RINA: An Architecture for Policy-Based Dynamic Service Management

International Journal of Communication Systems, 2006

More and more applications in the Internet are requiring an intelligent service infrastructure to provide customized services. In this paper, we present an infrastructure, which can transparently and effectively provide customized active-services to end users and dynamically adapt to changing customized policies in large distributed heterogeneous environments. The infrastructure consists of two components: the policy agent and middleware box. Particularly, our technologies include: (1) Generic active-service based infrastructure, where the policy agent can integrate policies requested by applications, and middleware boxes can transparently execute services and (2) Distributed policy processing in the middleware box. We study two policy partitioning schemes to achieve conflict-free policies for distributed policy processing and guarantee the correctness of the policy execution. We conduct extensive performance evaluations on different schemes proposed. Our experimental results demonstrate that our policy partitioning schemes can effectively generate partition-capable and conflict-free policy sets. The evaluation results also show that distributed policy processing can achieve over 70% increase in performance/price ratio with proper assignment of the policy distribution degree compared to a purely centralized approach.

Journal of the Brazilian Computer Society, 2001

This paper proposes an architecture for distributed management of high-layer protocols and network services. Based on the IETF Script MIB, the Trace architecture provides mechanisms for the delegation of management tasks to mid-level managers (MLMs), which interact with monitoring and action agents to have them executed. The paper introduces PTSL (Protocol Trace Specification Language), a graphical/textual language created to allow network managers to specify protocol traces. The specifications are used by mid-level managers to program the monitoring agents. Once programmed, these agents start to monitor the occurrence of the traces. The information obtained is analyzed by the mid-level managers, which may ask action agents for the execution of procedures (e.g. Perl scripts), making the automation of several management tasks possible.

Lecture Notes in Computer Science, 2008

Networks are becoming service-aware implying that all relevant business goals pertaining to a service are fulfilled, and also the network resources are used optimally. Future Internet Networks (FIN) have time varying topology (e.g. such networks are envisaged in Autonomic Internet [1], FIND program [2], GENI program [3], FIRE program [4], Ambient Networks [5], Adhoc networks [6]

The increasing complexity of the network management problem is due to both technological and human factors. Heterogeneity and globalization of network resources on the one hand, increasing user expectations for flexible and easy-to-use environments on the other hand, force to consider new approaches to the management issue. In the paper, we describe MESIS (Management Environment for Secure and Interoperable Services) for the integrated management of both network resources and services. MESIS permits the design and the implementation of new services for open, global, and untrusted scenarios, such as the Internet. The principal part of MESIS is a distributed processing environment composed of a wide set of facilities that can be exploited for building services at the application layer. The paper focuses on two crucial facilities: interoperability and security. The Interoperability Facility allows MESIS services to interwork with all resources and services, even provided by legacy systems. In particular, it considers interoperability with other management systems via compliance with CORBA. The Security Facility provides to MESIS the safe and protected operations necessary in global untrusted environments. MESIS permits authentication of principals, communication integrity and privacy, accountability, and secure interoperation with CORBA-compliant systems. MESIS is implemented in terms of mobile agents, to facilitate delegation and management automation, and to achieve efficiency and scalability through local access to managed entities.

Proceedings POLICY 2003. IEEE 4th International Workshop on Policies for Distributed Systems and Networks, 2003

The Internet is growing to the point of needing more serious, scalable management infrastructure. Telecommunications companies and Internet Service Providers alike face the pressures of upgrading and provisioning their networks while constraining their infrastructure costs to maintain profitability and to stay competitive in an industry that is financially stressed with tight profit margins. In order to be financially successful in this environment, service providers will have to support a variety of services and applications on a combined packet infrastructure, carrying increased varieties of traffic with different performance characteristics and predictable levels of managed Quality of Service (QoS). Multi-Protocol Label Switching (MPLS) traffic engineering enables service providers to engineer their networks to provide such QoS; however, this task brings along with it a plethora of management challenges. This paper discusses these management challenges and our experience with the design and implementation of a policy-based management system, PECAN, for managing MPLS networks. PECAN provides the ability for a network operator to define high-level policies that control the operation of the management system. These high-level policies control admission of traffic into the network based on the QoS guarantees required; placement of traffic flows on MPLS traffic engineered paths; and the feedback loop between network fault/performance monitoring and reconfiguration of the network to alleviate the effects of any observed problems.

Bell Labs Technical Journal, 2004

2008

As telecommunication networks have converged towards Internet technologies, the complexity of managing the telecommunications network equipment has rapidly increased. It is now difficult for telecommunications Service Providers to offer easy-to-manage and feature-enhanced services on a large scale. Many Service Providers are undergoing whole-of-business transformations to replace legacy management systems with next-generation technologies, to remain competitive and to improve cost efficiencies in their management processes. However, without adequate management systems, Service Providers will struggle to deliver complex services to consumers at a low cost and compete with any market entrants that emerge. Policy-based Management (PBM) systems offer an approach that help to simplify complexity, manage large-scale networks (with reduced effort), and provide personalisation through dynamic and adaptive service delivery. This occurs by describing the behaviour of the network and the management systems in response to environmental conditions or events. These policy descriptions allow changes to the network to be deployed at a large scale with near real-time performance. Thus, the behaviour of the services or management systems can be changed to deliver new services and features rapidly, and to support the personalisation of services for individual customers. This dissertation presents the PRONTO management system, which uses a policy-based service definition language. A service definition language describes the configuration and operation of a service that extends across the network. Complex network services can be described so that each service can be provisioned, monitored and managed in real-time without human assistance. By being policy-based, the language describes the behaviour of the network services in response to dynamic conditions and events that occur. Consequently, the behaviour of the service can change over time, to enhance the experience of the users in different environments and to suit the unique requirements of the users. It is necessary to overcome some of the challenges present with existing policy-based management systems. All PBM systems require improvements to the usability of the policy languages, so that the management of complex networks and services is eased and the most flexibility is provided to Service Providers. PBM systems also need to address scalability issues. Policies need to be distributed widely, and coordination is required to ensure that policies are enforced consistently across the network. This dissertation addresses, and provides a solution for, usability and scalability. An approach to distribution and coordination of policies across multiple management systems is proposed. Mean Value Analysis (MVA) queuing models and Discrete Event Simulations (DEVS) are used to predict the management system behaviour in terms of performance, scalability and its response to dynamic policy changes.

2000

IEEE Network, 2003

Next-eneration networks must be capable of supporting a multitude of service and quickly adapted over a common heterogeneous physical infrastructure, according to varying and sometimes conflictin customer requirements. In this context, network management must become more fyexible in order to cope with these emerging conditions. More specifically, new management architectures must offer service providers

Lecture Notes in Computer Science, 2000

Policy-based networks can be customized by users by injecting programs called policies into the network nodes. So if general-purpose functions can be specified in a policy-based network, the network can be regarded as an active network in the wider sense. In a policy-based network, two or more policies must often cooperate to provide a high-level function or policy. To support such building-block policies, two architectures for modeling a set of policies have been developed: pipeconnection architecture and label-connection architecture. It is shown that rule-based building blocks are better for policy-based network control and that the label-connection architecture is currently better. However, the pipe-connection architecture is better in regards to parallelism, which is very important in network environments.