Service-Oriented Computing

By bridging the physical and the virtual worlds, the Internet of Things (IoT) impacts a multitude of application domains, among which smart cities, smart factories, resource management, intelligent transportation, health and well-being to name a few. However, leveraging the IoT within software applications raises tremendous challenges from the networking up to the application layers, in particular due to the ultra-large scale, the extreme heterogeneity and the dynamics of the IoT. This paper more specifically explores how the service-oriented architecture paradigm may be revisited to address challenges posed by the IoT for the development of distributed applications. Drawing from our past and ongoing work within the MiMove team at Inria Paris, the paper discusses the evolution of the supporting middleware solutions spanning the introduction of: probabilistic protocols to face scale, cross-paradigm interactions to face heterogeneity, and streaming-based interactions to support the inherent sensing functionality brought in by the IoT.

Shared ledger technologies, as exemplified by Blockchain, provide a new framework for supporting business collaborations that is based on having a high-reliability, shared, trusted, privacy-preserving, nonrepudiable data repository that includes programmable logic in the form of “smart contracts”. The framework has the potential to dramatically transform business collaboration across numerous industry sectors, including finance, supply chain, food production, pharmaceuticals, and healthcare. Widespread adoption of this technology will be accelerated by the development of business-level languages for specifying smart contracts. This paper proposes that data-aware business processes, and in particular the Business Artifact paradigm, can provide a robust basis for a shared ledger Business Collaboration Language (BCL). The fundamental rationale for adopting data-aware processes is that shared ledgers focus on both data and process in equal measure. The paper examines potential advantages of the artifact-based approach from two perspectives: conceptual modeling, and opportunities for formal reasoning (verification). Broad research challenges for the development, understanding, and usage of a shared ledger BCL are highlighted.

In recent years there has been considerable interest in business process redesign. A process model may be redesigned by combining various tasks and services according to best practices so as to satisfy predefined business rules and constraints to achieve a specific purpose. This purpose may be stated in terms of functional goals (such as desired or acceptable process behavior) and non-functional goals like cost, time and quality of service. There are many ways to redesign a process instance by applying improvements such as: making a task optional, replacing a task by another faster task (or service), task postponement, task combination, task splitting, task restructuring, etc. Given many such alternatives, there is no systematic way of evaluating their costs and benefits, and the tradeoffs among them. We describe a novel approach based on a formal model to optimize the “benefits” or net effects of a redesign with respect to a baseline design and show how it can be used to evaluate and compare alternative models at both design and run time.

Economy of scale is a key driver behind the Cloud based adoption of a business process. Typically, the management of business process variants focuses on design variants, which permit (ideally small) variations in design (and hence, functionality) for achieving the same (functional) goal, under different functional constraints (such as the compliance obligations that have to be met in different jurisdictions). Little attention has been paid to: (a) variations in process design driven by non-functional considerations (e.g., performance, reliability and cost of operation) and (b) variations in process provisioning in Cloud. This paper seeks to develop means for identifying the correlation between both design and provisioning alternatives and the QoS of business processes deployed in the Cloud. Additionally, we explore the role of the context in determining the performance of a process. We use a set of data mining techniques (specifically decision tree learning, support vector machine and the k-nearest neighbour technique) to mine insights about these correlations. Proposed approaches are evaluated using a synthetic dataset as well as a real dataset.

Business processes are composed mainly of activities and events. The latter has gained much focus recently which has resulted in the drift towards Event-Driven Business Process Management (EDBPM). Events are used in both monitoring and controlling the execution of business processes. They are considered to be instantaneous and their content cannot be modified after they occur. However, this is not always the case in the real world. An event’s content can be modified at runtime under circumstances such as: earlier event information containing errors, or new information being obtained about the event. In such cases, the content modification for that event must be taken into consideration in the execution of the process. Additionally, the modified event’s content may affect other events within the process resulting in altering the content of those events as well. Therefore, it is important to determine the propagation of event content modification in an event network within a business process. In this work, we determine the types of event content modifications that can occur within processes, how content modification of one event affects other events within the process, and how the modification affects the process as a whole.

Process discovery is one of the main branches of process mining that allows the user to build a process model representing the process behavior as recorded in the logs. Standard process discovery techniques produce as output a procedural process model (e.g., a Petri net). Recently, several approaches have been developed to derive declarative process models from logs and have been proven to be more suitable to analyze processes working in environments that are less stable and predictable. However, a large part of these techniques are focused on the analysis of the control flow perspective of a business process. Therefore, one of the challenges still open in this field is the development of techniques for the analysis of business processes also from other perspectives, like data, time, and resources. In this paper, we present a full-fledged approach for the discovery of multi-perspective declarative process models from event logs that allows the user to discover declarative models taking into consideration all the information an event log can provide. The approach has been implemented and experimented in real-life case studies.

In the literature, hierarchical dimensions for procedural process models have been widely investigated as they provide different ways to relate, organize and classify models. Such a categorization is based on the dimensions of inheritance, behavioral equivalence, and modularization and can be used to better understand and modify models as well as handle their complexity. Unfortunately, in the context of declarative process models hierarchical dimensions have been sparsely investigated. This paper addresses such a research gap. More specifically, we study a formal semantics for the dimensions above and show how they naturally induce hierarchies on a declarative process language based on declare.

