Search

With increasing availability and use of Internet of Things (IoT) devices large amounts of streaming data is now being produced at high velocity. Applications which require low latency response such as video surveillance demand a swift and efficient analysis of this data. Existing approaches employ cloud infrastructure to store and perform machine learning based analytics on this data. This centralized approach has limited ability to support analysis of real-time, large-scale streaming data due to network bandwidth and latency constraints between data source and cloud. We propose RealEdgeStream (RES) an edge enhanced stream analytics system for large-scale, high performance data analytics. The proposed approach investigates the problem of video stream analytics by proposing (i) filtration and (ii) identification phases. The filtration phase reduces the amount of data by filtering low value stream objects using configurable rules. The identification phase uses deep learning inference to perform analytics on the streams of interest. The stages are mapped onto available in-transit and cloud resources using a placement algorithm to satisfy the Quality of Service (QoS) constraints identified by a user. The job completion in the proposed system takes 49\% less time and saves 99\% bandwidth compared to a centralized cloud-only based approach.

Dramatic changes in the technology landscape marked by increasing scales and pervasiveness of compute and data have resulted in the proliferation of edge applications aimed at effectively processing data in a timely manner. As the levels and fidelity of instrumentation increases and the types and volumes of available data grow, new classes of applications are being explored that seamlessly combine real-time data with complex models and data analytics to monitor and manage systems of interest. However, these applications require a fluid integration of resources at the edge, the core, and along the data path to support dynamic and data-driven application workflows, that is, they need to leverage a computing continuum. In this article, we present our vision for enabling such a computing continuum and specifically focus on enabling edge-to-cloud integration to support data-driven workflows. The research is driven by an online data-driven tsunami warning use case that is supported by the deployment of large-scale national environment observation systems. This article presents our overall approach as well as current status and next steps.

Combining edge processing (at data capture site) with analysis carried out while data is enroute from the capture site to a data center offers a variety of different processing models. Such in-transit nodes include network data centers that have generally been used to support content distribution (providing support for data multicast and caching), but have recently started to offer user-defined programmability, through Software Defined Networks (SDN) capability, e.g., OpenFlow and Network Function Visualization (NFV). We demonstrate how this multi-site computational capability can be aggregated to support video analytics, with Quality of Service and cost constraints (e.g., latency-bound analysis). The use of SDN technology enables separation of the data path from the control path, enabling in-network processing capabilities to be supported as data is migrated across the network. We propose to leverage SDN capability to gain control over the data transport service with the purpose of dynamically establishing data routes such that we can opportunistically exploit the latent computational capabilities located along the network path. Using a number of scenarios, we demonstrate the benefits and limitations of this approach for video analysis, comparing this with the baseline scenario of undertaking all such analysis at a data center located at the core of the infrastructure.

As devices existing at the edge of the network improve in their processing and data storage capacity, there is increasing potential to host and enact services on such devices. A workflow that was traditionally enacted on a data centre can be fragmented across both edge and data centre hosted resources. The following aspects are investigated in this work: (i) mechanisms for dividing a workflow across edge and cloud/data centre resources; (ii) service hosting environments that can be shared across edge and data centre resources; (iii) performance metrics that can influence service placement and selection. An "edge orchestrator" is a resource manager that makes such decisions on the behalf of a user application, and which may be centralised or distributed. An industry scenarios is used to illustrate decision points that influence such choices within an edge orchestrator. The overall objective considered is the completion of the workflow within some deadline constraint by the edge orchestrator.

Internet of Things (IoT) has been ever-growing over the last few years. The IoT devices generate a massive amount of data that should be transmitted to the cloud for computing. Cloud consolidation and centralization lead to many network hops between the IoT devices and its associated cloud which makes two critical problems: (i) high latencies (ii) high bandwidth consumption in the IoT domain. Network Function Virtualization (NFV), Software Defined Network (SDN) and fog computing have been emerged to address these problems. In the Fog-to-Cloud (F2C) architecture, Fog and cloud work together to provide computing, storage, and application services in the IoT domain. To build complex services a specific set of virtual network functions can be chained together in a specific order which is known as Service Function Chaining (SFC). The joint VNF placement and traffic routing are called SFC mapping. In this paper, we propose an Integer Linear Program (ILP) model to solve SFC mapping in the fog-to-cloud Computing System in order to minimize the overall end-to-end (e2e) latency of IoT devices. We observe that our approach reduces the overall e2e latency of IoT devices significantly. Moreover, our approach helps us to analyze the effect of a number of instances in the end-to-end latency of IoT devices.

