OMNeT++ Community Summit 2020

Successor of the International Workshop on OMNeT++

Virtual Summit – October 05-06

INTRODUCTION

Because of the changing world, the seventh OMNeT++ Community Summit will be held virtually from Monday to Tuesday, October 05-06, 2020.

OMNeT++ is a public-source, component-based, modular and open-architecture simulation environment with strong GUI support and an embeddable simulation kernel. It is designed to simulate discrete event systems, with the simulation of communication networks as one of its primary application areas.

In continuation of the history of the International Workshop on OMNeT++ and four very successful editions of its reincarnation, the OMNeT++ Community Summit, we continue to hold a yearly open meeting for all OMNeT++ community members with less costs and more interaction possibilities. The Community Summit provides a forum for tutorials, discussion sessions and presentations on recent developments and novel ideas in the area of simulation and modeling, with a focus on the OMNeT++ simulation environment.

The first six Summits in Hamburg, Zurich, Brno, Bremen, Pisa and Hamburg were huge successes with lively discussions, keynotes, tutorials, panels, presentations, and demonstrations of ongoing and finalized research covering many aspects of OMNeT++-related research.

OMNeT Community Summit 2020

HOW TO JOIN

The summit will take place on Zoom (video conferencing software). Registration deadline is Friday, October 2, midnight.

SUMMIT PROGRAM

The OMNeT++ Summit intends to be a lively event with a lot of discussions. A wide range of presentations and interactive sessions regarding specific topics will be held. There will be enough time to discuss about the different parts of OMNeT++ and related simulation and modeling questions. The event will be fully virtual. All times are Central European Summer Time (GMT+2).

Accepted Presentations

05. October 2020

All sessions will take place in virtual meetings. Access code will be sent to registered participants.
11:00 - 11:10 Welcome and organisation
11:10 - 11:45 Levente Mészáros (slides)
INET infrastructure for Time Sensitive Networking simulation With the advancements of in-vehicle communication systems and the ongoing automation of traditional manufacturing and industrial operations, Time-Sensitive Networking (TSN) became an essential target for network simulation. TSN is a set of standards, developed by the IEEE 802.1 working group, that define mechanisms for the time-sensitive transmission of data over deterministic Ethernet networks.

The current version of INET lacks important infrastructure components for TSN simulations: clock models, time synchronization protocols, a versatile queueing model for transmission selection, the general support for aborting transmissions, and an API for intra-node packet streaming. The presentation covers the latest developments regarding these topics.
11:45 - 12:00 Online coffee break and open discussion
12:00 - 12:20 Presentation 1: Mehmet Cakir (slides)
Simulation-Based Evaluation of a Delay-Based Forwarding Concept Clemm and Eckert propose Latency-Based Forwarding (LBF) as a novel approach to provide support for high-precision latency objectives. It prioritizes traffic with introducing packet metadata which carries latency objectives, and the metadata allows different actions to be taken at network nodes. A proof of concept simulation model has been developed using the Big Packet Protocol (BPP). In contrast to IntServ, LBF supports prioritizing specific packets, with the purpose of providing fairness among different applications. We have run OMNeT++ simulations and compared our simulation results to the emulation results, with the goal of validating the results of Clemm and Eckert.
12:20 - 12:40 Presentation 2: Attila Török (slides)
Proposed research topic: Using neural networks and machine learning
for deriving error models for WiFi and other wireless networks In wireless network simulations, an error model describes how the signal to noise ratio (SNR) affects the amount of errors at the receiver. Given that SNR is a multi-dimensional function over time and frequency, and given the diversity of coding and modulation schemes used in wireless networks, writing a good and efficient error model is a very difficult problem. Existing error models are often closed formulas derived from empirical observations, and in addition to being very limited in scope, they sometimes fall down even within their supposed limitations. In contrast, we propose using neural networks and deep learning techniques to produce an error model that, in addition to being universally applicable, can produce reasonable answers even for cases that it was not explicitly trained for.

We propose this as a research topic for those potentially interested in pursuing it. It is very promising as a research topic, practically feasible, and very useful at the same time. We have already spent some time trying out the idea and proven (at least to ourselves) that it is feasible and the approach outlined here can be made to work, but we don't have the resources (mostly, time) to elaborate it in-house. If you are a researcher (e.g. PhD student) looking for an exciting and rewarding topic to work on, we'd love to hear from you. We recommend this research topic primarily to those who are already knowledgeable in the field of neural networks and machine learning, and are not afraid of applying this knowledge to something new.
12:40 - 13:40 Lunch break
13:40 - 14:00 Presentation 3: Pierre Larrenie, Iyad Lahsen-Cherif, Olivier Venard, Jean-François Bercher
and Catherine Lamy-Bergot
Implementation of MANET routing protocols with OMNeT++ : A DSR-Flow showcase In unstructured and mobile networks such as MANETs, routing may become a real challenge due to frequent topology changes. Routing protocols are classified in 3 categories: proactive, reactive and hybrid. Proactive protocols send periodic messages to discover its neighborhood and construct the topology, whereas reactive ones ask for new routes only on demand. Finally, hybrid protocols combine both proactive and reactive features. Our aim is to conduct a comparison study of various routing protocols: AODV, OLSR, DSR under UAV-like mobility model regarding their performances mainly in terms of end-to- end delay, network load, and Packet Delivery Ratio (PDR). This requires the implementation of the DSR-flow extension in OMNeT++.