One of the key benefits of Cloud computing is elasticity, the ability of the system infrastructure to adapt to the workload changes by automatically adjusting the resources on-demand. Horizontal scaling refers to the method of adding or removing resources from the resource pool. As such it is appealing to enterprises who seek to migrate their legacy systems as it requires no application rewrite or refactoring. Vertical scaling approach offers a mechanism to maintain continuous performance while reducing resource cost through reconfiguration of the resource. The challenge is, however, in being able to automatically identify the right size of the target resource such as a VM or a container. Moreover, choice of scalability policies is not intuitive due to application complexity, topology and variability in system performance parameters that need to be considered.This paper presents a transformation model, FitScale, which provides scalability with minimum price of resources. The paper describes the framework that employs the application functional and operational properties to recommend the target sizing and scalability policies. We evaluate proposed approach in an on-premise and cloud environments, with a dataset of 2023 servers hosting 6737 applications. The experimental results show about 5 times cost reduction with minimum performance impact.

With the increasing scale of SaaS and the continuous growth in server failures, task scheduling problems become more intricate, and both scheduling quality and scheduling speed raise further concerns. In this paper, we first propose a virtualized and monitoring SaaS model with predictive maintenance to minimize the costs of fault tolerance. Then with the monitored and predicted available states of servers, we focus on dynamic real-time task scheduling in large-scale heterogeneous SaaS, targeting at jointly optimizing the long-term performance benefits and energy costs in order to improve scheduling quality. We formulate a dynamic programming problem, where both the state and action spaces are too large to be solved by simple iterations. To address these issues, we take advantage of Machine Learning theory, and put forward an approximate dynamic programming algorithm. We utilize value function approximation and candidate-heuristic method to separately solve state and action explosions. Thus, computation complexity is significantly reduced and scheduling speed is greatly enhanced. Finally, we conduct experiments with both random simulation data and Google cloud trace-logs. Qos evaluations and comparisons demonstrate that our approach is effective and efficient under bursty requests and high throughputs.

Cloud computing is currently the most popular model to offer and access computational resources and services. Many cloud providers use the REST architectural style (Representational State Transfer) for offering such computational resources. However, these cloud providers face challenges when designing and exposing REST APIs that are easy to handle by end-users and/or developers. Yet, they benefit from best practices to help them design understandable and reusable REST APIs.However, these best practices are scattered in the literature and they have not be studied systematically on real-world APIs. Consequently, we propose two contributions. In our first contribution, we survey the literature and compile a catalog of 73 best practices in the design of REST APIs making APIs more understandable and reusable. In our second contribution, we perform a study of three different and well-known REST APIs from three cloud providers to investigate how their APIs are offered and accessed. These cloud providers are Google Cloud Platform, OpenStack, and Open Cloud Computing Interface (OCCI). In particular, we evaluate the coverage of the features provided by the REST APIs of these cloud providers and their conformance with the best practices for REST APIs design.Our results show that Google Cloud follows 66 % (48/73), OpenStack follows 62 % (45/73), and OCCI 1.2 follows 56 % (41/73) of the best practices. Second, although these numbers are not necessarily high, partly because of the strict and precise specification of best practices, we showed that cloud APIs reach an acceptable level of maturity.

Integrating apps on mobile devices into applications running on other devices is usually difficult. For instance, using a messenger on a smartphone to share a text written on a desktop computer often ends up in a cumbersome solution to transfer the text, because many applications are not designed for such scenarios. In this paper, we present an approach enabling the integration of apps running on Android devices into applications running on other devices and even other platforms. This is achieved by specifying adapters for Android apps, which map their services to a platform-independent service interface. For this purpose, we have developed a domain-specific language to ease the specification of such mappings. Our approach is applicable without the need to modify the existing Android apps providing the service. We analyzed its feasibility by implementing our approach and by specifying mappings for several popular Android apps, e.g., phone book, camera, and file explorer.

The proliferation of cloud computing has enabled powerful virtualization capabilities and outsourcing strategies. Suitably, a vast variety of cloud resource configuration and management tools have emerged to meet this needs, whereby DevOps are empowered to design end-to-end and automated cloud management tasks that span across a selection of best-of-breed tools. However, inherent heterogeneities among resource description models and management capabilities of such tools pose fundamental limitations when managing complex and dynamic cloud resources. In this paper we thus propose the notion of “Domain-specific Models” – a higher-level model-driven approach for describing elementary and federated cloud resources as reusable knowledge artifacts over existing tools. We also propose a pluggable architecture to translate these artifacts into lower-level resource descriptions and management rules. This paper describes concepts, techniques and a prototypical implementation. Experiments on real-world federated cloud resources display significant improvements in productivity. As well as notably enhanced usability achieved by our approach in comparison to traditional techniques.

Open Cloud Computing Interface (OCCI) follows a set of guidelines (i.e. best practices) to create interoperable APIs over Cloud resources. In this paper, we identify a set of patterns that must be followed and anti-patterns that should be avoided to comply with the OCCI guidelines. To automatically detect (anti)patterns, we propose a Semantic-based approach, relying on SWRL (Semantic Web Rule Language) rules and in SQWRL (Semantic Query-Enhanced Web Rule Language) queries to describe the (anti)patterns symptoms. An evaluation, conducted on real world Cloud service APIs, shows the feasibility of the proposed approach by assessing their compliance to OCCI standard.

In customer-care service centers, upon remediation of customer issues, the human agents are expected to record their resolution summary in a clear, concise and understandable manner. These resolution summaries create a rich untapped source of unstructured information. In this work, we have addressed the problem of how to enable human agents to write better quality resolution text. This helps curate data artifacts which can reduce problem diagnosis time and create repeatable resolution recipes by a cognitive system. The problem is addressed through a two pronged approach: (i) On the fly automated scoring of the agent’s resolution summary and (ii) identifying concrete areas of improvement in the summary and offering appropriate recommendations. The model for automatic scoring is derived from a feature set that encodes all significant and relevant aspects of the domain and text. The model is trained using annotated data and achieves an accuracy of 88.2 % which is a significant improvement over naive method of text based classification (68.5 %).

