A Survey on M2M Service Networks
The number of industrial applications relying on the Machine to Machine (M2M) services exposed from physical world has been increasing in recent years. Such M2M services enable communication of devices with the core processes of companies. However, there is a big challenge related to complexity and to application-specific M2M systems called -vertical silos‖. This paper focuses on reviewing the technologies of M2M service networks and discussing approaches from the perspectives of M2M
... OPEN ACCESS Computers 2014, 3 131 and services, M2M communication and M2M security. Finally, a discussion on technologies and approaches potentially enabling future autonomic M2M service networks are provided. According to our conclusions, it is seen that clear definition of the architectural principles is needed to solve the -vertical silo‖ problem and then, proceeding towards enabling autonomic capabilities for solving complexity problem appears feasible. Several areas of future research have been identified, e.g., autonomic information based services, optimization of communications with limited capability devices, real-time messaging, creation of trust and end to end security, adaptability, reliability, performance, interoperability, and maintenance. Keywords: machine to machine communication; Internet of things; cyber-physical systems Computers 2014, 3 132 special access rights for each specific system, resulting in vendor specific closed systems. This has caused problems, for example in residential home environments, and prevented the emergence of home automation to a large extent. Smart grid solutions cannot interoperate with infrastructure and buildings/homes, even if it would be strongly required to reach higher level energy efficiency. Therefore, it is observed here that the technological complexity and vertical M2M silos are the cause of a grand research challenge for the development of a modern ecosystem. It is assumed here that the referred challenge is so fundamental that it requires novel approaches for the system architectures and application of novel paradigms. It is seen that Internet -IP everywhere‖ cannot alone solve the problem with industrial M2M systems. Something more is needed, such as horizontal approach and autonomic computing, which may have potential to create basis for solving this sizable challenge. Most of the existing vertical M2M solutions have difficulty in scaling, and therefore enabling the horizontal model is important for realizing embedded M2M  . Autonomic computing is a concept inspired from biological systems that aims to develop systems capable of automatic management for solving the complexity problem [4, 5] . There are multiple attempts to build such autonomic systems, e.g., [6, 7] , and different design approaches such as externalization and internalization  . According to IBM, the evolutionary path to autonomic computing is represented by five levels  , starting from basic, through managed, predictive, adaptive and finally to autonomic. Such approaches as software agents , active networks  and policy-based systems  have been developed to automate the management tasks to achieve higher response times with less management cost. Attempts have been made to solve the communication complexity problem by using be solved by built-in mechanisms to let systems manage their own communications  . The self-management capability aims to solve the rapidly growing complexity of computing system management and enable dynamic future growth of the system. There are four properties that enable autonomic capabilities in a system: self-configuration, self-optimization, self-healing and self-protection . Self-Configuration is the ability of the system to perform configurations according to pre-defined high level policies and seamlessly adapt to change caused by automatic configurations. Self-Optimization is the ability of the system to continuously monitor and control resources to improve performance and efficiency. Self-Healing is the ability of the system to automatically detect, diagnose and repair faults. Self-Protection is the ability of the system to pro-actively identify and protect itself from malicious attacks or cascading failures that are not corrected by self-healing measures. An autonomic system should satisfy these properties through a reactive or a proactive behavior. A reactive autonomic system tries to detect problems or meaningful events and then finds an appropriate action or solution after the problem has been already detected. A proactive autonomic system  uses preventive measures to maintain, improve or optimize the system performance. Those measures are based on the analysis of the current state, anticipated events and the predicted system reaction to them. Thus, we review here the technologies to evaluate their capabilities to solve the grand challenge in the context of M2M service networks. The approaches for M2M service networks are first discussed in Chapter 2 and the taxonomy for the review is defined. Then, the related M2M standards and technologies for each category are reviewed as follows: Chapter 3 M2M information and services, Chapter 4 M2M communication and Chapter 5 M2M security. After the review, a discussion is provided in Chapter 6, and finally concluding remarks are provided in Chapter 7. Computers 2014, 3 133 2. M2M Service Networks 2.1. M2M System A M2M system usually consists of a set of M2M asset devices attached in a M2M capillary network, a kind of M2M