Estimating energy costs for an industrial process can be computationally intensive and time consuming, especially as it can involve data collection from different (distributed) monitoring sensors. Industrial processes have an implicit complexity involving the use of multiple appliances (devices/ sub-systems) attached to operation schedules, electrical capacity and optimisation setpoints which need to be determined for achieving operational cost objectives. Addressing the complexity associated with an industrial workflow (i.e. range and type of tasks) leads to increased requirements on the computing infrastructure. Such requirements can include achieving execution performance targets per processing unit within a particular size of infrastructure i.e. processing & data storage nodes to complete a computational analysis task within a specific deadline. The use of ensemblebased edge processing is identifed to meet these Quality of Service targets, whereby edge nodes can be used to distribute the computational load across a distributed infrastructure. Rather than relying on a single edge node, we propose the combined use of an ensemble of such nodes to overcome processing, data privacy/ security and reliability constraints. We propose an ensemble-based network processing model to facilitate distributed execution of energy simulations tasks within an industrial process. A scenario based on energy profiling within a fisheries plant is used to illustrate the use of an edge ensemble. The suggested approach is however general in scope and can be used in other similar application domains.

Software-defined networks (SDNs) have proven to be an efficacious tool for undertaking complex data analysis and manipulation within data intensive applications. SDN technology allows us to separate the data path from the control path, enabling in-network processing capabilities to be supported as data is migrated across the network. We propose to leverage software defined networking to gain control over the data transport service with the purpose of dynamically establishing data routes such that we can opportunistically exploit the latent computational capabilities located along the network path. This strategy allows us to minimize waiting times at the destination data center and to cope with spikes in demand for computational capability. We validate our approach using a smart building application in a multi cloud infrastructure. We show how the in-transit processing strategy increases the computational capabilities of the infrastructure and influences the percentage of job completion without significantly impacting costs and overheads.

Cloud computing has emerged as attractive platform for computing data intensive applications. However, efficient computation of this kind of workloads requires understanding how to store, process, and analyse large volumes of data in a timely manner. Many “smart cities” applications, for instance, identify how data from building sensors can be combined together to support applications such as emergency response, energy management, etc. Enabling sensor data to be transmitted to a cloud environment for processing provides a number of benefits, such as scalability and on-demand provisioning of computational resources. In this chapter, we propose the use of a multi-layer cloud infrastructure that distributes processing over sensing nodes, multiple intermediate/gateways nodes, and large data centres. Our solution aims at utilising the pervasive computational capabilities located at the edge of the infrastructure and along the data path to reduce data movement to large data centres located “deep” into the infrastructure and perform a more efficient use of computing and network resources.

Pervasive computational ecosystems that combine data sources and computing/communication resources in self-managed environments, such as the ones powered by Internet of Things (IoT) devices, have the potential to automate and facilitate many aspects of our lives, and impact a variety of applications, from the management of extreme events to the optimization of everyday processes. However, this vision remains mostly unrealized despite the fact that the technology to achieve it exists, largely because of the gap between our ability to collect data and our ability to gain insight from it. In this paper, we discuss the challenges associated with providing a pervasive computational ecosystem. We then present our vision of how to best support data-driven computational ecosystems and propose a conceptual architecture that leverages ideas from software-defined environments in order to combine data, computing, and communication resources. In addition, we show how this proposed architecture enables the execution of data-driven workflows on top of these resources.

Cloud computing has generally involved the use of specialist data centres to support computation and data storage at a central site (or a limited number of sites). The motivation for this has come from the need to provide economies of scale (and subsequent reduction in cost) for supporting large scale computation for multiple user applications over (generally) a shared, multi-tenancy infrastructure. The use of such infrastructures requires moving data to a central location (data may be pre-staged to such a location prior to processing using terrestrial delivery channels and does not always require the use of a network-based transfer), undertaking processing on the data, and subsequently enabling users to download results of analysis. We extend this model using software defined networks (SDNs), whereby capability within the network can be used to support in-transit processing while data is in movement from source to destination. Using a smart building infrastructure scenario, consisting of sensors and actuators embedded within a built environment, we describe how an SDN-based architecture can be used to support real time data processing. This significantly influences the processing times to support energy optimisation of the building and reduces costs. We describe an architecture for such a distributed, multi-layered Cloud system and discuss a prototype that has been implemented using the CometCloud system, deployed across three sites in the UK and the US. Wevalidate the prototype using data from sensors within a Sports facility and making use of EnergyPlus.