-Cloud computing offers the potential to dramatically reduce the cost of software services through the commoditization of information technology assets and ondemand usage patterns. However, the complexity of determining resource provision policies for applications in such complex environments introduces significant inefficiencies and has driven the emergence of a new class of infrastructure called Platform-as-a-Service (PaaS). In this paper, we present a novel PaaS architecture being developed in the EU IST IRMOS project targeting real-time Quality of Service (QoS) guarantees for online interactive multimedia applications. The architecture considers the full service lifecycle including service engineering, service level agreement design, provisioning and monitoring. QoS parameters at both application and infrastructure levels are given specific attention as the basis for provisioning policies in the context of temporal constraints. The generic applicability of the architecture is being verified and validated through implemented scenarios from three important application sectors (film post-production, virtual augmented reality for engineering design, collaborative e-Learning in virtual worlds).

No abstract

A key task of emergency departments is to promptly identify patients who require hospital admission. Early identification ensures patient safety and aids organisational planning. Supervised machine learning algorithms can use data describing historical episodes to make ahead-of-time predictions of clinical outcomes. Despite this, clinical settings are dynamic environments and the underlying data distributions characterising episodes can change with time (data drift), and so can the relationship between episode characteristics and associated clinical outcomes (concept drift). Practically this means deployed algorithms must be monitored to ensure their safety. We demonstrate how explainable machine learning can be used to monitor data drift, using the COVID-19 pandemic as a severe example. We present a machine learning classifier trained using (pre-COVID-19) data, to identify patients at high risk of admission during an emergency department attendance. We then evaluate our model’s performance on attendances occurring pre-pandemic (AUROC of 0.856 with 95%CI [0.852, 0.859]) and during the COVID-19 pandemic (AUROC of 0.826 with 95%CI [0.814, 0.837]). We demonstrate two benefits of explainable machine learning (SHAP) for models deployed in healthcare settings: (1) By tracking the variation in a feature’s SHAP value relative to its global importance, a complimentary measure of data drift is found which highlights the need to retrain a predictive model. (2) By observing the relative changes in feature importance emergent health risks can be identified.

Abstract-In this paper we focus on how Quality of Service guarantees are provided to virtualised applications in the Cloud Computing infrastructure that is being developed in the context of the IRMOS 1 European Project. Provisioning of proper timeliness guarantees to distributed real-time applications involves the careful use of real-time scheduling mechanisms at the virtual-machine hypervisor level, of QoS-aware networking protocols and of proper design methodologies and tools for stochastic modelling of the application. The paper focuses on how we applied these techniques to a case-study involving a real eLearning mobile content delivery application that has been integrated into the IRMOS platform and its achieved performance.

No abstract

Abstract. This paper describes the design and prototype implementation of a novel architecture for integrated concept, metadata and content based browsing and retrieval of museum information. The work is part of a European project involving several major galleries and the aim is to provide more versatile access to digital collections of museum artefacts, including 2-D images, 3-D models and other multimedia representations. An ontology for the museum domain, based on the CIDOC Conceptual Reference Model, is being developed as a semantic layer with references to the digital collection as instance information. A graphical concept browser is an integral component in the user interface, allowing navigation through the semantic layer, display of thumbnails, or full representations of artefacts and textual information in appropriate viewers and the invocation of conventional content based searching or combined querying. Semantic Web technologies are used in system integration to describe how tools for analysis and visualisation can be applied to different data types and sources. This supports flexible and managed formulation, execution and interpretation of the results of distributed multimedia queries. Combined searches using concepts, content and metadata can be initiated from a single user interface.

Abstract. The paper describes the prototype design and development of a multimedia system for museums and galleries. Key elements in the system are the introduction of 3-D models of museum artefacts together with 3-D as well as 2-D content based retrieval and navigation facilities and the development of a semantic layer, centred on an ontology for museums, which aims to expose the richness of knowledge associated with the museum collections and facilitate concept based retrieval and navigation integrated with that based on content and metadata. Interoperability protocols are designed to allow external applications to access the collection and an example is given of an e-Learning facility which uses models extracted to a virtual museum.

Service Level Agreements (SLA) include quality of service (QoS) constraints and bounds that have to be honoured by the service provider. To maximise the Service Provider revenue while satisfying the QoS requirements of the agreed SLAs it is important to be able to perform a dynamic distribution of the service provider resources between the services and SLAs. This distribution should be based on the current status and predicted evolution of the QoS characteristics. This paper describes the experiences managing SLA obligations from a service provider perspective in a scenario where dynamic deployment of services can be undertaken. The main issues faced to deal with the management of SLAs in this context are detailed. The adopted solution, based on GRIA (a Service Oriented Architecture framework) is discussed.