Imunes

IMUNES Based Distributed Network Emulator
Z. Puljiz and M. Mikuc
Faculty of Electrical Engineering and Computing/Department of Telecommunications, Zagreb, Croatia zrinka.puljiz@fer.hr miljenko.mikuc@fer.hr
Abstract— In this paper we describe a new version of our distributed network emulator that extends an existing kernel level emulator called IMUNES. IMUNES is based on a lightweight virtualmachine concept and performs zero copying when packets traverse through the emulated topology. It works on a modified FreeBSD kernel and enables emulated nodes to use the standard UNIX applications. The main strengths of this tool are scalability, performance and high fidelity. We are developing a distributed network simulation to further increase the scalability by allowing parts of emulation tobe deployed across a peer-to-peer emulator cluster. The decentralized management of the emulator cluster improves availability and robustness of the system. We provide support for a multi-user and multi-experiment environment to maximize the benefit from newly increased resources. Distributed emulation, scalability, clustering

1. INTRODUCTION In network emulations the demands for resources arerising with the complexity of the emulated network. The resources available on only one machine are imposing the limit on scalability. As a result of dealing with this problem many powerful network emulators are designed as distributed systems [1], [9] and [11]. IMUNES is a network emulator that offers high scalability, performance and fidelity. Designing a distributed emulator based on IMUNESpushes the limits of scalability even further. IMUNES is a topology specification, management and GUI application. Core of IMUNES network emulation facility are kernel level lightweight virtual machines available in modified FreeBSD kernel. Distributing the simulation based on IMUNES implies the separation of topology specification and GUI application from the management utility, and extendingmanagement utility to provide the support for distributed simulations. Our distributed simulator works on a cluster and fulfills the following: • Improved scalability – this is the most important demand. We expect the scalability to improve proportionally to the number of distributed hosts. • Improved robustness – the state of the cluster is kept up-to-date on each host in the cluster. Even in the caseof a failure of any host in the cluster the state of the cluster remains available and accurate. • Improved utilization – all the available resources of the cluster should be used in an efficient manner. This means allowing more than one person to use the

cluster, and to emulate more than one topology at a time. • Improved portability – allowing user operating on some other OS, to use the IMUNESemulation capabilities. • Keeping the performance – we are preserving the simulation properties even in the case when the simulation is distributed among the hosts in the cluster. • Improved availability – we are providing remote access to the cluster. We have studied the centralized architecture that is used in some other distributed emulators [1] and [11], but in the end we decided to go for apeer-to-peer cluster. Our goal is to show that the proposed distributed network emulator is a powerful tool for simulation of large networks. For distributing the topology we used standard graph partitioning methods also used graph partitioning methods like in [13]. The rest of this article is organized as follows: in the second section we give an introduction to IMUNES, its building blocks,kernel structures and the properties. The distributed architecture in presented in the third section. In the fourth section are the results and the conclusion, whereas the future work is described in fifth section. Related work is presented in the last, sixth section. 2. RELATED WORK Many existing network emulators are designed as distributed systems. This is partially because the network emulation...