Evaluation of open source software for mobile ad hoc routing in military tactical networks Master’s thesis within Computer Systems and Networks Oscar Holmberg Department of Computer Science and Engineering Division Networks and Systems CHALMERS UNIVERSITY OF TECHNOLOGY Gothenburg, Sweden 2013 The Author grants to Chalmers University of Technology and University of Gothenburg the non-exclusive right to publish the Work electronically and in a non-commercial purpose make it accessible on the Internet. The Author warrants that he is the author to the Work, and warrants that the Work does not contain text, pictures or other material that violates copyright law. The Author shall, when transferring the rights of the Work to a third party (for example a publisher or a company), acknowledge the third party about this agreement. If the Author has signed a copyright agreement with a third party regarding the Work, the Author warrants hereby that he has obtained any necessary permission from this third party to let Chalmers University of Technology and University of Gothenburg store the Work electronically and make it accessible on the Internet. Evaluation of open source software for mobile ad hoc routing in military tactical networks Master’s thesis within Computer Systems and Networks OSCAR HOLMBERG c OSCAR HOLMBERG, October 2013 Examiner: PETER LUNDIN Chalmers University of Technology Department of Computer Science and Engineering Division Networks and Systems SE-412 96 Göteborg Sweden Telephone: +46 (0)31-772 1000 Department of Computer Science and Engineering Göteborg, Sweden October 2013 Abstract Evaluations of ad hoc routing protocols have been performed in sev- eral studies, both with hardware and in software simulators. How- ever, the network topology plays an important role for the protocol performance. This master thesis evaluates OSPF, OLSR, Babel and Batman-adv in a tactical network set up. A tactical network is not a completely ad hoc scenario which makes it different from other studies. Furthermore, the protocols’ settings are varied and they are mainly tested against convergence time and generated overhead traffic. The aim is to find which routing protocol is most suitable in the tactical network scenario that is tested. The tactical network set up, consisting of 36 nodes, is created with virtual machines and software that emulates realistically radio links. The results show that OSPF can be configured to adapt to a tactical network. OSPF and Babel shows the best performance with respect to convergence time and generated overhead traffic. However, OLSR offers a lot of configuration possibilities which pro- vides the potential to find a more suitable configuration than those that are used in the tests. Both Batman-adv and OLSR generates more overhead traffic than Babel and OSPF for the corresponding convergence time. Keywords: Ad hoc routing, tactical network, OSPF, Babel, Batman- adv, OLSR. Acknowledgements I would like to thank my supervisor at Saab, Anders Gunnar, for the opportu- nity to perform this master thesis and for his valuable input during the project. I would also like to thank my supervisor at Chalmers, Daniel Cederman, for his proofreading and comments on the report. i Contents Acknowledgements . .i List of figures . iv List of tables . .v Abbreviations . vii 1 Introduction 1 1.1 Background . .1 1.2 Objective . .2 1.3 Delimitations . .2 1.4 Methodology . .3 1.5 Related work . .3 1.6 Report structure . .4 2 Literature review 5 2.1 Routing in MANET:s . .5 2.2 Proactive or table-driven routing protocols . .5 2.2.1 OSPF . .6 2.2.2 DSDV . .8 2.2.3 OLSR . 10 2.2.4 Batman . 12 2.2.5 Babel . 13 2.3 Reactive or on-demand-driven routing protocols . 13 2.3.1 AODV . 14 2.3.2 DSR . 14 3 Test environment 17 3.1 Software . 18 3.1.1 VirtualBox . 18 3.1.2 Radio Link Emulator (RLE) . 19 3.1.3 Quagga . 20 3.1.4 Babeld . 20 3.1.5 OLSRd . 20 3.1.6 Batman-adv . 21 3.2 Network set up . 21 3.3 Test definition . 27 4 Results and Discussion 30 4.1 Results with default configurations . 30 4.2 OSPF . 34 ii 4.3 OLSR . 35 4.4 Babel . 37 4.5 Batman-adv . 39 5 Conclusion 42 Appendix A - Network topology of test environment 46 Appendix B - Overhead traffic for different configurations, relation between bytes and number of packets 47 iii List of Figures 2.1 A mobile ad hoc network where node D moves to a new location. Changes in the routing tables must always be updated. .9 2.2 The messages sent during a neighbour discovery procedure in OLSR. 10 2.3 The difference of using MPR nodes in the flooding procedure. 11 2.4 Example how an error message is created in order to inform the source node that the destination can’t be reached . 