Yoshiaki Kiriha

An automatic generation of management information base (MIB) for OSI based network management system

The authors propose an automatic generation of a management information base (MIB), which forms one part of the total NMS (network management system) development support tools. The proposed MIB has capabilities to manage the management information distributed over a network, and gives a convenient object-oriented interface for application programs to handle the distributed information in a transparent fashion. MIB libraries and tools have been developed, part of which can be used as the application interfaces, while others are used internally in the MIB. These libraries and tools cooperatively make possible the automatic generation of MIB. The implementation of the proposed MIB has been done, and it is being applied to real NMS products.<<ETX>>

Concurrent network management system using distributed processing techniques

This paper proposes a concurrent network management system which can integrate various management applications in an efficient way. The proposed system provides both techniques and tools to attain concurrent execution of management applications in a distributed processing environment. For instance, each concurrently executed management application has its own communication module, named the virtual object, that interacts with other applications. The virtual object as a much enhanced version of the well-known RPC stub is capable of managing application connections and communicated data structures, or handling communication errors. The efficiency of integration is further enhanced through the use of a virtual object generator, which automatically generates the virtual object program codes. The effectiveness of the proposed system has been tested and confirmed in a newly developed experimental system that integrates our management applications such as network design, control, and diagnosis.<<ETX>>

Open Programmable Layer-3 Networking

This paper discusses high speed (multi-gigabits/second) active networking techniques used to introduce active functionality into layer-3 processing in order to provide wide-ranging flexibility to internetworking. We propose here StreamCode, a compact object code for layer-3 programming, a StreamCode Processor that achieves a high-speed execution of the object code, and an architecture of StreamCode-based high-speed active networking node. The processor has a unique instruction fetch mechanism to prevail over promiscuous instructions and data flows inside the processor, and has resource management function to execute StreamCode programs safely. Our FPGA based prototype system with sample applications confirmed the feasibility of proposed StreamCode based programmable Layer-3 networking.

Accelerating Code Deployment on Active Networks

Active networks enable their users to specify how each packet is processed on network nodes. One of the essential techniques for active networks is the programmable node approach, which enables network nodes to evolve their packet processing functions by loading new software components into the nodes. Possible component loading strategies include demand loading. It reduces the usage of node resources and localizes possible problems, although the component loading time defers packet processing. This paper discusses the component loading time of the demand loading strategy, and proposes a scheme to shorten the loading time by masking the propagation delay of components. We have implemented the scheme pre-supplying on a Java-based system and have evaluated its effectiveness. The result shows that the proposed scheme shortens the loading time of a test component by as much as 70%.

A Self-Configuring Data Caching Architecture Based on Active Networking Techniques

This paper presents the design of a new Web cache architecture that uses the active network capabilities to provide a solution to cache dynamic data throughout the network. In our proposal, objects, viewed with the smal-lest granularity, are cached associated with a timestamp. But, instead of considering dates individually, we define some time classes which specify the level of objects time-sensitiveness. Each intermediate node is specialised into a unique time class according to its location within the network and caches dynamic data which belongs to corresponding time class. Nodes are shared out among two types (manager and cache) and are bound together to define a hierarchical time-sensitive cache tree, the timestamp tree. In our proposed cache architecture, the timestamp tree is automatically reconfigured according to users access history, applications load and network conditions. To achieve such self-configuring cache architecture, we have actually designed five types of capsules.

Hardware-oriented TMN systems

The TMN (Telecommunication Management Network) standard is an interoperable framework for the management of telecommunication networks and services. While its capability and functionality are recognized around the world, many networks and network elements lack the computing resources required to implement a complex TMN functionality. A number of reasonable approaches to this problem have been reported, such as a lightweight protocol, Q-adaptation, and Q-addition. However, these approaches suffer from the problem of a required trade-off between system overhead and TMN functionality. The hardware-oriented TMN proposed by the authors is one possible answer to this problem.

Policy transition mechanism: a new approach to multi-mode management

A policy management technology is one of promising technologies for the management of networking systems, which are getting complex and difficult to manage. Conventional policy languages provide a flat policy mechanism, under which all policy rules are always active. We argue that not all policy rules should be active in all situations. Some are defined for a certain situation and others for a different situation. For example, the set of active policy rules at office hours are different from those at off-time. We propose a policy transition mechanism, which modifies active policy groups, depending on a situation. As a result, we can achieve management depending on a change in the situation while the flat policy mechanism cannot.

Policy-based cooperation of services in ubiquitous environments

Various kinds of nodes, including cellular telephones and information appliances, are to become popular and are expected to provide a variety of services. Cooperation of these services will result in more convenient services than keeping them isolated would. A ubiquitous network is characterized by changeable system configurations. Because of this and the fact that a node is so frequently connected to and disconnected from the network, the global cooperation of services is difficult to describe in flow languages such as Web Services Flow Language (WSFL). One of the solutions to this problem is a policy technology. A policy attached to a node can be added or removed when the node is connected or disconnected. The policies can re-configure a changed system.

Node inquiry function for strategic front-end placement

In a front-end system, mirror servers called front-ends are deployed on nodes within the Internet to enhance the quality of on-line services. An on-line service provider can control its service quality by changing the number of front-ends. In addition to the number, the placement of front-ends is also important for the provider to maximize the profitability. This paper proposes a node inquiry function (NIF) that provides on-line service providers with strategic front-end placement plans. The NIF-tailored placement plan enables a service provider to obtain the best possible front-end distribution that maximizes investment efficiency. The NIF also helps them to find those heterogeneous nodes that can run their front-ends. The placement algorithm calculates in polynomial time and an experiment shows better approximation than a greedy approach.

Multi-tier Architecture for Service Creation, Deployment and Management

Customers now expect advanced IP services like video conferencing and application rental become customizable and easy to use. So, network service providers must be able to dynamically deploy such Internet services fitting the requirements of service customers while avoiding the expensive solution of over-provisioning the service. However, the current Internet technologies are lacking when it comes to developing and deploying new services, integrating together service offerings. This paper presents a service architecture which enables the creation, on-demand deployment and management of advanced IP services using the active network capabilities. In our approach, each business IP service is considered as a composition of basic service components which can be shared among different business IP services. Upon the requests of service users and Service Definition information, initialized by providers, mobile code is downloaded into active nodes and installed in the network as new basic service components. Finally, this paper presents the implementation of a Content Delivery Service using our proposed service architecture.