Topic allocation method on edge servers for latency‐sensitive notification service

Abstract

The importance of real-time notification has been growing for social services and Intelligent Transporting System (ITS). As an advanced version of Pub/Sub systems, publish-process-subscribe systems, where published messages are spooled and processed on edge servers, have been proposed to achieve datadriven intelligent notifications. In this paper, we present a system that allows a topic to be managed on multiple edge servers so that messages are processed near the publishers, even when publishers spread over a wide area. Duplicating messages on geographically distributed servers could enable immediate notification to neighboring subscribers. However, the duplicated message spool may cause exhaustion of resources. We prepare a formal model of our publishprocess-subscribe system and formulate the topic allocation as an optimization problem under the resource constraints of edge servers. As the optimization problem is NP-hard, we propose heuristics leveraging the locality and the pub/sub relationships observed between clients to use the edge server resources efficiently. Our performance evaluation shows that our method reduces the delay to deliver notifications and the effectiveness of the strategy exploiting the relationships between clients.

Topic-based distributed publish-process-subscribe system with metrics on geographic distance and permissible delay

“…The above model has relaxed restrictions on publishers compared to the previous model [3]. In the previous model, each client is assigned to one edge server and publishes/subscribes messages about all topics via that server.…”

“…The trade-off between latency reduction and storage consumption makes it difficult to decide where to allocate brokers. In the previous study [3], we introduced a formal model of our publish-process-subscribe system and formulated the topic allocation problem as an optimization problem of the total delay. This paper introduces two metrics.…”

“…This section briefly shows the major variables used in our model and points out the difference from the previous study [3]. The set of clients acting as publishers or subscribers is denoted as C = {c 1 , c 2 , .…”

“…The publishprocess-subscribe system is an extension that collects and analyzes the data published on each topic and delivers notifications to the subscribers [2]. Our group has proposed a distributed publish-process-subscribe system for edge computing environments that allows multiple edge servers to handle a single topic [3]. Apache Kafka [4] also assigns multiple brokers to a single topic for fault tolerance and scalability but does not consider geographic location.…”