machine) scheduling is a fundamental topic of virtualization, and fairness is its important design goal. Most VMMs (virtual machine monitor) provide PS (proportional share) schedulers. A PS scheduler assign a weight to every VM to declare the computational resource requirement, and VM is allocated CPU cycles proportional to weight. CPU-affinity is a property of VCPU (virtual CPU) to describe which PCPUs (physical CPU) it can run on. However, current definition of weight does not collaborate well with CPU-affinity. It behaves as if all PCPUs are available to all VMs. We expose the issue by using Xen's Credit scheduler, and show that CPU cycles can be fairly allocated to VMs in the default free-mapping case (no affinity restriction), but not so in restricted-mapping cases (when affinities are configured). This fairness issue makes it necessary to extend the meaning of weight, so as to reflect the resource requirements in all cases. In this paper we reconcile the meaning of a domain's weight with CPU-affinity, and improve Credit scheduler as an affinity-aware one. Experimental results show that our affinity-aware proportional share scheduler (named Credit-APS scheduler) brings a good fairness both in free-mapping and restricted-mapping cases.
This paper explores the relationship between I/O scheduling in a virtual machine monitor (VMM) and performance of Map Reduce applications running on the virtual machines (VMs). Traditionally, Map Reduce applications running on the virtual machines perform worse than on the physical machines due to competing for I/O. In this paper, we discuss the effect of I/O scheduling on performance of Map Reduce running on VMs. We present two strategies to improve the I/O scheduling and evaluate the performance through experiments. The results show that the strategies could reduce the percentage of I/O waiting and time consuming of Map Reduce applications effectively.
This research is aimed to reduce the complexity of heterogeneous spatial data in system that services user on demand. Heterogeneous include heterogeneous standards, heterogeneous systems, heterogeneous mode and the heterogeneous contents of data. For this purpose, we describe the data conformably with Grid based on Web Service. This research implemented the method, and provided a set of middleware and standards. And, on the data service convergence layer where services can be used, this paper proposes the solutions of service semantic heterogeneous phenomenon by the use of spatial information ontology.
This paper presents AMREF, an Adaptive Map Reduce Framework designed for an effective use of computational resources in data center networks to deal with real time data intensive applications. AMREF entails its adaptivity from adaptive splitter, adaptive mappers and adaptive reducers in a stochastic control manner. We use three methods, feedback control, stochastic learning with smooth filter and kalman filter to implement the framwork. Comparison among the three methods suggests they can be effectively and efficiently used to reduce the makspan in three different real-world workload scenarios.
Earth Observation (EO) is an effective way to monitor hazards for rapid response. The sensors such as optical, passive microwave, Synthetic Aperture Radar (SAR) can be used. For rapid response purpose based on EO data, fast data acquisition from data provides and processing in near real-time are two key issues. In this paper, the automatic data integration method based on metadata harvest interface is introduced. Distributed computing resources are collaborated based on gird computing to process the RS data into flood map. And grid workflow technique is used to manage and monitor the flood detecting processes. Flood is used as use case. The monitoring results from multi-sensor data can be used as time-serial monitoring of the flood extent and detection results from different sensors can be cross-validated. SAR and MODIS data from distributed websites and flood detection algorithms base on these data are used for test. An integrated web portal test bed based on data integration service and flood detection algorithms is developed. The test bed can be used to request and automatic access distributed RS data source and process RS data on demand to get flood extent in near real time.
With the development of 3G network technology, the rapid upgrading of mobile phone hardware, and the refined requirements for mobile phones by mobile phone users, it is essential for mobile phones to be able to promptly and accurately process images. A mobile phone terminal sends the service requests to cloud computer system which then processes the requests and sends the results via 3G network back to the mobile phone terminal. This paper presents a architectural model to realize the image processing system that connects mobile phone with clouding computing through 3G network. The method is to build intra-enterprise server group, distributed database system, and parallel computing system to form a private cloud computing system, which defines ontological semantic tags for images uploaded from mobile phone terminals and stores them in a distributed way (cloud storing), and to design the corresponding search algorithms to help the system to promptly and accurately query the images (cloud computing), and retrieve the appropriate images through sorting matching indexes, weights, and etc.
Traffic congestion has become a major problem in today’s society. It may cause a great waste of material and human casualties. We think the fixed-time-state traffic light control scheme is one important reason. An intelligent traffic light control scheme is necessary in order to give a solution to this problem. The paper proposes a model to describe a road cross. In terms of dynamic and real-time traffic information, weights of all the serial lanes are dynamically computed, so as to automatically manage the current status and duration of traffic lights. The proposed novel control scheme consider traffic situations of both upstream lanes and downstream lanes, in order to reduce average vehicle waiting time at a road cross. The paper further performs simulations for the proposed scheme and gives detailed analysis on parameters which have great influence on scheme performance. Results show that the proposed scheme outperforms the fixed-time-state traffic light control scheme on average vehicle waiting time.