In the dynamic service market, massive services and variations of their Quality of Services (QoS) and service contract make it difficult for customers to acquire the information of all the services comprehensively and timely. As a result, customers cannot raise accurate expectations. A customer has to choose services in terms of the incomplete information of the dynamic service market to achieve higher Satisfaction Degree (SD) as much as possible. Besides, because a customer’s preferences vary over time, his SD is also time-aware. Therefore, for service providers, to accurately recommend services to customers, it is necessary to sense the customer preferences varying against time and predict personalized customers’ satisfaction. To address this challenge, we propose a time-aware customer preference sensing and satisfaction prediction method based on customer’s service usage history and change history of services. Firstly, the customer satisfaction model on contract-based services is proposed to measure customers’ satisfaction for services. Then, we adopt the box-plot method and the frequency histogram to sense time-aware customer preferences. In addition, a time-aware personalized SD prediction algorithm called SDPred is presented to predict the missing values due to information asymmetry. Meanwhile, several experiments have been conducted based on a released data set, which verify the effectiveness of our methods. Besides, the impact of parameter settings in the SDPred algorithm is further studied, which provides more evidences to illustrate the superiority of our method.

Mobile ecosystem has penetrated into people’s daily life over these years and most web services are now using mobile application for service consumption. Permission system has been developed to protect the sensitive and valuable information stored in mobile. However, due to the complexity of permission framework, the permission over-privilege problem has become a serious problem bringing huge risk for the mobile ecosystem. Therefore, in this paper, we present a skewness-based framework for permission recommendation and risk evaluation, intending to facilitate the permission configuration and identify the risk applications. Specially, the topic model Latent Dirichlet Allocation is presented to build the mapping between app’s functionality and permission. Then a two-phase skewness-based filtering strategy is developed and combined with the collaborative filtering framework to remove the abnormal applications and permissions. Finally, the high risk permissions for each application are identified based on the difference between the malicious applications and popular applications. The experiments based on the Apps from Google Play shows that comparing with the state-of-the-art; our approach can effectively remove the abnormal applications and permissions, identify the unexpected and risk permissions, as well as generate the recommended permission configurations with better performance to reduce the permission over-privilege problem.

Developing IoT cloud platforms is very challenging, as IoT cloud platforms consist of a mix of cloud services and IoT elements, e.g., for sensor management, near-realtime events handling, and data analytics. Developers need several tools for deployment, control, governance and analytics actions to test and evaluate designs of software components and optimize the operation of different design configurations. In this paper, we describe requirements and our techniques on supporting the development and testing of IoT cloud platforms. We present our choices of tools and engineering actions that help the developer to design, test and evaluate IoT cloud platforms in multi-cloud environments.

Web service interfaces are considered as one of the critical components of a Service-Oriented Architecture (SOA) and they represent contracts between web service providers and clients (subscribers). These interfaces are frequently modified to meet new requirements. However, these changes in a web service interface typically affect the systems of its subscribers. Thus, it is important for subscribers to estimate the risk of using a specific service and to compare its evolution to other services offering the same features in order to reduce the effort of adapting their applications in the next releases. In addition, the prediction of interface changes may help web service providers to better manage available resources (e.g. programmers’ availability, hard deadlines, etc.) and efficiently schedule required maintenance activities to improve the quality. In this paper, we propose to use machine learning, based on Artificial Neuronal Networks, for the prediction of the evolution of Web services interface design. To this end, we collected training data from quality metrics of previous releases from 6 Web services. The validation of our prediction techniques shows that the predicted metrics value, such as number of operations, on the different releases of the 6 Web services were similar to the expected ones with a very low deviation rate. In addition, most of the quality issues of the studied Web service interfaces were accurately predicted, for the next releases, with an average precision and recall higher than 82 %. The survey conducted with active developers also shows the relevance of prediction technique for both service providers and subscribers.

The provision of IT solutions over electronic marketplaces became prominent in recent years. We call such marketplaces IT service markets. IT service markets have some core architectural building blocks that impact the quality attributes of these markets. However, these building blocks and their impacts are not well-known. Thus, design choices for IT service markets have been made ad-hoc until now. Furthermore, only single aspects of such markets have been investigated until now, but a comprehensive view is missing. In this paper, we identify common features and their interrelations on the basis of a systematic literature review of 60 publications using grounded theory. This knowledge provides an empirical evidence on the interdisciplinary design choices of IT service markets and it serves as a basis to support market providers and developers to integrate market features. Thereby, we make a first step towards the creation of a reference model for IT service markets that provides a holistic integrated view that can be used to create and maintain successful markets in the future.

We propose a new qualitative economic model based optimization approach to compose an optimal set of infrastructure service requests over a long-term period. The economic model is represented as a temporal CP-Net to capture the provider’s dynamic business strategies in qualitative service provisions. The multidimensional qualitative preferences are indexed in a k-d tree to compute the preference ranking of a set of incoming requests. We propose a heuristic based sequential optimization process to select the most preferred composition without the knowledge of historical request patterns. Experimental results prove the feasibility of the proposed approach.

