Anson Bastos, Shreeya Venneti, Anjaly Parayil
et al.
Reliability of large-scale cloud services is critical for user satisfaction and business continuity. Despite significant investments in reliability engineering, production incidents remain inevitable, often leading to customer impact and operational overhead. In large cloud companies, multiple services are deployed across regions necessitating robust health monitoring systems. However, the current monitor configuration process is manual, largely reactive and ad hoc, resulting in gaps in coverage and redundant alerts. In this paper, we present a comprehensive study of monitor creation in Microsoft, identifying key components in the existing process. We further design a modular recommendation framework that processes the graph structured service entities to suggest optimal monitor configurations. Through extensive experimentation on historical data and user study of recommendations for production services at Microsoft, we demonstrate the efficacy of our approach in providing relevant recommendations for monitor configurations.
Nowadays, there are many similar services available on the internet, making Quality of Service (QoS) a key concern for users. Since collecting QoS values for all services through user invocations is impractical, predicting QoS values is a more feasible approach. Matrix factorization is considered an effective prediction method. However, most existing matrix factorization algorithms focus on capturing global similarities between users and services, overlooking the local similarities between users and their similar neighbors, as well as the non-interactive effects between users and services. This paper proposes a matrix factorization approach based on user information entropy and region bias, which utilizes a similarity measurement method based on fuzzy information entropy to identify similar neighbors of users. Simultaneously, it integrates the region bias between each user and service linearly into matrix factorization to capture the non-interactive features between users and services. This method demonstrates improved predictive performance in more realistic and complex network environments. Additionally, numerous experiments are conducted on real-world QoS datasets. The experimental results show that the proposed method outperforms some of the state-of-the-art methods in the field at matrix densities ranging from 5% to 20%.
This paper presents a grid-aware probabilistic approach to compute the aggregated flexibility at the grid connection point (GCP) of active distribution networks (ADNs) to allow the participation of DERs in ancillary services (AS) markets. Specifically an optimal power flow (OPF) method using a linear network model is used to compute the aggregated capability for the provision of multiple AS. We start from the method proposed in [1] and extend it to allow for optimizing the provision of multiple services simultaneously, ensure cost-effectiveness of the used DERs and handle uncertainties in a probabilistic way. The allocation of individual DERs power flexibilities accounts for the operational costs associated to the provision of different services and ensures cost-effectiveness while maximizing the value of the advertised aggregated flexibility, assuming known service prices. Empirical uncertainty sets are obtained to achieve a predefined coverage of the probability distribution in line with recent developments in the Nordic AS markets. Finally, a feeder-decomposition approach is proposed to ensure the methods applicability to realistic distribution networks with a large number of buses. Different case studies show the effectiveness of the method, highlight the importance of accounting for network constraints and illustrate its applicability to realistic distribution systems.
The Architecture Analysis & Design Language (AADL) is an architecture description language for design of cyber-physical systems--machines controlled by software. The AADL standard, SAE International AS5506D, describes Run-Time Services (RTS) to be provided to execute AADL models in accordance with semantics defined by the standard. The RTS of primary concern are transport services and timing services. Although, the study presented in [1] sets a foundation for the formal semantics of AADL, but without modeling time. This paper extends and simplifies this formalization using a modal logic defined by a Kripke structure, to explicitly include time. The RTS defined in the AADL standard are also expanded to support reactive state-transition machines of the Behavior Specification annex standard language (BA) and its closely-related, formally-defined counterpart, the Behavior Language for Embedded Systems with Software (BLESS). An example of AADL RTS with time, implemented by the High Assurance Modeling and Rapid Engineering for Embedded Systems (HAMR) for state-transition machine behavior written in BLESS, is also presented.
Phishing websites are a common phenomenon among Tor onion services, and phishers exploit that it is tremendously difficult to distinguish phishing from authentic onion domain names. Operators of onion services devised several strategies to protect their users against phishing. But as we show in this work, none protect users against phishing without producing traces about visited services - something that particularly vulnerable users might want to avoid. In search of a solution we review prior research addressing this problem, and find that only two known approaches, hash visualization and PAKE, are capable of solving this problem. Hash visualization requires users to recognize large hash values. In order to make hash visualization more practical we design a novel mechanism called recognizer, which substantially reduces the amount of information that users must recognize. We analyze the security and privacy properties of our system formally, and report on our prototype implementation as a browser extension for the Tor web browser.
