1,721,017 research outputs found
Scalable classification of QoS for real-time interactive applications from IP traffic measurements
Full text linkMeasurement of network Quality of Service (QoS) has attracted considerable research effort over the last two decades. The recent trend towards Internet Service Providers (ISP's) offering application-specific QoS is creating possibilities for more sophisticated QoS metrics to be offered by ISP's in service level agreements. This in turn could be used for the purposes of improved network optimization and billing according to application specific QoS guarantees. We report a scalable near real-time approach using passively logging IP traffic data for classification of application latency and packet loss across a range of real-time interactive applications. We run six experiments involving Minecraft, Quake 3 Urban Terror, VLC video streaming and the commercial Wirofon VOIP application. We use a mixture of laboratory and real-world deployments, with run times ranging from hours to days, and observe a combination of real and simulated ISP latency and packet loss events. Our binary classification (i.e. classes 'OK' or 'lag') 10-fold cross validation F1 scores are between 0.80 and 0.93 depending on application type. Our multi-class classification (i.e. classes representing discrete packet loss or latency ranges) 10-fold cross validation F1 scores for Minecraft are 0.89 for latency and 0.90 for packet loss. With new business models between ISP's and application developers being actively considered this work represents a significant contribution to the debate by providing scientific evidence relating to a novel approach to scalable QoS measuremen
Multi-level data fusion of environmental data in future internet applications
No full textThe rapid increase in environmental observations which are conducted by SMEs, communities and volunteers using affordable in situ sensors at various scales, together with the more established observatories set up by environmental and space agencies using airborne and space-borne sensing technologies is generating serious amounts of BIG data at ever increasing rates. Furthermore, the emergence of Future Internet technologies and the urgent requirements for the deployment of specific enablers for the delivery of processed environmental knowledge in real-time with advanced situation awareness to citizens has reached greater imminence. It is now highly critical to build and provide services which automate the aggregation of data from various sources, while surmounting the semantic gaps, conflicts and heterogeneity in data sources.The early stages of aggregation of data enable the preprocessing of data generated from multiple sources with the reconciliation between temporal gaps in observation time series, and alignment of their respective asynchronicities. As a result, multi-level processes of fusion need to be implemented and made accessible to large communities of users using future internet services.This paper presents the process and the preliminary results using RBF networks methods for the spatial fusion of water quality observations and measurements from asynchronous space-borne, in situ and validated models simulation data sources in the Irish Sea
Experiences monitoring and managing QoS using SDN on testbeds supporting different innovation stages
Full text linkIn recent years there has been a big increase in the number of network-related experiments using software defined networking (SDN) technology. We report on our practical experience over 2 years running network experiments on three classes of testbed facility, each supporting researchers working at a different innovation stage. We run experiments using the commercial Amazon EC2 cloud facility, pre-commercial federated testbed of FIWARE Lab instances and the OFELIA experimental facility. We run an idealized common network experiment on each testbed, reducing its scope where needed to match testbed capabilities, and report details of the practical experience gained using a set of qualitative metrics for direct comparison across classes of testbed. We conclude with a concrete recommendation for pre-commercial testbed facilities to allow better support for network experiments in the future
Classifying space-time images obtained from distributed acoustic sensing
No full textIn this paper we present a classifier that was trained on space-time images obtained from distributed acoustic sensing, for the purpose of monitoring earthquakes. The model is capable of discriminating between actual and non-earthquake events
An extended approach to impact assessment in a Horizon2020 digital manufacturing project
No full textThis paper presents an extended approach to impact assessment (IA) within European Union funded large-scale projects within the manufacturing domain, which may offer value to other research projects and SME organisations seeking to develop detailed organizational reporting, as well as provides an overview of the impact data returned and assessment results. As applied during the European Connected Factory Platform for Agile Manufacturing (EFPF) project (part of the EU Horizon2020 digital manufacturing grant), this paper will detail the processes undertaken as part of this extended approach, demonstrating how project Outcome Indictors and impact assessment criterion can be aligned through an extensive review and integration of existing impact domains, objectives, measures and evidence sources with project documentation to provide a comprehensive IA/KRI framework. It will also report on the results of the process, before concluding by detailing how organisational research may best utilise the approach, and discussing the wider implications for Industry4.0
- …
