Advances in Services Computing

With the popularity of smartphones, an increasing number of applications (app) are installed on common users’ smartphones. As a result, it is becoming difficult to find the right apps to use promptly. Based on the observation from the real data, it can be found the correlative relationship exists between the usage of app and the context, specifically, time and location contextual information. According to this analysis, a context-aware usage prediction model is proposed to predict the probability of launching apps and present this prediction by an ordered list. Furthermore, a dynamic desktop application for android platform is developed to adjust the app icon order on the desktop according to the current time and location information, which facilitates the smartphone users always capable finding their needed ones in the first page. The experiments show that our prediction model outperforms other approaches.

The problem of service rule verification has attracted some attention in recent years. In this paper, we consider service rules in simple Boolean logic and present a new method for business rule verification using simultaneous minimal support set computation. As we show here, the problem is similar in flavor to the problem of prime implicant generation of a given Boolean function which has alluded researchers for several decades and significant efforts in this direction have been reported in literature, with proposals of widely varying algorithms and data structures. In this paper, we revisit this problem in the context of business rules and present a new method that aids in rule verification and also in query execution at runtime. Our method builds on the classical binary decision diagram data structure for representing business rules and generates the test scenarios by a simple traversal algorithm. Experimental results on simulated benchmark rules show the efficacy of our approach.

Dynamic resource provisioning is an important way of ensuring performance and Service Level Agreement (SLA) guarantees for applications under changing workload. However, it is always hard to meet exactly the amount of resources required at every second. Thus, how to optimize the resource provision becomes the key problem. In this paper, we propose a Reactive-Predictive Hybrid Resource Provision Method (RPHRPM), which combines reactive and predictive methods together to benefit from both. We take advantage of ARIMA model to predict the workload and get resources pre-provisioned. Meanwhile, a reactive method is also enabled to deal with the unpredictable situations. More importantly we describe a novel mechanism which will be involved when conflicts between these two methods happen. It can help to keep better performance when encounter could burst. The experiment results show that RPHRPM not only has better performance compared with other provision schemes, but also be energy-efficient.

With the popularity of social network, an increasing number of users attempt to find their interested web services through service recommendation, e.g., Collaborative Filtering (i.e., CF)-based service recommendation. Generally, the traditional CF-based service recommendation approaches work, when the target user owns one or more similar neighbors or friends (Neighbor and friend are interchangeable in the rest of paper) (i.e., user-based CF), or the target user’s invoked services own similar services (i.e., item-based CF). However, in certain situations, similar neighbors and similar services are absent from the user-service invocation network, which brings a great challenge for accurate service recommendation. In view of this challenge, a novel recommendation approach SBT-SR (Social Balance Theory-based Service Recommendation) is put forward in this paper. Concretely, for the target user, we first determine his/her “enemies” (antonym of “friend”, i.e., the users who have opposite preference with target user), and then look for the “potential friends” of target user, based on the “enemy’s enemy is friend” rule in Social Balance Theory. Afterwards, the services preferred by “potential friends” are recommended to the target user. Finally, through a case study and a set of experiments, we demonstrate the feasibility of our proposal.

Network systems employ policies that are inherently dynamic in nature and that depend on temporal conditions defined in terms of external events such as the measurement of bandwidth, use of hosts, intrusion detection or specific time events. Software-defined networking (SDN) offers the opportunity to make networks easier to configure by providing richer configuration methods. To reduce network monitoring costs and traffic overheads, herein, we propose a software-defined cloud resource management framework that uses a Fuzzy Analytical Hierarchy Process (Fuzzy-AHP) to customize the network resource allocation. The framework can be incorporated into SDN-enabled cloud infrastructures by using an Application Program Interface (API). Using real-time data, we demonstrate that our framework can improve network resource management and is capable of handling increasing traffic requests. We also validate our framework efficiency through simulations.

