1,720,978 research outputs found
Handling Data Handoff of AI-based Applications in Edge Computing Systems
Full text linkEdge computing aims at better supporting low-latency applications. One of its key techniques is computation offloading, the process that outsources computing tasks from resourced-constrained mobile devices and moves them to edge data centers. In this paper, we tackle an emerging problem within the umbrella of computation offloading, i.e., migration of offloaded inference tasks of Artificial Intelligence (AI) trained models. Such context tailors migration aspects of data-sensitive services where i) the value of the updates is inversely proportional to the data age and ii) outage is highly detrimental to accuracy. To tackle this challenge, we propose Mobile Edge Data-handoff (MED) a framework able to relocate inference or online training tasks from one edge datacenter to another by moving only the necessary data to minimize any accuracy drop during the process. We implemented MED in a well-known edge computing emulator, openLEON, and experimentally verified its performance with an AI-based Industry 4.0 application that forecasts the gas flow in a chemical plant. For our experiments, we use a real, open-source dataset that contains sensors readings. Collected results show that MED, employing proactive data handoff algorithms, is able to minimize the packet loss during the handoff thereby providing guarantees on the inference accuracy
A Practical way to Handle Service Migration of ML-based Applications in Industrial Analytics
No full textNowadays, Machine learning (ML) plays a significant role in Industrial Analytics. It enables predictive analytics, and helps uncovering essential insights to transform industries. As a result, real-time data analytics has become an essential requirement for industrial engineering jobs. Edge computing enables local intelligence and real-time analytics that are key for industry processes to take autonomous decisions locally at the edge of the network. However, outages in edge datacenters can jeopardize the whole plant security. In this paper, we proposed a practical approach to effectively handling service and data migration of ML-based applications in Industrial Analytics scenarios in the presence of a lack of computing resources at the edge. We argue that in this context the value of data is inversely proportional to their age and is very important to work with fresher data. In this paper, we describe our architectural approach for service and data handoff and show a predictive diagnostics case study deployed in an edge-enabled IIoT infrastructure. We evaluate our proposed approach in terms of drop of accuracy in a well-known edge computing emulator, i.e., openLEON. The experimental results show the benefit of our solution with respect to standard techniques
xSTART: xApp Simulated Evaluation Environment for Developers
Full text linkThe advent of the Open Radio Access Network (Open RAN) in 5G, as delineated in the standards introduced by 3GPP and O-RAN Alliance, posed a pivotal shift in revolutionizing the telecommunications landscape. Central to this transformation is the Open RAN control-plane architecture. Control plane encompasses containerized network functions, operating as intelligent controllers for RAN nodes and resources with non-Real-Time and near-Real-Time constraints. These functions, called Near Real-Time RAN Intelligent Controller (Near-RT RIC) and Non Real-Time RAN Intelligent Controller (Non-RT RIC), include operators-defined applications, respectively named xApps and rApps, serving as common cloud-native applications controlling and optimizing Open RAN elements. Understanding the entire development process and runtime behaviour for these applications becomes an essential prerequisite for providing support for next-generation networks. However, the lack of ready-to-use tools that allow developers to test and evaluate their applications in simulated and real environments makes it difficult to study and experiment with xApps. In this work, we propose xSTART, a ready-to-deploy environment based on Docker technology that allows xApp developers to quickly deploy and incorporate their xApps in a simulated ns-3 environment to test and evaluate their functionalities. Then, we evaluate an IIoT Network use-case scenario showing a machine learning (ML)-based xApp applied to the O-RAN environment. The results compare the use of different ML techniques and show the correct behavior of the simulation. Finally, we make xSTART available to the community to ease the development and evaluation of future xApps
Advanced and Future Network Access Technologies for the Metaverse
No full textThe Metaverse, a fully immersive, shared, and persistent 3D virtual space integrating virtual and real-world environments, is gaining significant attention. This chapter focuses on the crucial role of the access infrastructure and of the employed network protocols in supporting the Metaverse, with a specific emphasis on Edge Computing and multi-access Edge Computing (MEC) paradigms, as well as network acceleration technologies. The Edge Computing and MEC paradigms, bringing computational capabilities closer to the network edge, further enhance processing capabilities and enable ultra-low latency. Next-generation networks, such as beyond 5G and 6G, are required to provide distributed orchestration and management capabilities, integrating physical and virtual computing resources. Network protocols, such as 5G/6G, Wi-Fi 6, and TSN, play a crucial role in efficient and reliable communication within the Metaverse while networking acceleration technologies, both software-based (DPDK, XDP) and hardware-based (RDMA), enhance performance and enable ultra-reliable low-latency communication. The chapter also introduces SELENE, a solution designed to provide a technology-agnostic middleware application programming interface (API). SELENE allows developers to specify their quality of service (QoS) communication requirements, while dynamically selecting the most suitable acceleration technology based on the hosting edge node. Additionally, the chapter demonstrates the development of SELENE-based applications, such as an image streaming framework, with minimal code complexity. Extensive performance evaluations reveal that SELENE introduces negligible ns-scale overhead to the underlying network acceleration technologies, further highlighting its effectiveness in supporting the Metaverse demanding requirements and facilitating seamless and immersive experiences within this virtual realm