The evaluation of dynamic performance of cloud services relies on continual assessments from cloud users, e.g., ordinary consumers and testing parties. In order to elicit continual and truthful assessments, an effective incentive mechanism in cloud environments should allow users to provide uncertain assessments when they are not sure about the real performance of cloud services, e.g., when users do not access cloud services on time, rather than providing untruthful or arbitrary assessments. Different from all prior works, we propose a novel uncertain assessment compatible incentive mechanism. Under this mechanism, a user not only has sufficient incentives to continually provide truthful assessments, but also would prefer providing uncertain assessments over untruthful or arbitrary assessments since uncertain assessments can bring more benefits than untruthful or arbitrary assessments. We theoretically analyze the proposed incentive mechanism and evaluate it through simulations under different circumstances. The theoretical analysis demonstrates the effectiveness of our approach. Moreover, the experimental results based on simulations strongly support the results from the theoretical analysis.

Successful Web services must evolve to remain relevant (e.g. requirements update, bugs fix, etc.), but this process of evolution increases complexity and can cause the Web service interface design to decay and lead to significantly reduced usability and popularity of the services. Maintaining a high level of design quality is extremely expensive due to monetary and time pressures that force programmers to neglect improving the quality of their interfaces. A more fundamental reason is that there is little support to automatically identify design defects at the Web service interface level and reduce the high calibration effort to determine manually the threshold value for each quality metric to identify design defects. In this paper, we propose to treat the generation of interface design defects detection rules as a bi-level optimization problem. To this end, the upper level problem generates a set of detection rules, as combination of quality metrics, which maximizes the coverage of a base of defects examples extracted from several Web services and artificial defects generated by the lower level. The lower level maximizes the number of generated artificial defects that cannot be detected by the rules produced by the upper level. The statistical analysis of our experiments over 30 runs on a benchmark of 415 Web services shows that 8 types of Web service defects were detected with an average of more than 93 % of precision and 98 % recall. The results confirm the outperformance of our bi-level proposal compared to state-of-art Web service design defects detection techniques and the survey performed by potential users and programmers also shows the relevance of the detected defects.

In the scenario of service recommendation, there are multiple object types (e.g. services, mashups, categories, contents and providers) and rich relationships among these objects, which naturally constitute a heterogeneous information network (HIN). In this paper, we propose to recommend services for mashup creation by exploiting different types of relationships in service related HIN. Specifically, we first introduce meta-path based measure for similarity estimation between mashups along different types of paths in HIN. We then design a recommendation model based on collaborative filtering and meta-path based similarities, and employ Bayesian ranking based optimization algorithm for model learning. Comprehensive experiments based on real data demonstrate the effectiveness of the HIN based service recommendation approach.

With the recognition of the significance of OSNs (Online Social Networks) in the recommendation of services in e-commerce, there are more and more e-commerce platform being combined with OSNs, forming social e-commerce, where a participant could recommend a product to his/her friends based on the participant’s corresponding purchasing experience. For example, at Epinions, a buyer could share product reviews with his/her friends. In such platforms, a buyer providing lots of high quality reviews is very likely to influence many potential buyers’ purchase behaviours. Such a buyer is believed to have strong social influence. However, dishonest participants in OSNs can deceive the existing social influence evaluation models, by mounting attacks, such as Constant (Dishonest advisors constantly provide unfairly positive/negative ratings to sellers.) and Camouflage (Dishonest advisors camouflage themselves as honest advisors by providing fair ratings to build up their trustworthiness first and then gives unfair ratings.), to obtain fake strong social influence. Therefore, it is crucial to devise a robust social influence evaluation model that can defend against attacks and deliver more accurate social influence evaluation results. In this paper, we propose a novel robust Trust-Aware Social Influencer Finding, TrustINF, method that considers the evolutionary trust relationship and the variations of historical social influences of participants, which can help deliver more accurate social influence evaluation results in social e-commerce. Our experiments conducted on four real social network datasets validate the effectiveness and robustness of our proposed method, which is greatly superior to the state-of-the-art method.

In a service system comprising of knowledge intensive tasks, a pull-based allocation strategy (where knowledge workers decide on tasks to commit to, as opposed to having these commitments decided for them) can often be quite effective. Such a scenario is characterized by different types of tasks and workers with varying efficiencies. As workers and tasks change with time, a key challenge faced by knowledge workers is in deciding the most suitable tasks to commit to. Organizational roles of workers provide them the privilege of working on the tasks that the role is authorized to perform, but the suitability of a worker to perform a task varies because workers could have varying operational performance on different types of tasks. Past allocations, when correlated with execution histories annotated with quality of service (or performance) measures, can provide insights on the suitability of a task for a worker. It has been recognized that the effectiveness of a resource in performing a task often depends on the context in which the task is executed. In this work, we present a context-aware collaborative filtering recommender system that predicts a worker’s suitability for a task, in different contexts or situations. The context-aware recommender uses information on the performance of similar resources in similar contexts to predict a resource’s suitability for a task. Experiments performed on real-world execution logs demonstrate the effectiveness of the proposed approach.

This paper deals with the problem of web service recommendation. We propose a new content-based recommendation system. Its originality comes from the combination of probabilistic topic models and pattern mining to capture the maximal common semantic of sets of services. We define the notion of semantic patterns which are maximal frequent itemsets of topics. In the off-line process, the computation of these patterns is performed by using frequent concept lattices in order to find also the sets of services associated to the semantic patterns. These sets of services are then used to recommend services in the on-line process. We compare the results of the proposed system in terms of precision and normalized discounted cumulative gain with Apache Lucene and SAWSDL-MX2 Matchmaker on real-world data. Our proposition outperforms these two systems.

