Session details: Theme: Information systems: SFECS - Sustainability of fog/edge computing systems track
Symposium on Applied Computing(2019)
摘要
ABSTRACTFog/Edge Computing paradigms are widely used in enterprises to address the emerging challenges of big data analysis, because of their underlying scalable, flexible and distributed data management schemes. The data centers in the Clouds are facing great challenges on the burden of the consequent increasing the amount of data to be man- aged and the additional requirements of location awareness and low latency at the edge of network necessary by smart cites and factories. These are the reasons why a centralized model cannot be an efficient solution for generated or required data by the IoT devices in those applications and there is the progressive shift towards fog nodes and smarted edge nodes mediating between the cloud and the IoT devices. The Fog/Edge computing paradigm is a decentralized model that transfers a part of low computing data analysis from the cloud to the intermediate (fog) nodes or the edges, performing only high computing tasks in the cloud. This new approach tries to minimize the three factors that negatively compromise the effective and efficient application of the Cloud computing to smart cities and factories, or similar application domains: the network bandwidth usage, decentralization of the data processing tasks and reduced response latency for clients (IoT devices). Fog/Edge computing is a hierarchical approach where the overall infrastructure is structured in multiple layers, each responsible of offering a good coordination and data management to the nodes at the lower layer. The lowest layer is usually composed of sensors and/or actuators that measure and/or control the environment or a given business process, implemented as mobile devices that are running a sensing/controlling application. In this case, combining Sustainable computing with Fog and Edge computing represents a new approach for increasing quality-of- service and efficiency of the system, creating the capability to present temporal and geo-coded information, and increasing innovation, and co-designing sustainable future large scale distributed systems. This new paradigm appears to offer a good approach in handling the scale factor of the data size, reducing the network bandwidth usage and the response latency of the system. In order to support specifically the Fog/Edge architectures, there is a need, for instance, of location-awareness and computation placement, replication and recovery. In many cases Edge resources would be required for both computation and data storage to address the time and locality constraints. There are multiple kinds of orchestration management solutions for virtualization in this type of architecture with different characteristics and drawbacks. This results in different restrictions for application definition, scalability, availability, load balancing and so on. Also, virtualization may be needed at multiple levels in a Fog/Edge architecture as it consists of the following levels of abstraction: at the sensing level we have the IoT devices/smart things, at the Edge level there are the gateways to a first collection and the data from the IoT devices and their preliminary processing, at the Fog level we have an additional data management layer, and at the Cloud level there is the compute/storage infrastructure with applications on top. Last, but not least, the energy efficiency is particularly important at the IoT and edge level since the devices may be equipped with a limited battery, possible difficult or impossible to be charged. So, optimizing the energy consumption is a must. To address several open research is- sues regarding sustainability of future Fog/Edge systems, this track aims at solicit contributions highlighting challenges, state-of-the-art, and solutions to a set of currently unresolved key questions including - but not limited to - performance, modeling, optimization, energy-efficiency, reliability, security, privacy and techno-economic aspects of Fog/Edge systems. Through addressing these concerns while understanding their impacts and limitations, technological advancements will be channeled toward more sustainable/efficient platforms for tomorrow's ever-connected systems.
更多查看译文
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络