Rasha Faqeh, Andrè Martin, Valerio Schiavoni
et al.
Fault-tolerance techniques depend on replication to enhance availability, albeit at the cost of increased infrastructure costs. This results in a fundamental trade-off: Fault-tolerant services must satisfy given availability and performance constraints while minimising the number of replicated resources. These constraints pose capacity planning challenges for the service operators to minimise replication costs without negatively impacting availability. To this end, we present PCRAFT, a system to enable capacity planning of dependable services. PCRAFT's capacity planning is based on a hybrid approach that combines empirical performance measurements with probabilistic modelling of availability based on fault injection. In particular, we integrate traditional service-level availability mechanisms (active route anywhere and passive failover) and deployment schemes (cloud and on-premises) to quantify the number of nodes needed to satisfy the given availability and performance constraints. Our evaluation based on real-world applications shows that cloud deployment requires fewer nodes than on-premises deployments. Additionally, when considering on-premises deployments, we show how passive failover requires fewer nodes than active route anywhere. Furthermore, our evaluation quantify the quality enhancement given by additional integrity mechanisms and how this affects the number of nodes needed.
Applications providing location-based services (LBS) have gained much attention and importance with the notion of the internet of things (IoT). Users are utilizing LBS by providing their location information to third-party service providers. However, location data is very sensitive that can reveal user's private life to adversaries. The passive and pervasive data collection in IoT upsurges serious issues of location privacy. Privacy-preserving location-based services are a hot research topic. Many anonymization and obfuscation techniques have been proposed to overcome location privacy issues. In this paper, we have proposed a hybrid location privacy scheme (H-LPS), a hybrid scheme mainly based on obfuscation and collaboration for protecting users' location privacy while using location-based services. Obfuscation naturally degrades the quality of service but provides more privacy as compared to anonymization. Our proposed scheme, H-LPS, provides a very high-level of privacy yet provides good accuracy for most of the users. The privacy level and service accuracy of H-LPS are compared with state-of-the-art location privacy schemes and it is shown that H-LPS could be a candidate solution for preserving user location privacy in location-based services.
Mahdi Fahmideh, Jafar Habibi, Fereidoon Shams
et al.
Service Oriented Software Engineering is based on concepts and principles for constructing complex enterprise systems in which services as building block of the system, are distributed in large networks. The main goal of the service oriented methodologies is to define a process for development and maintenance of service based systems. Most of the Service Oriented methodologies are not mature enough compared with traditional software development methodologies such as Object Oriented or Component-Based. Hence, defining an evaluation framework will be useful for comparing methodologies for identifying their strengths and weaknesses, defining new methodologies or extending existing Service Oriented methodologies. At the time being, there is no complete evaluation framework for evaluating Service Oriented methodologies. The principal objective of this paper is to introduce a comprehensive evaluation framework for evaluating Service-Oriented methodologies. This evaluation tool is appropriate for methodology engineers to develop new methodologies, as well as project managers to select an appropriate methodology at a specific project.
Mahdi Fahmideh, Mohsen Sharifi, Fereidoon Shams
et al.
Software systems development nowadays has moved towards dynamic composition of services that run on distributed infrastructures aligned with continuous changes in the system requirements. Consequently, software developers need to tailor project specific methodologies to fit their methodology requirements. Process patterns present a suitable solution by providing reusable method chunks of software development methodologies for constructing methodologies to fit specific requirements. In this paper, we propose a set of high-level service-oriented process patterns that can be used for constructing and enhancing situational service-oriented methodologies. We show how these patterns are used to construct a specific service-oriented methodology for the development of a sample system. Keywords. Service-Oriented Software Development Methodologies, Process Patterns, Process Meta-Model, Situational Method Engineering
Network function virtualization (NFV) fosters innovation in the networking field and reduces the complexity involved in managing modern-day conventional networks. Via NFV, the provisioning of a network service becomes more agile, whereby virtual network functions can be instantiated on commodity servers and data centers on demand. Network functions can be either mandatory or best-effort. The former type is strictly necessary for the correctness of a network service, whereas the latter is preferrable yet not necessary. In this paper, we study the online provisioning of NFV-enabled network services. We consider both unicast and multicast NFV-enabled services with multiple mandatory and best-effort NF instances. We propose a primal-dual based online approximation algorithm that allocates both processing and transmission resources to maximize a profit function, subject to resource constraints on physical links and NFV nodes. The online algorithm resembles a joint admission mechanism and an online composition, routing and NF placement framework. The online algorithm is derived from an offline formulation through a primal-dual based analysis. Such analysis offers direct insights and a fundamental understanding on the nature of the profit-maximization problem for NFV-enabled services with multiple resource types.