REST APIs together with JSON are commonly used by modern web applications to export their services. However, these services are usually reachable in a pull mode which is not suitable for accessing changing data. Turning a service from a pull to a push mode is therefore frequently asked by web developers that want to get notified of changes. Converting a pull API into a push one obviously requires to make periodical calls to the API but also to create a patch between each successive version of the data. The latter is the most difficult part and this is where existing solutions have some imperfections. To face this issue, we present a new patch algorithm supporting move and copy change operations. Our evaluation done with real industrial data shows that our algorithm creates small patches compared with other libraries, and creates them faster.

This work proposes an approach for User Interface (UI) derivation based on a role-enriched Business Process (BP) model with the capability to describe the details of the control flow and data operations in a BP. For each user role, data relationships are extracted according to the identified control flow patterns and data operation patterns. A set of mandatory and recommended UI derivation rules are specified as the cornerstones to derive the UI logic from a BP. The algorithm for UI derivation is provided. This UI derivation approach provides the basis for UI development and maintenance.

Case Management (CM) is a BPM technology for supporting flexible services orchestration. CM approaches like CMMN, an OMG standard, and GSM, one of CMMN’s core influences, use Event-Condition-Action rules, which can be inconsistent due to cyclic inter-dependencies between the rules; repairing such an inconsistent case management schema is difficult. To avoid the problem of inconsistencies altogether, we provide a technique for automatically deriving consistent GSM case management schemas from higher-level business policies defined as DCR graphs, an alternative CM approach. Concretely, we define a behaviour-preserving mapping that (1) removes the burden from the modeller of GSM schemas to prove consistency and define the ordering of rules, (2) provides high-level patterns for modelling GSM schemas, and (3) gives a way to define a notion of progress (liveness) and acceptance for GSM instances. The mapping is illustrated by a running example of a mortgage loan application; and a prototype implementation available at http://dcr.itu.dk/icsoc16.

Configurable process models are recently gaining momentum as a basis for process design by reuse. Such models are designed in a generic manner to group common and variable parts of similar processes. Since these processes are usually large and complex, their configuration becomes manifestly a difficult task. This is why, an increasing attention is being paid to help achieving the process models configuration in a correct and domain-compliant manner. In this work, we propose an Event-B based formal approach that guides the process analyst to easily derive correct process variants while considering business domain constraints provided by configuration guidelines. To show the effectiveness of our approach, we conduct experiments on a case study.

Service-oriented social networks gain increasing popularity among a huge user base in recent years. In social networks, trust prediction is significant for recommendations of high-quality service providers as well as in many other applications. In the literature, trust prediction problem can be solved by several strategies, such as matrix factorization, trust propagation, and K-NN search, etc. However, most of the existing works have not considered the possible complementarity among these mainstream strategies to optimize their effectiveness and efficiency. In this paper, we propose a novel trust prediction approach named iSim: an integrated similarity based collaborative filtering approach leveraging on user similarity, which integrates three kinds of factors to measure user similarity, including vector space similarity, matrix factorization, and propagated trust. This paper is the first work in the literature employing matrix factorization and propagated trust in the study of similarity. Additionally, we use several methods like adding inverted index to reduce the time complexity of iSim, and provide its theoretical time bound. Finally, the extensive experiments with real-world dataset show that iSim achieves great improvement for both efficiency and effectiveness over the state-of-the-art approaches.

With the increasing number of Web services, the personalized recommendation of Web services has become more and more important. Fortunately, the social network popularity nowadays brings a good alternative for social recommendation to avoid the data sparsity problem that is not treated very well in the collaborative filtering approach. Since the social network provides a big data about the users, the trust concept has become necessary to filter this abundance and to foster the successful interactions between the users. In this paper, we firstly propose a trusted friend detection mechanism in a social network. The dynamic of the users’ interactions over time and the similarity of their interests have been considered. Secondly, we propose a Web service social recommendation mechanism which considers the expertise of the trusted friends according to their past invocation histories and the active user’s query. The experiments of each mechanism produced satisfactory results.

Business process model discovery (BPMD) is one of the most important research topics in the business process mining area. Many outstanding BPMD algorithms which perform well in most cases have been developed in the last years. As one of the most widely used BPMD algorithms, the Heuristics Miner meets great challenges while dealing with event logs that contain complex behaviours. As a result, process models with low fitness values might be obtained. In this paper, we propose a new technique that is able to locate the process behaviours recorded in an event log which cannot be expressed by the Heuristics Miner and then transform them into expressible behaviours so that a high-fitness model can be built.

In the field of business process management, the interaction between business actors or services are modeled via business process choreographies. However, enforcing or implementing business process choreographies is a challenge particularly related to the choreography’s exclusive gateways, which are used to model shared decisions among business actors. Since there is no central locus of control, participants may interpret the data relevant for decision making differently. To tackle this problem, this paper offers a solution by delegating the decision making to a decision service. This service is based on the recently published Decision Model and Notation standard and is provided to the choreography participants via a REST interface. The RESTful decision service assures a correct implementation of choreographies’ exclusive gateways and provides a blueprint for RESTful services that offer decision-making solutions based on the DMN standard.

To help the independent-living elders or patients nursed in a single isolated ward, we propose a proof-of-concept prototype named WiCare, a non-intrusive and device-free healthcare service system based on ubiquitous WiFi signals. It extracts Channel State Information (CSI) from the physical layer and detects the unique variations of CSI values caused by human activities. We implement WiCare on two laptops equipped with the commercial 802.11n network interface cards. Two potential application scenarios are considered: a living room and a bedroom. The results demonstrate that the proposed scheme achieves overall recognition accuracies of 92.3 % in living room and 87.6 % in bedroom with low false positive rates. Moreover, WiCare can send alarm messages when the server recognizes the occurrences of emergency activities, which assist the users in getting help as quickly as possible.

