Policy-constrained bio-inspired processes for autonomic route management

Abstract

Autonomic networking systems must be designed to achieve an appropriate balance between the operation of decentralized algorithms and processes that seek to maintain optimal or near-optimal behavior in terms of global stability, improved performance and adaptability, robustness and security, with the requirement for top-down control of the system by humans to ensure business goals are met. Taking a communications networking survivability case study, we show how the operation of decentralized algorithms, inspired by the operation of biological systems, can be controlled and constrained through the deployment of management policies authored by network administrators. We present survivability-related routing algorithms (inspired by chemotaxis, reaction-diffusion and quorum sensing biological processes) which work together to effectively reconfigure network resources when transient link failures occur and demonstrate how these algorithms can be re-parameterized via policies to improve performance given prevailing network conditions. Simulation results show how the combined operation of these algorithms, as controlled by policies, allows the network to react well to survive link failure events.

Introduction

As communications networks have become more dynamic, more heterogeneous, less reliable and larger in scale, researchers and industry practitioners have started to actively investigate how best to engineer systems embodying autonomic principles. The aim is to simplify network management processes by automating and distributing the decision making processes involved in optimizing network operation. This allows expensive human attention to focus more on business logic and less on low-level device configuration processes, with consequent revenue savings and increased efficiency.

Much of the work to date on autonomic network management has focused on the deployment of decentralized processes and algorithms into the network infrastructure that seek to maintain optimal or near-optimal behavior in terms of global stability, improved performance and adaptability, robustness and security. As described by Babaoglu et al. [1], many of these processes and algorithms can be profitably modeled on various biological processes found in the natural world. However, to ensure that they act in accordance with business goals, we argue that such processes and algorithms should themselves be modeled so that their operation can be automatically configured via appropriate management policies. This allows the human administrators of the network to explicitly constrain the behavior of the processes and algorithms via appropriate parameterization, so that the system operates in accordance with (changing) business goals and/or changing operational context. Integrating models of the algorithms and processes in system information models allows specification of the impact their parameterization has on services, products, customers, and ultimately on the revenue generated by the system.

In this paper we discuss how policy-constrained bio-inspired algorithms and processes for routing can allow a network to gracefully react to transient link failures, which are described by Iannaccone et al. [2] as the primary cause of service degradation in current networks. More specifically, we specify routing algorithms and processes based on biological principles and describe how they can be integrated within a policy-based network management framework for autonomic networking. We show how the operation of the algorithms under the control of management policies allows network elements to react to link failures by cooperatively re-routing traffic in a manner that ensures prioritized services are impacted minimally.

The paper is structured as follows. In Section 2 we briefly summarize related work in the areas of autonomic computing and communications, bio-inspired algorithms, policy-based network management and network survivability. Section 3 introduces our policy-based network management architecture for control of bio-inspired routing processes, whilst Section 4 specifies the bio-inspired routing algorithms and processes whose operation is controlled and constrained via policies. Section 5 outlines how the operation of these algorithms is modeled in the context of a wider information model, which is then used to (semi-)automatically generate appropriate policy authoring and analysis tools. Section 6 describes how our prototype implementation was used to explore a case study of traffic re-routing following a link failure event. Finally, Section 7 summarizes the work and outlines areas for further exploration.