gateway, M2M communication infrastructure and a set of M2M services and applications as shown on Figure 1. The red line represents a typical scenario related to the remote monitoring and control process in M2M service networks. Such a scenario requires operation of the complete M2M system including the functionalities of M2M information and service, M2M communication and M2M security, which are therefore included in a sensible level into this survey. Figure 1. A view of a Machine to Machine (M2M) system. M2M services and applications can be divided further to multiple levels such as, e.g., information, service platform and applications. The applications are usually domain specific business logic, high level management of the M2M system, devices, and domain specific information. The information level usually contains information management services and exchange transactions between the stakeholders of the system. A standardized common information model can enable smooth information exchange and business interactions between stakeholders of the domain. An example of this kind of common information model is CIM standardized for energy grid  . M2M Service Platform includes service solutions and frameworks, which may be applied in multiple domains. The service solutions can contain generic service elements such as e.g., event notification, environment monitoring, service discovery and delivery, generic profiling, access control, generic storage and device management  . ETSI M2M has specified a set of service capabilities related to application, communication, reachability, addressing and repository, remote management, security, history and data containers  . A standardized M2M service platform could enable smooth application development and interaction between service platforms of different vendors, and services of M2M asset devices to be applied in multiple cases. M2M communication infrastructure contains heterogeneous networks including local M2M asset network (M2M capillary network) such as personal/body area network, vehicular network and wireless sensor network (WSN), and Internet including various overlay networks. The overlay network can logically connect the M2M asset devices, M2M gateways, infrastructure servers and user with each Dynamic configuration and discovery methods have been applied in large scale within peer to peer (P2P) content delivery systems, e.g., Napster, KaZaA, Gnutella, Morpheus and BitTorrent  . However, content delivery differs essentially from the discovery of services exposed from M2M asset devices. There are some service discovery systems targeted to local scale and embedded devices such as UPnP and Bluetooth SDP. However, there is lack of solutions for large-scale wireless and hybrid networks  . Overlay networking is based on the virtual communication layer, which is built on top of other transport media and/or physical network  . Overlay networking has been applied to improve the robustness and availability of Internet paths between hosts (e.g., MIT RON), enable smooth transition to the improved technology (e.g., 6Bone) or to reduce network load by peer assisted data delivery (e.g., BitTorrent). Overlays have been used to route control messages and connect different entities (e.g., SIP and XMPP), and also to implement data forwarding and dissemination (e.g., Chord, Tapestry, and Pastry). Another motivation for overlay networking arises from security challenges. For example, Virtual Private Ad Hoc Networking (VPAN) has been developed to create virtual overlay networks between trusted IP capable devices  . In an M2M system, M2M asset devices, network and users can be mobile. In these conditions, information exchange with the M2M asset device need to solve such challenges as unreliable communication channels, temporal presence, limited power and computing capabilities. Solutions for wireless sensors networks in some specific domain and application have been developed  . Usually, these solutions have vendor dependent optimized solutions for computing platform, radio technology, communication and services. However, there is at least one potential emerging standard for low power communication between wireless sensors called as Bluetooth 4.0 , Bluetooth Smart, which may be applicable for multiple domains. However, there is still a need for generic standard based communication and service protocols working with limited capability embedded M2M assets and mobile gateways. Discussion The M2M standardization approaches demonstrate the heterogeneity of the M2M service networks area. The area is very large; our aim is to review the available technological solutions at least from the following perspectives: M2M Information and services, M2M communication and M2M security. In this review we categorize the available technologies into these perspectives. Within each category, the key contribution of each technology is analyzed. After the review, the potential contributions of the technologies to support horizontal and autonomic M2M are discussed. M2M Information and Services M2M Information and Service technologies are reviewed in this chapter. After the review, two potential technologies: sensor web enablement (SWE) and ETSI M2M service capability layer are shortly overviewed. Finally, autonomic features and related technologies are discussed. M2M Information and Service Technologies The selected set of M2M information and service level technologies are reviewed in Table 2 , and some of them are shortly analyzed in the following.