A service mining framework is proposed that enables discovering interesting relationships in Internet of Things services bottom-up. The service relationships are modeled based on spatial-temporal aspects, environment, people, and operation. An ontology-based service model is proposed to describe services. We present a set of metrics to evaluate the interestingness of discovered service relationships. Analytical and simulation results are presented to show the effectiveness of the proposed evaluation measures.

It is crucial to ensure correct process model executions. However, existing process testing approaches struggle with the verification of concurrent resource access patters that can lead to concurrency faults, such as, deadlocks or data corruption during runtime. Thus, we provide a concurrency verification approach that exploits recorded executions to verify the most frequently occurring concurrent resource access patterns with low test execution time. A prototypical implementation along with real life and artificial process execution logs is utilized for an evaluation.

Complex Event Processing (CEP) is becoming more and more popular in service-oriented practices, especially to monitor the behaviour of continuous tasks within manual business processes, such as in logistics. The inference mechanisms of CEP engines are completely guided by rules, which are specified manually by domain experts. We argue that this user-based rule specification is a limiting factor that complicates the integration of CEP within the realm of Business Process Management (BPM) in a seamless way. Therefore, we present autoCEP as a two-phase data mining-based approach that automatically learns CEP rules from historical traces. In the first phase, complex temporal patterns are learned using early classification on time series techniques, then these patterns are algorithmically transformed into CEP rules in the second phase. Satisfactory results from evaluations on real data demonstrate the effectiveness of our framework.

In this paper, we investigate optimal pricing models for profit maximization from the perspective of cloud providers in the presence of multiple classes of IaaS (Infrastructure as a Service) services. We propose an iterative model in which a cloud provider iteratively posts updated prices for the multiple classes of IaaS instances to users until reaching convergence that maximizes its profit. During this process, any interested user can determine the optimal class of IaaS instances and the optimal quantity to buy according to its own private utility function. In particular, we propose two algorithms to implement the iterative pricing process: a Genetic based near-optimal algorithm, and a hill climbing based cost-effective algorithm. The experimental results show that our iterative pricing algorithms can achieve advanced profitability in pricing multiclass IaaS instances in cloud environments.

Nowadays databases have become the backbone of cloud-based applications. Cloud-based applications are used in about every industry today. Despite their popularity and wide adoption, little is still known about the energy footprint of these applications and, in particular, of their databases. Yet, reducing the energy consumption of applications is a major objective for society and will continue to be so in the near to far future. In this paper, we study the energy consumption of three databases used by cloud-based applications: MySQL, PostgreSQL, and MongoDB, through a series of experiments with three cloud-based applications (a RESTful multi-threaded application, DVD Store, and JPetStore). We also study the impact of cloud patterns on the energy consumption because databases in cloud-based applications are often implemented in conjunction with patterns. We measure the energy consumption using the Power-API tool to keep track of the energy consumed at the process-level by the variants of the cloud-based applications. We report that the choice of the databases can reduce the energy consumption of a cloud-based application regardless of the cloud patterns that are implemented.

Data analytic applications and services are becoming increasingly important, especially in this age of Big Data. QoS properties such as latency, reliability, response time of such services can vary based on the attributes (e.g., size, number of dimensions, data types) of the dataset being processed. The existing QoS-based web service selection methods are not adequate for ranking this type of services because they do not consider these dataset attributes. In this paper, we have proposed a method to predict the QoS values for data analytic services based on the attributes of the dataset by incorporating a meta-learning approach. We could then rank these services according to the predicted QoS values. Our experiment results prove the effectiveness of this approach and the improvement in service ranking when compared with the traditional service selection approach.

The widespread distribution of smart mobile devices offers promising perspectives for the timely collection of huge amounts of data. When realizing sophisticated mobile data collection applications, numerous technical issues arise. For example, as many real-world projects require the support of different mobile operating systems, platform-specific peculiarities must be properly handled. Existing approaches often rely on specifically tailored mobile applications. As a drawback, changes to the data collection procedure result in costly code adaptations. To remedy this drawback, a model-driven approach is proposed, enabling end-users (i.e., domain experts) to create mobile data collection applications themselves. This model relies on complex questionnaires called instruments. An instrument not only contains all information about the data to be collected, but additionally comprises information on how it shall be processed on different mobile operating systems. For this purpose, we developed an advanced mobile (kernel) service being capable of processing sophisticated instruments on various platforms. This paper discusses fundamental kernel requirements and introduces the developed architecture. Altogether, the mobile service allows for the effective use of smart mobile devices in data collection application scenarios (e.g., clinical trials).

With the increasing popularity of web service technology, many users turn to look for appropriate web services to further build their complex business applications. As an effective manner for service discovery, service recommendation technique is gaining ever-increasing attention, e.g., Collaborative Filtering (i.e., CF) recommendation. Generally, the traditional CF recommendation (e.g., user-based CF, item-based CF or hybrid CF) can achieve good recommendation results. However, due to the inherent sparsity of user-service rating data, it is possible that the target user has no similar friends and the services preferred by target user own no similar services. In this exceptional situation, traditional CF recommendation approaches cannot deliver an accurate recommendation result. In view of this shortcoming, a novel Social Balance Theory (i.e., SBT)-based service recommendation approach, i.e., Rec _SBT is introduced in this paper, to help improve the recommendation performance. Finally, through a set of simulation experiments deployed on MovieLens-1M dataset, we further validate the feasibility of Rec _SBT in terms of recommendation accuracy and recall.

