Introduction: Machine Learning as a Service (MLaaS) is a capture term for a range of cloud-based platforms that use machine learning tools to produce solutions that help machine learning professionals. Many cloud-based service providers have led the road in recent years to provide I.T. specialists with comparatively cheap and instantly available machine learning solutions to simplify machine learning solutions. As a result, there is a greater need to compare cloud-based services that deliver machine learning solutions. As a result, we chose to evaluate AWS sagemaker and Azure ML cloud services in terms of user experience when loading, training, and serving ML models.                                            Background: The use of cloud computing has been increased these days, the companies that provide these services have been gradually increased. Although there are many cloud services available on the market, users should always select the more flexible and efficient ones to use. As a result, our research is focused on comparing cloud services in terms of user experience. Assessment approaches and concepts such as Cognitive Walkthrough and Usability heuristics apply to our study as we delve deeper into user interaction and experience. In this case, the user interfaces of Microsoft Azure Machine Learning Studio and Amazon Web Services sagemaker are compared while loading, training, and serving machine learning models. Objectives: The main objective of this thesis is to compare and evaluate the two cloud services such as AWS sagemaker and Microsoft Azure ML while loading, training, and serving machine learning models to decide which of these two cloud services has the best user interface from the users' perspective using Cognitive Walkthrough. Methods: Determining the best cloud service in terms of user experience between AWS Sage Maker and Microsoft Azure ML is done using Cognitive Walkthrough by executing selected tasks in both cloud services, and comparison is done using Usability heuristics to reach our research conclusions. Results: The results originated from the cognitive walkthrough, and comparison with Usability heuristics are presented in graphical formats such as pie charts. The results of cognitive walkthrough are obtained after completion of each task and best cloud service in users’ perspective is obtained. Conclusions: Finally, we conclude Microsoft Azure machine learning studio is better than AWS sagemaker in terms of user-experience while performing the specified tasks such as loading, training and serving ML models in both the cloud services.

<p>It was an online defence session.</p>

ith the revolution of the internet, new applications have emerged in our daily life. People are dependent on services for transportation, bank matters, and communication. Services availability is crucial for their survival and competition against other service providers. Achieving good availability is a challenging task. The latest trend is migrating systems to the cloud. The cloud provides numerous methods to prevent downtimes, such as auto-scaling, continuous deployment, continuous monitoring, and more. However, failures can still occur even though the preemptive techniques fulfill their purpose. Monitoring the system gives insights into the system's actual state, but it is up to the maintainer to interpret these insights. This thesis investigates how machine learning can predict future crashes of Kubernetes pods based on the metrics collected from them. At the start of the project, there was no available data on pod crashes, and the solution was to simulate a 10-tier microservice system in a Kubernetes cluster to create generic data. The project applies two different models, a Random Forest model and a Temporal Convolutional Networksmodel, where the first-mentioned acted as a baseline model. They predict if a failure will occur within a given prediction time window based upon a 15-minutes of data. The project evaluated three different prediction time windows. The five-minute prediction time window resulted in the best foresight based on the models' accuracy. The Random Forest model achieved an accuracy of 73.4 %, while the TCN model achieved an accuracy of 77.7 %. Predictions of the models can act as an early alert of incoming failure, which the system or a maintainer can act upon to improve the availability of its system.

The IEEE Time-Sensitive Networking (TSN) standards offer a promising solution to deal with the challenge of supporting high-bandwidth, low-latency, and predictable communication in distributed embedded systems. Although TSN provides a gate mechanism to support the low-jitter transmission of high-priority time-triggered traffic, it also brings complexity to the network design as the configuration of such mechanism together with support for low-priority transmission is non-trivial. Moreover, the combination of the gate mechanism and the Credit-based Shaper (CBS) mechanism in TSN deals with many configuration parameters, hence finding the most suitable configuration is complex. To avoid this complexity, the Best-effort (BE) class is sometimes used as an alternative channel to the classes that undergo the CBS mechanism, through which the real-time traffic without strict deadlines is transmitted with a minimum level of Quality of Service (QoS). On the other hand, the end stations that operate based on the legacy communication standards might not support the TSN's traffic shaping mechanisms, hence the designers need to assign the legacy traffic to use the BE class in a TSN network. To the extent of our knowledge, there is no implicit mechanism to support the QoS of BE in a TSN network. Hence, utilizing BE as an alternative to other classes must be guaranteed in terms of meeting the timing requirements, i.e., response times and end-to-end delays. Therefore, the work in this thesis aims at developing techniques and solutions to support the QoS of the lower-priority classes in TSN. In this regard, this work improves the scheduling solutions of high-priority time-triggered traffic to reduce the latency of BE traffic and develops techniques to verify the timing properties of BE traffic considering the impact of all other traffic classes in TSN. Furthermore, the work in this thesis extends the existing end-to-end data-propagation delay analysis for distributed real-time systems based on TSN networks. Finally, the applicability of the proposed techniques is verified and demonstrated by automotive application use cases.

Function-as-a-Service (FaaS) is an increasingly common platform for many kinds of applications and services, replacing the need to maintain and setup hardware or virtual machines to host functionality in the cloud. The billing model for FaaS is commonly based on actual usage, which makes the ability to estimate the performance and latency of an application before invoking it valuable. This thesis evaluates if previously defined algorithms for offline latency estimation, can be adapted to work with online data. Performing online estimation of latency potentially enables cheaper estimations, as no extra executions are neccessary, and latency estimation of applications and functions that can not be executed spu- riously. The experiments show that for a set of test applications, the previously defined algorithms can achieve greater than 95% accuracy, and that a non-graph based estimation using exponential moving average can achieve greater than 98% accuracy.