Service reuse must follow certain constraints in order to correctly interact with Web Services. Violations of constraints can cause fatal errors or incorrect results in the service reuse. However, constraints are often not formally specified and are thus not available in the service reuse. To address this issue, this paper focuses on two common types of such constraints, including location constraints on Web Services and object constraints on simple parameters. An approach is proposed to clarify the two constraints automatically, via a hybrid analysis of heterogeneous information, including the social tags and the service documentations. Then an improved method is presented to identify collaborative relations among Web Services, integrating constraints compatibility into semantic matching. One experiment is carried out on 509 Web Services crawled from the Internet to evaluate the effectiveness of our approach. The other experiment is conducted on the same dataset to assess impacts of the two constraints on service relations. Experimental results show that our approach can clarify the two types of constraints effectively and achieve adequate recall and precision. Moreover, it is indicated that the two types of constraints, especially object constraints, have significant impacts on improving the quality of identified service relations, thus provide a strong guarantee for accurate service reuse.

Recent years have witnessed an increasing number of services published on the Internet. How to find suitable services according to user queries remains a challenging issue in the services computing field. Many prior studies have been reported towards this direction. In this paper, we propose a novel service discovery approach by mining and matching common topic groups. In our approach, we mine the common topic groups based on the service-topic distribution matrix generated by topic modeling, and the extracted common topic groups can then be leveraged to match user queries to relevant services, so as to make a better trade-off between the number of available services and the accuracy of service discovery. The results of experiments conducted on a publicly available data set show that compared with other widely used methods, our approach can improve the performance of service discovery by decreasing the number of candidate services.

Context-Aware Recommender System is known to not only recommend items or services similar to those already rated with the highest score, but also consider the current contexts for personalized Web services recommendation. Specifically, a key step for CARS methods refers to previous service invocation experiences under the similar context of the user to make Quality of Services prediction. Existing works either considered the influence of regional correlations on user preference, or combined the location-aware context with the matrix factorization method. However, the user preference expansion triggered by instant update of user location is not fully observed. For instance, when making Web service recommendation for a user, it is expected to be aware of rapid change of the user location immediately and the expansion of user preference as well. In this paper, we propose a Web services recommendation approach dubbed as CASR-UPE (Context-aware Web Services Recommendation based on User Preference Expansion). First, we model the influence of user location update on user preference. Second, we perform the context-aware similarity mining for updated location. Third, we predict the Quality of Services by Bayesian inference, and thus recommend the best Web service for the user subsequently. Finally, we evaluate the CASR-UPE method on WS-Dream dataset by evaluation matrices such as RMSE and MAE. Experimental results show that our approach outperforms several benchmark methods with a significant margin.

In cloud, resources are virtualized and the software delivery way is becoming something like a “service” to provide end user and operator benefits including on-demand self-service, resource pooling, rapid elasticity and service metering capability. As a part of network function virtualization, firewall virtualization can greatly increase the firewall configuration flexibility for the cloud environment. In this paper, we focus on FWaaS (Firewall as a Service) and we design a parallel firewall system called CPFirewall (Cloud Parallel Firewall System). In CPFirewall, the firewall resources are virtualized and multiple tenants can build up their own parallel firewall by renting virtual firewalls. This needs solve some challenges. We adopt a rule-splitting algorithm to build a rule anomaly set (We call it Wrapset.) for detecting rule anomaly. We design the rule-allocation algorithm to achieve the cloud-native features, including load balance and dynamic scale. And we also improve the system performance using Exponential Smoothing (ES) forecasting method. Experiment results have verified that CPFirewall has a higher efficiency than other firewall schemes and is much more suitable for the Cloud network environment.

As a fundamental phrase in the life cycle in SOA, service identification has a huge impact in building up SOA based applications. Several service identification methods focus on the definitions of loosely coupled and a high cohesion inside services. There is a majority using business process as input. Because of the simplification of relation between process in most of the process modelling language, dependency between business process is ignored. However, dependency is an inevitable factor to performance of future system. In this paper, we proposed a procedure of dependency aware process analysis for service identification method to ensure not only the characteristics of SOA but also the dependency between services. With this procedure, we tried to have a group of services with visible dependency from analysing the business process and requirements.

Cloud computing provides dynamic resource allocation using virtualization technology to greatly improve resource efficiency. However, current resource reallocation solution seldom considers the stability of VM placement pattern. Varied workloads of applications would lead to frequent resource reconfiguration requirements due to repeated occurrence of hot nodes. In this paper, a multi-objective genetic algorithm (MOGA) is presented to significantly improve the stability of VM placement pattern with less migration overhead. The group encoding scheme is employed in MOGA to express the mapping of physical nodes and virtual machines (VMs). Fitness function is designed based on the stability and migration overhead of group. Our simulation results demonstrate that, our MOGA is much more efficient than other algorithms for resource reallocation with good stability.