Martin Klein, Lyudmila Balakireva, Harihar Shankar
Services and applications based on the Memento Aggregator can suffer from slow response times due to the federated search across web archives performed by the Memento infrastructure. In an effort to decrease the response times, we established a cache system and experimented with machine learning models to predict archival holdings. We reported on the experimental results in previous work and can now, after these optimizations have been in production for two years, evaluate their efficiency, based on long-term log data. During our investigation we find that the cache is very effective with a 70-80% cache hit rate for human-driven services. The machine learning prediction operates at an acceptable average recall level of 0.727 but our results also show that a more frequent retraining of the models is needed to further improve prediction accuracy.
Mohammad Riyaz Belgaum, Safeeullah Soomro, Zainab Alansari
et al.
Cloud computing has been gaining popularity in the recent years. Several studies are being proceeded to build cloud applications with exquisite quality based on users demands. In achieving the same, one of the applied criteria is checkpoint based load balancing in real time scheduling through which suitable cloud service is chosen from a group of cloud services candidates. Valuable information can be collected to rank the services within this checkpoint based load balancing. In order to attain ranking, different services are needed to be invoked in the cloud, which is time consuming and wastage of services invocation. To avoid the same, this chapter proposes an algorithm for predicting the ranks of different cloud services by using the values from previously offered services.
Gianluca Reali, Mauro Femminella, Emilia Nunzi
et al.
This paper provides a global picture about the deployment of networked processing services for genomic data sets. Many current research make an extensive use genomic data, which are massive and rapidly increasing over time. They are typically stored in remote databases, accessible by using Internet. For this reason, a significant issue for effectively handling genomic data through data networks consists of the available network services. A first contribution of this paper consists of identifying the still unexploited features of genomic data that could allow optimizing their networked management. The second and main contribution of this survey consists of a methodological classification of computing and networking alternatives which can be used to offer what we call the Genomic-as-a-Service (GaaS) paradigm. In more detail, we analyze the main genomic processing applications, and classify not only the main computing alternatives to run genomics workflows in either a local machine or a distributed cloud environment, but also the main software technologies available to develop genomic processing services. Since an analysis encompassing only the computing aspects would provide only a partial view of the issues for deploying GaaS system, we present also the main networking technologies that are available to efficiently support a GaaS solution. We first focus on existing service platforms, and analyze them in terms of service features, such as scalability, flexibility, and efficiency. Then, we present a taxonomy for both wide area and datacenter network technologies that may fit the GaaS requirements. It emerges that virtualization, both in computing and networking, is the key for a successful large-scale exploitation of genomic data, by pushing ahead the adoption of the GaaS paradigm. Finally, the paper illustrates a short and long-term vision on future research challenges in the field.
In Interactive Information Retrieval (IIR) different services such as search term suggestion can support users in their search process. The applicability and performance of such services is either measured with different user-centered studies (like usability tests or laboratory experiments) or, in the context of IR, with their contribution to measures like precision and recall. However, each evaluation methodology has its certain disadvantages. For example, user-centered experiments are often costly and small-scaled; IR experiments rely on relevance assessments and measure only relevance of documents. In this work we operationalize the usefulness model of Cole et al. (2009) on the level of system support to measure not only the local effect of an IR service, but the impact it has on the whole search process. We therefore use a log-based evaluation approach which models user interactions within sessions with positive signals and apply it for the case of a search term suggestion service. We found that the usage of the service significantly often implicates the occurrence of positive signals during the following session steps.
