Kenichi Kamada

Security architecture for control networks using IPsec and KINK

There are many kinds of control networks which have been used in various nonIP network areas, such as BA (building automation), FA (factory automation) and PA (process automation). These do not incorporate reasonable security mechanisms as they have been mainly used for closed networks. Recently the security of control networks is becoming important because of the popularization of the Internet, the deployment of wireless technologies and the security requirements of such infrastructures. Control networks require security mechanisms which 1) enable end-to-end security that do not depend upon specific network topology, 2) work with multiple control network technologies, and 3) are suited to small embedded devices commonly used in control networks. This paper shows security mechanisms which can meet the above requirements, assuming that IP is applied to the control networks.

A translation method between 802.15.4 nodes and IPv6 nodes

There are many kinds of control networks based on non-IP, such as BA (building automation), FA (factory automation) and PA (process automation). The IPv6 and wireless technologies are expected to improve those networks. The IEEE 802.15.4 is a candidate for wireless technology in control networks because of its feature, e.g. low power consumption and small implementation. Seamless communication between both technologies is important, though it is not easy for IEEE 802.15.4 packet to carry IPv6 packet due to its limited features. This paper shows a translation method between IPv6 nodes and IEEE 802.15.4 nodes

Network Simulation for Control Systems

Maintaining the quality of network traffic of a control system is crucial; however, this is becoming much more challenging with the emerging IP-based technologies. Operators must constantly monitor the state of traffic flow in the control system and decide whether the system is functioning properly. To achieve this, we must know the characteristics of normal traffic and extract boundary conditions, i.e., thresholds, at which the quality of services begins to deteriorate to intolerable levels; this helps distinguish safe traffic from an undesirable one. The authors propose to use a network simulation technique to analyze the traffic behavior of IP-based control systems and construct flexible simulation architecture. We also evaluate the architecture model by implementing an example scenario that estimates the necessary network capacity for the experimental control system

Design and evaluation of a client-friendly cross-realm framework for Kerberos 5

One problem with the cross-realm operation of Kerberos 5 is that clients need to traverse the authentication paths between realms. Traversal is a burden for clients with limited resources. In this paper, the authors constructed a cross-realm framework, called the client-friendly cross-realm framework, which releases clients from traversal of realms. This framework enables a client to obtain a service ticket in one message exchange with its local key distribution center (KDC). On the other hand, the framework introduces overhead on KDCs. The authors implemented PKCROSS as a component of the framework and showed that the overhead is insignificant.

Industrial data-driven plant optimization modeling

In recent years, the importance of EnMS (Energy Management System) that can monitor the balance of supply and demand of energy, and can efficiently operate plant, is increasing. In order to improve the plant operation by appropriate introduction of EnMS, customers want precise pre-diagnostic. In order to calculate the improvement potential precisely, it is effective to optimization using a plant simulating model. However, since the advanced technical knowledge of equipment is required for the construction of the plant model, it is a problem to increase modeling man-hours. Therefore, we have developed Data Driven Plant Optimization Modeling that can automatically create equipment models from operating data. The greatest feature of this technology is that it can extract exhaustively equipment characteristic equations even if any operating data are assigned. This new modeling technology is able to reduce more than 80% of the conventional modeling man-hours and create high precision model for certain plant.