Mashup is emerging as a promising software development method for allowing software developers to compose existing Web APIs to create new or value-added composite Web services. However, the rapid growth in the number of available Mashup services makes it difficult for software developers to select a suitable Mashup service to satisfy their requirements. Even though clustering based Mashup discovery technique shows a promise of improving the quality of Mashup service discovery, Mashup service clustering with high accuracy for discovery of Mashup services is still a challenge problem. In this paper, we propose a novel Mashup service clustering method for Mashup service discovery with high accuracy by exploiting LDA topic model built from multiple data sources. It enables to infer topic probability distribution of Mashup services, which serves as a basis of computation of similarity of Mashup services. K-means and Agnes algorithm are used to perform Mashup service clustering in terms of their similarities. Compared with other service clustering approaches, experimental results show that our approach achieves significant improvement in terms of precision, recall and F-measure rate, which will improve Mashup service discovery.

Execution replay of virtual machines is a useful method for debugging applications in the cloud computing environment. The traditional methods to reproduce a bug is recording every details during the system runtime. However, these methods will incur much overhead and affect the system performance, especially in a multicore processor system. In this paper, we present a virtualization-based execution replay method consisting of three steps. First, we only record some necessary events in the runtime and take a memory checkpoint in a regular interval. Second, we search for execution paths between every two adjacent checkpoints. Third, we reproduce the bugs according to these paths. We can decrease the logging overhead in the runtime by searching instead of logging. We have implemented the method and evaluate it on Xen. The experimental results demonstrate that our method can reduce the runtime overhead by 30 % effectively.

Service composition is a widely-used approach in the development of applications. However, well-designed service composition approaches always lacks the consideration of execution environment, and the approach designed for application execution is usually incomplete and lacking necessary business consideration. In order to improve the comprehensiveness covered both design and execution stages, a service composition approach based on process mining is proposed. First, a meta-model is designed to connect the information of execution environment and business requirement. Next, the scene model based on this meta-model is generated by leveraging process mining. Then the scene model is applied to do service composition, including service selection from the Service Registry. After that, BPEL instance is converted based on aggregated scene information so as to enable application execution. Finally, a cloud-based logistics platform is implemented to verify the approach, and the result shows that the approach has high requirement accuracy and execution effectiveness.

Existing triple-failure-tolerant codes assume that failures are independent and instantaneous. Such assumptions overlook the underlying mechanism of multi-failure occurrences and ignored the effect of reconstruction window. These codes are not adapted to the occurrence pattern of failure in real-world applications. As a result, the third parity drive is almost idle as it set to handle the triple-failure scenario only with lower-level failure situations unattended. Furthermore, the problem of single failure rebuild deteriorates with the increasing disk capacity, and the system’s reliability will decrease with user experience impaired. Aiming at these problems, a fast reconstructable coding scheme extended from RAID-6 has been developed in this study. RAID-6Plus maintains a smaller reconstruction window by recoding the third parity drive. Existing codes provide absolute reliability for triple failures via full combinations. As a contrast, RAID-6Plus employs short combinations which are able to greatly reuse overlapped elements during reconstruction to remake the third parity drive. The short combinations shorten the reconstruction window of single failure, which avoids multi-failure overlapping in the reconstruction window. The capability of multi-failure degradation provides RAID-6Plus with (1) a better system performance comparing to RTP and STAR and (2) an enhanced reliability comparing to RAID-6.

The combination of social network and search engine is the trend of internet in coming years. Through introducing the widely utilized sharing and recommending mechanism in social network to search engine, this paper proposes a new searching ranking model. This model judges the quality of web pages and decide what extent do they meet users’ personalized need through analyzing the records of users’ social circle’s behavior of sharing and recommending. Then, it can make search engine provide users with personalized results sequences. Both the experiment and the theoretical analysis show the proposed model can automatically help users to select the high quality search results, and provide users with better personalized service.

Over the past decade, though web contents increase rapidly, the architecture of the web services still remains the same. As the growth of users, web servers must provide huge network bandwidth and computing power. This paper realizes the WebCDN system, which achieves a content distribution network by web users. Through the WebRTC and HTML5 technology, the web site needn’t require their users installing anything to use this service. This paper describes the design and implementation of the system in terms of both server and client side in detail. Through simulating the web resources and user behavior, the experiment result shows that the WebCDN system greatly reduces the network traffic with acceptable service latency.