With a huge amount of APIs on the Internet, understanding users’ complex needs and preferences for APIs becomes an important task. In this paper, we aim to uncover users’ implicit needs for APIs and recommend suitable APIs for users. Specifically, first different similarity scores between APIs are computed according to heterogeneous functional aspects of APIs. Next, users’ preferences for APIs is combined with similarities of APIs measured with different functional aspects, and matrix factorization technique is used to learn the latent representation of users and APIs for each functional aspect. Then we use a personalized weight learning approach to combine the latent factors of different aspects to get the predicted preferences of users for APIs.

We make an empirical study on App evolution, especially on inter-App interface evolution and inner-App feature evolution, both from externally observable exhibitions of Apps. Interfaces are extracted from .apk files, and statistical methods are used to discover underlying patterns of interface evolution. Furthermore, potential trend on how interface evolutions of Apps result in the evolution of Global inter-App Network (GAN) is observed. Latent Dirichlet Allocation (LDA) is applied to extract updated features (“topics”) from “What’s New” of each version to explore the underlying patterns of feature evolution. A set of significant phenomena have been observed from the empirical study.

With the number of services growing explosively, it has been a serious problem selecting appropriate services for mashup development. In this paper, we come up with a Multi-level Relational Network (MRN) based approach for service recommendation in mashup development, which captures deep relationships among services on top of latent topic, tag and service network. Specifically, by modeling the correlation among services, representing it as a Quadratic Knapsack Problem and solving it using Branch and Bound algorithm, we are able to recommend a package of services, which are complementary and possible to be used together in a mashup. Experiments on a realistic mashup data set have shown its effectiveness.

As a powerful computing paradigm for constructing complex distributed applications, service composition is usually addressed as a planning problem since the goal is to optimize a path for combining services to satisfy special requirements. Some planning methods assume that the state of running environment can be fully observed and monitored. However, the dynamic internet environment and opaque internal status, such as QoS attributes and invoking results, make the assumption too strict and not generally applicable. In this paper, we introduce a Partially Observable Markov Decision Process (POMDP) to model a service composition, which views the environment as partially observable and generates a policy with incomplete information. The partial observability relaxes the previous assumption and can handle the difficulties occurring in a dynamic and unpredictable environment. Based on this model, we propose a reinforcement learning algorithm to compute the optimal strategy. We conduct a series of experiments to verify the proposed algorithm, and compare it the comparison with other two algorithms. The results show the correctness and effectiveness of our algorithm.

This paper deals with semantic composition of ubiquitous computing (Ubicomp) services in multi-domain heterogeneous environments. A new semantic framework enabling the specification of multi-domain composite services and proving their correctness is proposed. The proposed framework reduces the gap between safety critical systems including ubiquitous services and best effort engineering practices. It consists of an extensible Semantic Conceptual Model (SCM) for Ambient Intelligence systems and a composition formal system based on the Basic Constructive Description Logics $$\mathcal {BCDL}_0$$BCDL0. The soundness property proof of the proposed formal system, as well as, the services composition correctness proofs are demonstrated along with the interactive theorem prover Isabelle/HOL.

Service crowdsourcing follows typical social collaboration processes with stochastic and dynamic characteristics. In this paper, the “bug-fix” social collaboration on GitHub is used as a case scenario of crowdsourcing, and 53,475 issues in 10 OSS projects are collected to conduct an empirical study on features and patterns of service crowdsourcing. Seven collaboration features (CFs) are proposed to delineate social characteristics of crowdsourcing. In terms of these CFs, social collaboration processes are clustered and results show that these features have significant distinguishability. An extended Generalized Sequential Pattern (GSP) algorithm is put forward to identify two types of collaboration patterns called participant-oriented pattern (PP) and role-oriented pattern (RP), and the richness and individualized degree of collaboration patterns in different OSS projects are analyzed and compared.

Web services partake in various types of interactions during their lifetime such as recommendation, substitution, and composition, hence giving rise to social behaviors. In this paper, we propose a social-aware approach for service communities. Communities are built around socially active services called leaders. The remaining services, called followers, use past interactions to elect their leaders and join communities. We introduce a clustering algorithm for multi-relation networks and define heuristics to identify community leaders and followers. We also define a new metric, called interoperability degree, to determine the degree to which members of a community are likely to socially interact. We conduct experiments to illustrate that leveraging social behaviors may help clump together services that are suited to interoperate.

In the background of actual production demands, we develop data services to solve the problem of “information isolated island” in the tire production for achieving the unified management for data from diverse production systems. Based on the data services, the management system for tire production is designed. The system uses the decision tree algorithm with data fitting and data screening technologies to analyze the data from the whole production process and realize the forecast of product quality and defects analysis. The system has been applied to the production by Shandong Linglong Tire Co., Ltd. The practice has proved that our data services and system not only improve the tire pass rate and production efficiency, but also help enterprises to achieve the efficient management of production. In addition, we apply the service to the actual manufacturing industry, which plays a positive role in the promotion and improvement of service application.

Technical support agents working in the IT support services field resolve IT problems. They are often faced with the daunting task of identifying the correct solution document through a search system from large corpora of IT support documents. Based on the observation that system logs may contain critical information for identifying the root cause of IT problems, we explore the idea of automatic query expansion by using system logs as a bridge to link queries with the most relevant documents. Given the original query from a user such as a technical support agent, an intermediate query is first formed by adding key terms extracted from system logs using domain-specific rules. Based on topic models, further key terms are selected from corpora of IT support documents, which are combined with the intermediate query to form the final query. Our experimental results show that expanding queries using system logs together with topic models yields better performance in retrieving relevant IT support documents than using topic models only.

