{"title":"EdgePub:用于远边缘的自适应分布式MQTT代理覆盖","authors":"Chamseddine Bouallegue, Julien Gascon-Samson","doi":"10.1109/FMEC57183.2022.10062858","DOIUrl":null,"url":null,"abstract":"The MQTT protocol, based on a topic-based publish/subscribe paradigm, plays an important role in the Internet of Things (IoT), as it enables flexible and highly decoupled communications between the different entities of an IoT system. Further, several IoT applications require low latencies (e.g., tele-surgery, connected vehicles) – hence, a centralized MQTT (publish/subscribe) infrastructure can be impractical. In this paper, we present EdgePub, a dynamic, highly distributed, and self-adaptable edge-based publish/subscribe middleware that provides drop-in compatibility with existing MQTT-based client applications and brokers. EdgePub transparently builds a one-hop dissemination overlay over embedded MQTT brokers deployed at the far-edge (i.e., on the client devices themselves), and provides a load balancing strategy that continuously minimizes the average publication latency, while ensuring that the bandwidth constraints of the edge client devices are met. We provide an implementation through the form of an MQTT.js-compatible Node.JS library, and we evaluate EdgePub over different deployment scenarios (i.e., local to world-wide deployments), over a test-bed of Raspberry Pi devices. We report 18%-77% lower average latencies compared to centralized edge and cloud-based deployments, without exceeding the limited bandwidth constraints of the edge brokers.","PeriodicalId":129184,"journal":{"name":"2022 Seventh International Conference on Fog and Mobile Edge Computing (FMEC)","volume":"36 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2022-12-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"EdgePub: A Self-Adaptable Distributed MQTT Broker Overlay for the Far-Edge\",\"authors\":\"Chamseddine Bouallegue, Julien Gascon-Samson\",\"doi\":\"10.1109/FMEC57183.2022.10062858\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"The MQTT protocol, based on a topic-based publish/subscribe paradigm, plays an important role in the Internet of Things (IoT), as it enables flexible and highly decoupled communications between the different entities of an IoT system. Further, several IoT applications require low latencies (e.g., tele-surgery, connected vehicles) – hence, a centralized MQTT (publish/subscribe) infrastructure can be impractical. In this paper, we present EdgePub, a dynamic, highly distributed, and self-adaptable edge-based publish/subscribe middleware that provides drop-in compatibility with existing MQTT-based client applications and brokers. EdgePub transparently builds a one-hop dissemination overlay over embedded MQTT brokers deployed at the far-edge (i.e., on the client devices themselves), and provides a load balancing strategy that continuously minimizes the average publication latency, while ensuring that the bandwidth constraints of the edge client devices are met. We provide an implementation through the form of an MQTT.js-compatible Node.JS library, and we evaluate EdgePub over different deployment scenarios (i.e., local to world-wide deployments), over a test-bed of Raspberry Pi devices. We report 18%-77% lower average latencies compared to centralized edge and cloud-based deployments, without exceeding the limited bandwidth constraints of the edge brokers.\",\"PeriodicalId\":129184,\"journal\":{\"name\":\"2022 Seventh International Conference on Fog and Mobile Edge Computing (FMEC)\",\"volume\":\"36 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2022-12-12\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2022 Seventh International Conference on Fog and Mobile Edge Computing (FMEC)\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/FMEC57183.2022.10062858\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2022 Seventh International Conference on Fog and Mobile Edge Computing (FMEC)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/FMEC57183.2022.10062858","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
EdgePub: A Self-Adaptable Distributed MQTT Broker Overlay for the Far-Edge
The MQTT protocol, based on a topic-based publish/subscribe paradigm, plays an important role in the Internet of Things (IoT), as it enables flexible and highly decoupled communications between the different entities of an IoT system. Further, several IoT applications require low latencies (e.g., tele-surgery, connected vehicles) – hence, a centralized MQTT (publish/subscribe) infrastructure can be impractical. In this paper, we present EdgePub, a dynamic, highly distributed, and self-adaptable edge-based publish/subscribe middleware that provides drop-in compatibility with existing MQTT-based client applications and brokers. EdgePub transparently builds a one-hop dissemination overlay over embedded MQTT brokers deployed at the far-edge (i.e., on the client devices themselves), and provides a load balancing strategy that continuously minimizes the average publication latency, while ensuring that the bandwidth constraints of the edge client devices are met. We provide an implementation through the form of an MQTT.js-compatible Node.JS library, and we evaluate EdgePub over different deployment scenarios (i.e., local to world-wide deployments), over a test-bed of Raspberry Pi devices. We report 18%-77% lower average latencies compared to centralized edge and cloud-based deployments, without exceeding the limited bandwidth constraints of the edge brokers.