Workflows can orchestrate multiple applications that need resources to execute. The cloud computing has emerged as an on-demand resource provisioning paradigm, which can support workflow execution. In recent years, Amazon offers a new service option, i.e., EC2 spot instances, whose price is on average more than 75 % lower than the one of on-demand instances. Therefore, we can make use of spot instances to execute workflows in a cost-efficient way. However, the spot instances is cut off when their price increases and exceeds the customer’s bid, which will make the task failed and the execution time becomes unpredictable. We propose a deadline driven scheduling (DDS) algorithm which is able to use both on-demand and spot instances to reduce the cost while the deadline of workflows can also be guaranteed with a high probability. Especially, we use an attribute, called global weight, to represent the interdependency relations of tasks and schedule the tasks whose interdependent tasks need longer time first to reduce the whole execution time. The experimental results demonstrate that DDS algorithm is effective in reducing cost while satisfying the deadline constraints of workflows.

A Wireless Sensor Network (WSN) is mission dependent network, deployed in an interesting area in order to collect data about a relevant observable environmental phenomenon and send them to end user through a base station. Due to their potential promising development, WSN increasingly attract researcher’s attention in order to ensure them the expected maturity of widespread deployment. However, many obstacles inherent to intrinsic sensor node characteristics may prevent achieving this goal. So, energy depleting is the most important hindering since node initial energy budget is poor. In this paper, we propose new hierarchical routing protocol sensitive to energy consumption and based on nodes duty-cycle management. This protocol improves WSN life time duration and data packets loss rate. The proposal was integrated to the well know LEACH protocol to enhance its performance. Simulation results via NS2 simulator showed that the proposal is convincing and outperforms the classical LEACH capabilities.

The research of life pattern is a hot topic in the field of LBSN (Location Based Social Network). Periodic behavior is also a life pattern. In view of multiple periodic behaviors existed in time series, an algorithm which can mine all periods in time series is proposed in this paper. In view of periodic behaviors always occurred at the same time interval and the random access of matrix’s characteristic, the algorithm creates a suspected periodic matrix which can store all suspected periods. By judging the validity of a suspected period in the matrix, the true periods can be mined accurately. Updating the suspected periodic matrix dynamically can reduce executing time.

Since its release in 2013, WeChat payment service has gradually become one of the most popular mobile payment services in China. Different from other mobile payment platforms, WeChat payment bundles with the most popular social network service in China, WeChat. It is then becoming interesting to investigate the reason beneath its popularity by combination of social network and mobile payment. In this research, we applied the technology acceptance model to predict the acceptability of WeChat payment and to identify the variables which attribute to the popularity of WeChat payment. Besides the primary explanatory variables of TAM, the proposed framework is further extended to include the constructs of Social Interaction, Trust, Perceived Enjoyment and Use Context. Online survey has been collected by respondents chosen randomly among users of WeChat payment. The results have shown that the proposed model is able to explain the variance in user’s behaviour intention to use WeChat payment services. We hope this study can provide insights to understand the adoption behaviour of social aware mobile payment and service and help further improve their services.

Elasticity is very important to the scale-out distributed systems running on today’s large-scale multi-tenant clouds, regardless public or private. An elastic distributed data processing system must have the capability of: (1) dynamically balancing the computing load among workers due to their performance heterogeneity and dynamicity; (2) fast recovering the lost memory state of failure workers with acceptable overheads during the regular execution.

Unfortunately, we found that the design of the state-of-the-art distributed graph computing system only works well in small sized dedicated clusters. We implement a distributed graph computing prototype, X-Graph, and demonstrate the capabilities of being elastic in three ways. First, we present menger, a novel two-level graph partition framework, which further splits one worker-level partition into several sub-partitions as the basic migration units, and each has the “migration affinity” with one of the other workers. Second, we implement a dynamical load balancer based on menger, which prefers the worker that has the affinity of the sub-partition to be migrated as the destination, and completely avoids the costly sophistical graph re-partitioning algorithms. Third, we implement a differentiated replication frame-work, which supports parallel recovery for lost partitions just like general-purpose dataflow systems.