Software Defined Networking for Quality-aware Management of Multi-hop Spontaneous Networks
No full textThe Software Defined Networking (SDN) approach has recently demonstrated its effectiveness in simplifying the dynamic management of networking capabilities of infrastructure environments such as datacenters, e.g., by greatly enhancing the flexibility of dispatching features provided by industrial-grade switches. Inspired by the previous and more traditional scenario, we propose SDN adoption in infrastructure-less distributed multi-hop spontaneous networks based on the impromptu collaboration of fixed/mobile devices, with the goal of significantly improving the Quality of Service perceived by final users. To this purpose, the paper outlines our primary guidelines and reference architecture to support quality-aware packet dispatching. In particular, we present how collaborative nodes can exploit the SDN approach to appropriately manage the quality of different traffic flows, by avoiding undesired interference and by taking into consideration network capabilities/conditions and application-level requirements
On the Efficiency of Service and Data Handoff Protocols in Edge Computing Systems
No full textThe Multi-access Edge Computing (MEC) enables a new layer of edge middleboxes, acting as local proxies with virtualized resources deployed at edge localities. To support scalable, low-latency, and locally managed service provisioning, MEC relies on computation offloading, the process that outsources computing tasks from resourced constrained mobile devices and moves it to edge data centers. In this paper, we tackle a specific sub-problem within the umbrella of computation offloading. We argue that it is convenient to migrate a service because of the lack of computing resources in the anchor edge data center even if a device, such as industrial IoT devices, is not moving. In this paper, we extensively evaluate the efficiency of data and service handoff protocols. Specifically, we thoroughly assess protocols, that we designed in our past work, in a well-known edge computing emulator, i.e., openLEON. These protocols migrate data and service either in a reactive fashion, i.e., upon realizing of resource exhaustion, or proactively, i.e., beforehand to swiftly minimize the downtime. We experimentally verify their performance for a typical MEC use case, i.e., video. Our results show that by being proactive, the service interruption downtime reduces by a factor of 4 times
5G-Kube: Complex Telco Core Infrastructure Deployment Made Low-Cost
Full text linkNetwork Function Virtualization (NFV) along with Software Defined Networking (SDN) have brought an evolution in telecommunications laying out the bases for 5G networks and its softwarization. Accordingly, new implementations of telecom standards, such as the 3GPP 5G Core, are defined as fully-virtualized infrastructures consisting of different components and leveraging a cloud-native approach. At the same time, standard-oriented solutions, such as ETSI Management and Orchestration (MANO), have emerged to master the complexity of Virtualized Network Functions (VNFs) orchestration, including 5G Core VNFs. While MANO operates at the NFV level, it also leverages existing cloud infrastructures for the deployment of VNFs by interoperating with resource orchestrators at the cloud level. From the business perspective, that requires telco operators to interact with different technology providers, from NFV/MANO software producers to cloud computing providers, and to hire technicians proficient in the technologies of both telco and computing worlds, that are a rather difficult human resourcing to find. The main claim of the article is that the Development and Operations (DevOps) tools in the IT world are mature enough to leverage them directly in the telco world, without superimposing other interlaced standard/software. That allows to significantly reduce OPEX cost of complex telco infrastructures by supporting all needed automation and by avoiding the combined use of (too) complex layered standards/software stacks, such as in the case of MANO. Accordingly, in this article, we leverage container-based technologies and Kubernetes to design and evaluate a novel deployment approach, called 5G-Kube, for softwarized 5G core networks. 5G-Kube, which is openly to the community, has been also evaluated in two different use cases of the 5G Core and Kubernetes deployment fitting, namely, Industry 4.0 and Smart Cities
DRIVE: Discovery seRvice for fully-Integrated 5G enVironmEnt in the IoT
No full textA lot of research is being carried out about Internet of Things (IoT) and in the following years it will emerge more and more in our lives. Furthermore, with the advent of future fully-integrated 5G networks, new constraints need to be satisfied such as ultra-reliability and low-latency. With the help of Fog computing and the Multi-access Edge Computing (MEC) framework, services can be offered to the end users in a fast and practical way. Our work presents DRIVE a framework for service discovery in a 5G environment. However, in order to guarantee dynamic distribution and best management of services, we plan to deploy those services as container (e.g. Docker container). Moreover, we propose distribution of edge services at three different layer of communication: Application, Service, and Communication Layer. Given the above considerations, we propose an edge node, placed at the edge of network, that acts as the «brain» and take over the computation. The main innovative elements of the proposed framework, compared to the existing literature, include the possibility to select the working layer, the dynamic reconfiguration of the edge node and the field experimental results about the performance achieved by our solution over rapidly deployable environments with resourcelimited edge nodes such as Raspberry Pi devices
- …