In order for an Information Technology (IT) service provider to respond to a client’s request for proposals of a complex IT services deal, they need to prepare a solution and enter a competitive bidding process. A critical factor in this solution is the pricing of various services in the deal. The traditional way of pricing such deals has been the so-called bottom-up approach, in which all services are priced from the lowest level up to the highest one. A previously proposed more efficient approach and its enhancement aimed at automating the pricing by data mining historical and market deals. However, when mining such deals, some of the services of the deal to be priced might not exist in them. In this paper, we propose a method that deals with this issue of incomplete data via modeling the problem as a machine learning recommender system. We embed our system in the previously developed method and statistically show that doing so could yield significantly more accurate results. In addition, using our method provides a complete set of historical data that can be used to provide various analytics and insights to the business.

The Internet of Things (IoT) embodies the evolution from systems that link digital documents to systems that relate digital information with real-world physical items. It provides the infrastructure to transparently and seamlessly glue heterogeneous resources and services together by accessing sensors and actuators over the Internet. By connecting the physical world and the digital world, IoT creates numerous novel opportunities for many applications such as smart homes, smart cities, and industrial automation. However, on the other hand, IoT poses challenges to business process development, which unfortunately, have rarely been studied in the literature. In this paper, we present WITSCare, a research prototype of Web-based Internet of Things Smart home systems, with the aims of helping older people live in their own homes independently longer and safer. WITSCare exploits the heterogeneous contextual information (e.g., daily activities) captured and learned from IoT devices, then exposes the contexts as services to be integrated into personalized care management processes, and to support automatic and better decision making in an effective and user-friendly manner. The practical experiences gained from this project provide insights on developing real-world IoT applications.

In recent years several description tools and formats have been introduced for describing REST Web APIs both in human and machine readable formats. Although these descriptions provide functional information about the APIs (e.g. HTTP methods, URIs, model schema, etc.), the information that qualifies the properties of APIs (e.g. classification of input arguments and response data) is missing. We envisage that providing a complete set of information to the users will facilitate the composition of APIs to fulfil users’ specific needs.This paper analyses the current state of the art in Web API Descriptions and Semantic Annotations to show that although there are solutions with semantic capabilities, most of them fails to add semantic annotations automatically or semi-automatically. Moreover, advanced technical skills are needed to manage semantics and compose different Web APIs, which reduce the number of potential users of such solutions. The goal is to enhance actual API descriptions by creating a simple description format to annotate properties at semantic level to support semi-automatic composition. To achieve this goal, we propose an extension of the Open API Initiative (OAI) specification to create comprehensive descriptions. The approach focuses on the emerging concept of API Profiling to add descriptive information of data semantics by addressing Dublin Core Application Profile (DCAP) guidelines.

Pervasive computing promotes environments where smart, communication-enabled devices cooperate to provide services to people. Due to their inherent complexity, many pervasive applications are built on top of service-oriented platforms, providing a set of facilities simplifying their development and execution. In this paper, we present such a platform, iCasa, extended with an autonomic, service-oriented context module. This module is programmed with a domain-specific service-oriented language built on top of iPOJO, the Apache service-oriented component model. It is validated on smart home applications developed with the Orange Labs.

Services organizations maintain a pipeline of sales opportunities with different maturity level (belonging to progressive sales stages), lifespan (time to close) and contract values at any time point. As time goes, some opportunities close (contract signed, or lost) and new opportunities are added to the pipeline. Accurate forecasting of contract signing by the end of a time period (e.g., quarterly) is highly desirable to make appropriate sales activity management with respect to the projected revenue. While the problem of sales forecasting has been investigated in general, two specific aspects of sales engagement for services organizations, which entail additional complexity, have not been thoroughly investigated: (i) capturing the growth trend of current pipeline, and (ii) incorporating current pipeline build-up in updating the prediction model. We formulate these two issues as a dynamic curve-fitting problem in which we build a sales forecasting model by balancing the effect of current pipeline data and the model trained based on historical data. There are two challenges in doing so, (i) how to mathematically define such a balance and (ii) how to dynamically update the balance as more new data become available. To address these two issues, we propose a novel discrete-constraint method (DCM). It achieves the balance via fixing the value of certain model parameters and applying a leave-one-out algorithm to determine an optimal free parameter number. By conducting experiments on real business data, we demonstrate the superiority of DCM in sales pipeline forecasting.

The goal of a Service System in an organization is to deliver uninterrupted service towards achieving business success. Ticketing system is an example of a Service System which is responsible for handling huge volumes of tickets generated by large enterprise IT (Information Technology) infrastructure components and ensuring smooth operation. Instead of manual screening one needs to extract information automatically from them to gain insights to improve operational efficiency. To ensure better operation we propose a framework to cluster incident tickets based on their textual context that can eliminate manual classification of them, which is labor intensive and costly. Further we label each of the clusters by generating meaningful keywords as logical itemsets, extracting candidate labels from Wikipedia articles, and finally scoring each of labels against each cluster. These labels can reflect an adequate and concise specification of each cluster. Further we experiment our approach with industrial ticket data from three different domains and report on the learned experience. We believe that our framework for clustering and labeling will enable enterprises to prioritize the issues in their IT infrastructure and improve the reliability and availability of their services.