15 3.1 Describes how the communication is allowed in a battalion. The communication is hierarchical and must follow chain of command. 17 3.2 Two virtual machines running on a host computer. The virtual ma- chines are connected with an internal network inside VirtualBox. 19 3.3 Overview of the routing nodes and how they are connected in a tac- tical network set up. 21 3.4 Overview of the routing nodes distributed among different radio net- works. 22 3.5 Example of how the RLE and routing nodes are connected together. 23 3.6 A layered view that shows a small part of the network and how it is running inside VirtualBox. 24 3.7 The three physical computers and how they are connected with the use of VLAN. 25 3.8 A logical view of how the traffic is separated in the three lab computers. 26 3.9 A logical view over the routing nodes where the edges are controlled by the RLEs. The dotted edges have a packet loss of 100 percent at some times during the test. Network traffic is logged at all gray nodes. 27 4.1 The ping ratio for all protocols with their default configurations. 30 4.2 Overhead traffic on intermediate nodes when all links have packet loss from 0 to 25 percent. All protocols have their default configuration. 32 4.3 Overhead traffic on end nodes when all links have packet loss from 0 to 25 percent. All protocols have their default configuration. 33 4.4 The relation between overhead traffic and convergence time for dif- ferent settings for OSPF. 34 4.5 The ping ratio for different settings for OSPF. 35 4.6 The relation between overhead traffic and convergence time for dif- ferent settings for OLSR. 36 4.7 The ping ratio for different settings for OLSR. 37 4.8 The relation between overhead traffic and convergence time for dif- ferent settings for Babel. 38 4.9 The ping ratio for different settings for Babel. 39 iv 4.10 The relation between overhead traffic and convergence time for dif- ferent settings for Batman-adv . 40 4.11 The ping ratio for different settings for Batman-adv. 41 v List of Tables 2.1 Message types of the OSPF protocol . .7 2.2 The routing table for node D before moving to the new location. .9 2.3 The routing table for node D after moving to the new location. .9 3.1 How the end nodes send ICMP messages to each other and possible shortest path between the nodes. 28 3.2 Describes the packet loss on the edges for the five test cases. 28 4.1 Convergence times measured at the node N2RLE3 with default con- figurations for all protocols. 31 vi Abbreviations ANSN Advertised neighbour sequence number AODV Ad hoc on-demand distance vector Batman Better approach to mobile ad hoc networking DSDV Destination-sequence distance vector DSR Dynamic source routing EIGRP Enhanced interior gateway routing protocol ICMP Internet control message protocol IHU I hear you IP Internet protocol MANET Mobile ad hoc network MDR Manet designated router MPR Multi-point relay NIC Network interface card OLSR Optimized link state routing OLSRd Optimized link state routing daemon OGM Originator message OR Overlapping relay OS Operating system OSPF Open shortest path first OSPFd Open shortest path first daemon RAM Random access memory RFC Request for comments RIP Routing information protocol SSH Secure shell TC Topology control TTL Time to live VLAN Virtual local area network vii Chapter 1 Introduction 1.1 Background There are many different routing protocols that are designed for different intents and purposes. Some protocols are designed to cope well with changes while some are optimized for other objectives. In order to choose the most suitable protocol for your intents, one has to be aware of the network topology properties. For example, today’s Internet is a static network with respect to the routers and network infrastructure that create the backbone of the Internet. Even though new nodes are connected and others are removed, the backbone network topology is essentially stable. In contrast, an ad hoc network does not have a fixed network topology and it does not rely on any fixed infrastructure [1]. This poses new requirements on the routing protocols that are being used. For the Internet, all routers in the network contain an up to date routing table in order to select the best route for the traffic. The routing protocol that ensures that this table is up to date is not designed to cope with rapid changes in the network topology. The routing protocols for the Internet are designed for stable and stationary networks. When the routers have found each other once, there is not much change in the routing table which imposes low demands on the update frequency for the routing protocol in use.