Jeman PARK¹, Inwhee JOE¹, Won-Tae KIM^2
1 Department of Electronics Computer Engineering, Hanyang University
17 Haengdang-dong, Seongdong-gu, Seoul, Korea
mirrsam@hanyang.ac.kr; iwjoe@hanyang.ac.kr
² CPS Research Team, ETRI
138 Gajeongno, Yuseong-gu, Daejeon, Korea
wtkim@etri.re.kr

Abstract: A Cyber Physical System (CPS) is an autonomous embedded system based on high reliability with real-time control of distributed physical systems through wired/wireless networks. There is usually large volume of data which needs to be delivered to right places at the right time. In addition, large number of controllers in the automation and control systems are usually distributed which increases the complexity that there needs to be more point-to-point Ethernet-connections in the network. Because the controllers in the network may share control data and interact with each other from different communication protocols, including higher level operator systems. The interdependencies between these nodes may potentially create a complex architecture of the network in the distributed system especially if the point-to-point connection needs to be established. Publish-subscribe model shows some appealing properties, such as connectionless and multicast, that can be used to reduce some of the visible complexity in the software systems. Data distribution middleware for CPS should be based on a data-centric approach and guarantee real-time performance. In this regard, OMG’s DDS is the best proximity middleware. RTPS (Real-Time Publish/Subscribe) is proposed for real-time service discovery in DDS. However, legacy discovery protocols cannot completely support the CPS system with a large-scale network (approx. 100,000 entities) like a warship, because service discovery messages are proportional to the square of the number of participants in RTPS. This paper proposes a scalable and fast service discovery protocol with improved discovery time for large-scale cyber physical systems based on the boot-strap algorithm and adaptive PDP message period. As a result, the proposed protocol improves reliability and real-time for service discovery in cyber physical systems. In this paper, mathematical analysis and test-bed experiments are conducted to evaluate the performance of the proposed protocol. Consequently, mathematical analysis and test-bed experiments provide almost identical results. The performance results prove that our protocol works to scale for large-scale CPS networks by minimizing the discovery time as well as traffic simultaneously.

Keywords: Service Discovery, CPS, Optimal Discovery Time, Boot-strap algorithm, Adaptive period.

>Full text
CITE THIS PAPER AS:
Jeman PARK, Inwhee JOE, Won-Tae KIM, An Efficient Discovery Protocol of Large-Scale CPS Middleware for Real-Time Control System, Studies in Informatics and Control, ISSN 1220-1766, vol. 23 (1), pp. 23-30, 2014. https://doi.org/10.24846/v23i1y201403

1. Introduction

The integration of physical systems and processes with networked computing has led to the emergence of a new generation of engineered systems [1]. A cyber physical system (CPS) is a computing system that interacts with physical processes [2, 3, 4]. Recent years have seen a growing development of embedded systems. An embedded system is a special-purpose computer system designed to perform dedicated functions for specific physical devices. They are generally assumed to be standalone. The last decade has been an explosive development period of the Internet all over the world. The advent of the Internet has raised numerous research challenges and has brought arguably the greatest technological impact on society, allowing a new communication paradigm. A CPS will bring other research challenges and social impacts, suggesting yet another new paradigm of controlling information and physical devices. There are various other research issues to consider, including real-time, privacy, reliability etc. To effectively manage such data deliver. CPS may need data-centric middleware, such as the Data Distribution Service (DDS) specification [5, 6], which makes it easy to deal with complicated data distribution. A data-centric middleware model makes it easier to address such requirements of CPS applications. The OMG Data Distribution Service (DDS) specification defines publish-subscribe middleware, which enables CPS nodes to satisfy user’s QoS requirements. At the core of DDS is the Data-Centric Publish-Subscribe (DCPS) model, which defines standard interfaces that allow applications running on heterogeneous systems to read/write data to/from a global data space in a networked system. Also, Real-Time Publish Subscribe (RTPS) support interoperability and real-time transmission among different types of DDS in a networked system.

The RTPS specification [7] defines a discovery service for domain participants called the Participant Discovery Protocol (PDP) and another for matching data readers and data writers called the Endpoint Discovery Protocol (EDP).