Because of the bottleneck of disk I/O, the distributed file system based on disk is limited in the performance on data throughput and latency. It is a big challenge for such a system to meet the high performance requirement of the massive small-file storage. Cache has been widely used in storage system to improve the data access performance. In order to support the storage of massive small files, we have integrated memcached into our distributed file system to optimize the storage of massive small files. However eviction problem arose from LRU replacement algorithm in memcached. It means that the non-stale objects might be replaced due to large short-lived objects. Therefore, we proposed Prioritized Cache (PC) and Prioritized Cache Management (PCM) to solve the problem. The cache of memcached is reorganized and classified into permanent cache and temporary cache. Furthermore, in order to alleviate side effects on hit rate in sequential access, temporary cache is partitioned into different parts with different priorities and managed according to the priorities. We have implemented and evaluated the integrated prototype system. The experimental results show that the improved distributed file system with distributed object cache can deliver high performance on small-file storage. Compared with the original system, the read of small files increased by a factor of 2.65 ~ 8.05, without write performance degradation.
Replica management has become a hot research topic in data grid, It improves the availability and fault tolerant of data in grid computing environment. Based on the deep research on the mainstream replica management technology in traditional data grid, we propose a dynamic replica management strategy. Our replica management strategy consists of the creation method of dynamic replica that can automatically increase replica according to the frequency of the file access, the selection method of replica based on the GridFTP and the replacement method of the replica combining with the establishment time, number of access and file size. The simulation experiments indicate that our dynamic replica management strategy behaves much better than the five built-in replica management strategies in OptorSim grid simulator when the resource access is coherent to time.
The instability of the network and storage nodes may lead to the failure of the storage process in the network storage environment. In this paper, we establish the realibility model of the storage node availability, and the probability model of a success file block storage process, give the Store-After-Partition Strategy, compare the probabilities of a success file storage process at different partition granularities, and finally put forward the optimal strategy of file partition among network storage environment. This strategy is valuable for network storage facing massive files.
This paper designs and implements a load balancing scheme based on dynamic resource allocation policy for virtual machine cluster, which are running under para-virtualization mode on a cluster of physical machines (PM) in shared storage architecture. It monitors the real-time resources utilization of VMs and PMs, including CPU, memory and network, then uses instant resource reallocation for virtual machines (VM) running on same PM to achieve local VMs’ load balancing, while uses live migration of VMs among PMs to achieve global VMs’ load balancing. It optimize the resource allocation of VMs to achieve global load balancing of virtual machine cluster. The resource utilization of physical machines will be improved as well. Compared to traditional load balancing schemes based on task scheduling, it is application independent and works seamless on VMs hosting different kinds of applications.
With the blast of information available on the web and the relevant documents set returned by search engineer can not meet the user’s needs, the question answer system research is promoted. In this paper, we introduce the structure of Question Answer system based on SOA, define the structure of the dictionary of Keywords and searching template, the word segmentation algorithm based on the dictionary of keyword, the storage of searching template and the algorithm of template matching. On the foundation, we implement a QA system for Railway domain application, the experimental result show that QA system based on techniques we employed has high accuracy and recognition rate.
Constellation networking is the prominent technology for the next generation satellite systems. The communication schedule is a key problem for constellation maintenance. However, the traditional SRS (Satellite Range Scheduling) model do not use inter-satellite links. In this paper, a communication scheduling model is proposed, in which the ISLs (Inter Satellite Links) is built. And the NP complicity of SNRS is proved. Meanwhile, a two stage greedy algorithm is designed to calculate the SNRS. And heuristic strategies in the proposed algorithm are compared by simulation. The experiment results show that EDF strategy combined with minimum distance route can bring into an acceptable schedule.
As software continues to expand the scale and more and more reliability requirements come from users, how to measure reliability of each component of the software has become a key issue in software development. So, in this paper, a novel guarantee-verification method based on cloud service was proposed from the aspect of component service. First, the component’s trustworthiness was defined in terms of component service guarantee-verification. Secondly, an effective component model was defined. Thirdly, based on this model, the process of verifying a component service by atomic service extension was given and illustrated in an application sample. This model can provide an effective verification method which can automatically determine whether a component service satisfies a certain user’s requirement or not. And it can also help improve the usability of a software component and the efficiency of component based software development.
The paper proposes a publicly verifiable threshold decryption scheme with no center based on the situation that a trust center does not exist in many special occasions. The scheme asks decryption member for submitting commitment in the decryption process in order to make sure the identity is publicly verifiable. Moreover, the scheme has several advantages such as protecting the shadow without revealing and the members can be deleted dynamically.