5G mobile networks encompass the capabilities of hosting a variety of services such as mobile social networks, multimedia delivery, healthcare, transportation, and public safety. Therefore, the major challenge in designing the 5G networks is how to support different types of users and applications with different quality-of-service requirements under a single physical network infrastructure. Recently, network slicing has been introduced as a promising solution to address this challenge. Network slicing allows programmable network instances which match the service requirements by using network virtualization technologies. However, how to efficiently allocate resources across network slices has not been well studied in the literature. Therefore, in this paper, we first introduce a model for orchestrating network slices based on the service requirements and available resources. Then, we propose a Markov decision process framework to formulate and determine the optimal policy that manages cross-slice admission control and resource allocation for the 5G networks. Through simulation results, we show that the proposed framework and solution are efficient not only in providing slice-as-a-service based on the service requirements, but also in maximizing the provider's revenue.
With the increasing processing capability of mobile platforms and advancements in Internet of Things, modern mobile devices have shown a favorable prospect for on-the-go service provisioning. However, there is much to be done to realize this. A detailed, dynamic, and lightweight service description is an important requirement for automatic and efficient discovery, selection, and subsequent provisioning of services over mobile devices. Traditional approaches for service description are usually not directly adaptable to mobile environments owing to the latter's dynamic and distinct nature. In this paper, we propose a dynamic, lightweight, extensible, and detailed service description especially designed for mobile environments, considering crucial aspects such as isolated data source, collaborator partners, and hardware aspects along with the functional, non-functional, business, and contextual aspects. The description has been partitioned along these lines and various parts of the description are distributed between service registries and the mobile service providers. An up-to-date and light weight description has been achieved by this, without compromising on the overall consistency of the description. A prototype of the proposed system has been implemented with the intent of validating the feasibility of the approach. Further, the proposed approach is suitable for a heterogeneous environment comprising both wired and wireless systems.
Cloud computing has achieved great success in modern IT industry as an excellent computing paradigm due to its flexible management and elastic resource sharing. To date, cloud computing takes an irrepalceable position in our socioeconomic system and influences almost every aspect of our daily life. However, it is still in its infancy, many problems still exist.Besides the hotly-debated security problem, availability is also an urgent issue.With the limited power of availability mechanisms provided in present cloud platform, we can hardly get detailed availability information of current applications such as the root causes of availability problem,mean time to failure, etc. Thus a new mechanism based on deep avaliability analysis is neccessary and benificial.Following the prevalent terminology 'XaaS',this paper proposes a new win-win concept for cloud users and providers in term of 'Availability as a Service' (abbreviated as 'AaaS').The aim of 'AaaS' is to provide comprehensive and aimspecific runtime avaliabilty analysis services for cloud users by integrating plent of data-driven and modeldriven approaches. To illustrate this concept, we realize a prototype named 'EagleEye' with all features of 'AaaS'. By subscribing corresponding services in 'EagleEye', cloud users could get specific availability information of their applications deployed in cloud platform. We envision this new kind of service will be merged into the cloud management mechanism in the near future.
Service oriented architecture is more and more used in the companies, The importance of the service orientation and its advantages with the information system of the company, confront us to a new challenge. It is primarily to ensure the decision aspect of the information system of company by adopting Services Orientated Architecture like support architecture. This engineering must ensure on the one hand the flexibility of the information system and on the oth-er hand, avoids the redevelopment of the system by the decisions request. In the actual position, several obstacles force the installation of the SOA within the company. It is basically a question of the lack of method to be implemented to define the services architecture within the information system of the company which takes account the decisional aspect. Moreover, there's no existing works which treat this challenge. On the basis of these notes, we are interested, in this paper, on the develop-ment of a three-dimensional new architecture for the integration of the deci-sional aspect in SOA architectures, so that they are used perfectly. The proposal takes support, mainly, on the use of a coupling MAS-SOA. To demonstrate the application of our proposition, we use two study cases: project management and Evapo-transpiration.
Data collection is a key component of an information system. The widespread penetration of ICT tools in organizations and institutions has resulted in a shift in the way the data is collected. Data may be collected in printed-form, by e-mails, on a compact disk, or, by direct upload on the management information system. Since web services are platform-independent, it can access data stored in the XML format from any platform. In this paper, we present an interface which uses web services for data collection. It requires interaction between a web service deployed for the purposes of data collection, and the web address where the data is stored. Our interface requires that the web service has pre-knowledge of the address from where the data is to be collected. Also, the data to be accessed must be stored in XML format. Since our interface uses computer-supported interaction on both sides, it eases the task of regular and ongoing data collection. We apply our framework to the Education Management Information System, which collects data from schools spread across the country.