Serverless computing is becoming widely adoptedamong cloud providers, thus making increasingly popular theFunction-as-a-Service (FaaS) programming model, where thedevelopers realize services by packaging sequences of statelessfunction calls. The current technologies are very well suited todata centers, but cannot provide equally good performance indecentralized environments, such as edge computing systems,which are expected to be typical for Internet of Things (IoT)applications. In this paper, we fill this gap by proposing a frameworkfor efficient dispatching of stateless tasks to in-networkexecutors so as to minimize the response times while exhibitingshort- and long-term fairness, also leveraging information from avirtualized network infrastructure when available. Our solution isshown to be simple enough to be installed on devices with limitedcomputational capabilities, such as IoT gateways, especially whenusing a hierarchical forwarding extension. We evaluate theproposed platform by means of extensive emulation experimentswith a prototype implementation in realistic conditions. Theresults show that it is able to smoothly adapt to the mobilityof clients and to the variations of their service request patterns,while coping promptly with network congestion.
A Decentralized Framework for Serverless Edge Computing in the Internet of Things
C Cicconetti;M Conti;A Passarella
2020
Abstract
Serverless computing is becoming widely adoptedamong cloud providers, thus making increasingly popular theFunction-as-a-Service (FaaS) programming model, where thedevelopers realize services by packaging sequences of statelessfunction calls. The current technologies are very well suited todata centers, but cannot provide equally good performance indecentralized environments, such as edge computing systems,which are expected to be typical for Internet of Things (IoT)applications. In this paper, we fill this gap by proposing a frameworkfor efficient dispatching of stateless tasks to in-networkexecutors so as to minimize the response times while exhibitingshort- and long-term fairness, also leveraging information from avirtualized network infrastructure when available. Our solution isshown to be simple enough to be installed on devices with limitedcomputational capabilities, such as IoT gateways, especially whenusing a hierarchical forwarding extension. We evaluate theproposed platform by means of extensive emulation experimentswith a prototype implementation in realistic conditions. Theresults show that it is able to smoothly adapt to the mobilityof clients and to the variations of their service request patterns,while coping promptly with network congestion.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.