IOCD-based Semantic Enablement for Service-oriented M2M Architecture
Yan Li, Xuemei Hu, Yulin Zhang, Abdul-Wahid Mohammed, Yang Xu
School of Computer Science and Engineering
University of Electronic Science and Technology of China
Chengdu, Sichuan, P.R.China, 611731
xuyang@uestc.edu.cn
Abstract—To build a simple picture of M2M communication
under service-oriented M2M architecture, a key innovation
underlines the enablement of M2M device abstraction and
semantics support. However, state of the art structures of
M2M devices/software lack common M2M Service Layer for
interoperability. Devices/software rely heavily on specific API
implementation-oriented data services and the data definition
of their own. In this paper, we propose a generic M2M design
pattern called IOCD by referring to the SSO pattern used in
Semantic Sensor Network ontology to enable interoperability
among M2M devices and their services. This briefly abstracts
and describes machines from four key properties: input (I),
output (O), capability (C) and deployment (D). Based on the
design pattern, we extend these four properties to build our
IOCD design model. To enable IOCD semantic support, we
build our M2M semantic server under oneM2M architecture,
and extend the oneM2M Generic semantic functional model.
The M2M semantic server retrieves data from the oneM2M
platform, and semantically annotates and reasons on the data
according to IOCD design model.
Keywords-M2M, Semantic, Ontology, IOCD design pattern.
I. INTRODUCTION
Machine-to-Machine (M2M) technology is a hot research
area, and usage of M2M devices dominates in many realms
[1] [2]. Although the M2M market is much alive and grow-
ing, vertical M2M solutions dominate, and so lacks common
Service-oriented M2M Architecture for interoperability a-
mong large scale heterogeneous devices and services [3].
This inevitably impacts or even impedes large-scale M2M
deployment. Enabling the semantic support for service-
oriented M2M architecture is therefore key in this research
field. To this end, oneM2M [4] is the current standards
organisation seeking to integrate global M2M standards and
vertical industry applications.
Open Geospatial Consortium (OGC) has so far been
able to specify publicly available encoding and interface
standards, which allow developers create applications, plat-
forms, and products involving Web-connected devices so
as to make complex spatial information and services ac-
cessible and useful with all kinds of applications. In order
to add real-time sensor dimension to the Internet, Sensor
Web Enablement (SWE) makes the Web a unique and
revolutionary framework of open standards for exploiting
Web-connected sensors and sensor systems of all types.
From resource perspective, the IOT.A project [5] has also
proposed a resource model that captures different resource
types including sensor, actuator, processor, storage and tag.
This resource model is limited in representational scope, and
more resource types need to be defined in order to link up
with external ontologies. Semantic Sensor Network ontology
[6] represents a high-level schema model, which provides
sensor’s device description, capabilities, platform and related
attributes in the semantic webs and applications. However, it
does not include modeling aspects for features of actuators
and other M2M devices like RFID tags.
In this paper, we propose a semantic model–IOCD to en-
able interoperability among M2M devices and their services.
This is a more generic framework, which is not limited to
vertical applications, and describes the common properties
of M2M devices from four aspects: input, output, capability
and deployment. To achieve semantic support for service in
this approach, we also implement an M2M semantic server
under oneM2M standard functional model architecture based
on the IOCD model.
II. IOCD DESIGN PATTERN
The key requirement of providing common service is
to convert application demands into specific manipulations
for both applications and machines. In consequence, we
extend the SSO design pattern [7] to achieve a more generic
framework called IOCD design pattern.
Figure 1. IOCD design pattern
As shown in Fig.1, the IOCD design pattern has 4 core
elements to satisfy the essential requirement of modeling a
machine: input(I), output(O), capability(C) and deploymen-
t(D). This design pattern can reflects 3 main perspectives:
1) data: with a focus on the data flow and related metadata,