A Dataflow Application Deployment Strategy for Hierarchical Networks
| Autor: | Satoshi Tanita, Shintaro Ishihara, Toyokazu Akiyama |
|---|---|
| Rok vydání: | 2019 |
| Předmět: |
050101 languages & linguistics
business.industry Dataflow Computer science Distributed computing 05 social sciences Cloud computing 02 engineering and technology Software Software deployment Component (UML) 0202 electrical engineering electronic engineering information engineering 020201 artificial intelligence & image processing 0501 psychology and cognitive sciences Use case business Component placement Edge computing |
| Zdroj: | COMPSAC (1) |
| DOI: | 10.1109/compsac.2019.00055 |
| Popis: | Massive dataflows are generated from IoT devices for sensing and controlling real-world environments. At the same time, reducing network traffic and delay utilizing hierarchical network characteristics including edge resources becomes important for maintaining long-running IoT application costs. A dataflow platform considering hierarchical networks is an ongoing research area, and it enables application developers to develop a dataflow application which usually consists of loosely coupled reusable software components for easy and flexible developments. However, the existing dataflow platform does not mention component management including software deployment and placement selection over the hierarchical network. In this paper, we proposed a component management method using a major open source product and a component deployment layer selection method considering network and resource conditions. During the definition of the component placement problem, we also extracted parameters required to estimate component deployment costs. Furthermore, we implemented a simple dataflow application based on two use cases and confirmed the validity of the proposed method. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|