The use of open-source systems in software development is more of a general practice in firms, institutions and large enterprises. Such systems produce and oper- ate on a large amount of enterprise data. Thus the protection of the access, integrity and confidentiality of enterprise information assets is of high importance. Equally so is avoiding cyber attacks and reducing their impact in case of such occurrence. This study, "Predicting Security Vulnerabilities in open-source systems", is to use mod- ern methods such as machine learning, to predict potential security vulnerabilities by analysing the source code of the system. The project uses a five-stage approach, viz., Data Collection, Data cleansing, feature selection, Building ML models and Cross Validation. Data collected from the public domains of open-source systems such as Apache Tomcat. The source code of all Apache Tomcat versions were collected and were used to derive individual code metrics using aniche-ck’s tool. The tool resulted in a grand total of 43 metrics including the quantities of various components in a class, coupling between objects and many more. This project makes use of tradi- tional machine learning models to report which among them would prove effective to predict security vulnerabilities. Models are generated using different algorithms such as Naive Bayes Classifier, Logistic Regression, Decision Tree. In further ex- perimentation, K-fold cross validation is introduced with a value of 10, making it 10-fold validation. It is noted that the models have a higher precision and recall val- ues for predicting security vulnerabilities using source code metrics features. Finally, the KNN classifier proved to be the most efficient of all the models that were taken in to account. The decision tree classifier also performed well but failed when it came to multi-class classification

Distributed cloud is a key technology for 5G networks and is emerging as an alternative cloud infrastructure for hosting latency critical and traffic-intense applications. Placing computational resources on the edge of networks allows applications to be hosted closer to end users and traffic generating sources, which will reduce latency and traffic deeper in the network. This thesis presents a two-phase approach to solve the combinatorial op-timization problem of latency and traffic aware container placement in distributed cloud. Each phase is evaluated using a phase-specific simulated environment. The first phase involves placing containers in data centers and is solved using an integer programming model. Three different objective functions are presented and evaluated using acceptance ratio and average cost as performance metrics. In the second phase, containers are placed in servers. A traffic-aware heuristic is presented and evaluated against traditional bin packing heuristics. The traffic-aware heuristic managed to drastically reduce all traffic-related metrics at the cost of a few additional active servers in comparison to the bin packing heuristics. The traffic-aware heuristic can therefore be a good approach when placing traffic-intense applications in data centers in order to avoid network congestion.

This report describes the development of a Windows Store application which runs on a Windows 8 tablet PC. The application was developed for a company that targets television based services and have the requirements of displaying a television guide based on data from a device connected to the television. Both the application and the device communicates with each other via an intermediary web based service. Established design patterns for object-oriented programming languages were used along with a method called test-driven development in order to accomplish this. Furthermore a comparison between mocking frameworks will be presented for choosing a suitable framework when developing Windows Store applications. The communication model between application and the device is also discussed and evaluated for whether or not it is a suitable method of communication.

Context -- The rapid development of Industry 4.0 and Industrial Cyber-Physical Systems is leading to the exponential growth of unprocessed volumes of data. Industrial cloud computing has shown great potential as a solution that can provide the necessary resources for processing these data. However, in order to be widely adopted, it must provide satisfactory levels of QoS. The lack of a standardized model of quality attributes to be used for assessing QoS raises significant concerns. Objective -- This study aims to provide a map of current research on QoS in industrial cloud computing, focusing on identifying and classifying the quality attributes that are currently most commonly used to evaluate QoS. Method -- To achieve our objective, we conducted a systematic mapping study of the state-of-the-art of QoS in industrial cloud computing. Our search yielded 1063 potentially relevant studies that were subject to a rigorous selection process, resulting in a final set of 42 primary studies. Key information from the primary studies was extracted according to the categories of a well-defined classification framework. Results -- The analysis of the extracted data highlighted the following main findings: (i) research largely focuses on providing solution proposals that require a more solid validation, (ii) the adoption of cloud technologies is closely related to performance indicators, while research on other quality attributes is quite limited, (iii) there is a lack of research on security in industrial cloud computing, (iv) approaches are in most cases not targeting explicitly a specific industrial domain, (v) there is a strong focus on the impact of virtualization solutions on QoS, and (vi) research efforts are oriented towards the improvement of QoS through scheduling. Conclusion -- These results can help the research community identify trends, limitations, and research gaps on QoS in industrial cloud computing, and reveal possible directions for future research.

Taking a driver license is an expensive and time consuming process. This process can be facilitated by the use of modern technology. This report will address how it is possible to take driving lessons to the next level by practicing in a virtual environment. The goal of this project is to create a driving school simulator with modern virtual reality technology which will put the driver in realistic traffic situations for an educational purpose. / Att ta körlektioner har länge varit en kostsam och tidskrävande process. Den här processen kan underlättas med användning av ny teknik. Denna rapport kommer gå in på hur man kan ta körlektioner till nästa nivå genom att öva i en virtuel miljö. Målet med projektet är att skapa en körskolesimulator med hjälp av dagens virtual reality teknik som kan sätta föraren i realistiska trafiksituationer i utbildningssyfte.