In this work, we provide guidelines to OMNeT++ users to implement a routing protocol with INET/INETMANET projects. We describe how we managed to implement a simple version of the well-known routing protocol DSR and its flow extension. Our contribution is threefold. First, we defined and extended data-structures and classes of INET/INETMANET to get a classic version of the DSR protocol. Then, we implemented the flow extension by using the classic DSR implementation and finally, we used specific UDP applications to provide stream traffic generation in order to compare features and performance of both protocols. We are particularly interested by the end-to-end management feature that could guarantee a certain QoS level for a dedicated service. Hence, we focused on core concepts of the DSR- Flow extension (i.e flow labelling) and did not implemented yet rare options such as "default flow" feature and "Destination and Flow ID" option which do not enter our domain of interest. We plan to open-source our implementation in the future.
14:00 - 14:20 Joint discussion - generic topics
14:20 - 15:00 Deep discussion of presentations (open-ended)
Presentation 1 in room A, Presentation 2 in room B, Presentation 3 in room C

06. October 2020

All sessions will take place in virtual meetings. Access code will be sent to registered participants.
11:00 - 11:20 Presentation 4: András Varga (slides)
Proposed research topic: Zero configuration automatic parallel simulation OMNeT++ already supports parallel simulation. However, the network must be partitioned manually, and the partitions run concurrently in separate processes communicating over MPI. Here we propose a parallel simulation approach relies on multi-threaded execution instead of communicating processes, and therefore, by eliminating MPI overhead, it can make much better use of shared-memory multiprocessors which are e.g. commonly available on everyone's desk as multi-core CPU laptop/desktop computers. Moreover, it requires no prior knowledge or manual configuration about the simulation model. This makes it potentially both significantly more efficient (higher speedup), and also more accessible and convenient to use than the existing solution.

We propose this as a research topic for those potentially interested in pursuing it. It is very promising as a research topic, practically feasible, and very useful at the same time. We have already spent some time trying out the idea and proven (at least to ourselves) that it is feasible and the approach outlined here can be made to work, but we don't have the resources (mostly, time) to elaborate it in-house. If you are a researcher (e.g. PhD student) looking for an exciting and rewarding topic to work on, we'd love to hear from you.
11:20 - 11:40 Presentation 5: Ujjval Rathod (slides)
Timing Analysis of SpaceWire using OMNeT++ based Simulator The SpaceWire is a standard for communication for on-board spacecraft. SpaceWire is nominated by ESA and planned to be the only communication standard for spacecrafts in the future. It is already in use by several missions which are currently in operation. The SpaceWire protocol is not a deterministic protocol. It supports arbitrarily large sizes of messages with possibility of different datarates. It has wormhole routing in routers which may block critical messages for large amounts of time. It also has very low buffer memory and no virtual channel support. These conditions makes SpaceWire’s timings very uncertain. Embedded system developers need a realistic bound on the end-to-end delay for messages sent over the SpaceWire to design software with more predictable timing properties. Worst-case analysis of the end-to-end delay reports very pessimistic bounds, which are not applicable for many applications. A good simulator can provide less pessimistic bounds. It can provide accurate modeling of the network and its timing behavior.

First, we will show the implementation of SpaceWire’s routing strategies in OMNeT++. Then an implementation of wormhole routing will be presented. Based on routing strategies, end-to-end delay for messages considering different network scenarios will be analyzed. The blocking of packets in the router will also be demonstrated. Future work may focus on network emulation and test and validation process.
11:40 - 12:00 Presentation 6: Nicholas J. Omumbo (slides)
Computational Evaluation of the Effects of Openflow and Ospf Routing
on Quality of Service Metrics in a Large Simulated Mobile Core Internet Protocol Network
Using OMNeT++ Mobile Internet Protocol (IP) phone networks are arguably one of the most important innovations of modern times. More access systems are increasingly being defined using IP and as a result one of the challenging tasks that mobile IP phone service companies have to deal with is an upsurge of user equipment (UE) which in turn places excessive demand on the routing infrastructure and degrades quality of service (QoS) values at the mobile core network. Studies show that majority of mobile IP phone subscriber companies struggle to meet acceptable QoS values due to their insistence on using classical routing methods. Classical routing is not easily scalable and often lacks central source of information. There is an attempt to explore emerging Software Defined Network (SDN) technologies to enhance scalability and promote centralized controller approach. OpenFlow routing is an instance of SDN that has been found to improve routing in wired and small wireless networks.

The aim of this study was to evaluate the computational effects that OpenFlow and OSPF would have on routing in a large mobile core IP network. Specifically, the study evaluated the effects of the use of OpenFlow and OSPF routing on jitter, packet delivery ratio (PDR), throughput and end to end delay. This research adopted exploratory research design where OMNeT++ network emulator was used in addition to SimuLTE add on to model two mobile IP networks; OSPF routed and OpenFlow. Applications that present high demand for processing and routing used; interactive gaming, VoIP, audio streaming and Internet Protocol Television (IPTV) were used for the purposes of testing routing efficiency. Scenario manager was used to simulate network failure and renewed convergence was observed. The setup test environment consisted of 1000 UEs, 80 OSPF routers, 80 OpenFlow switches. The study used already existing tools in OMNeT++ namely scenario manager, UEs, OSPF and OpenFlow routers to build test environment.

OpenFlow was observed to manage better the QoS metrics in large mobile core network compared to OSPF routing. Notably jitter was improved by 10 milliseconds when OpenFlow routing was used in comparison to OSPF. This improvement was consistent across the network with addition of more end user equipment (UE). In instances where OSPF improved routing speeds the value is less significant with standard deviation of 24 mbs. This study provides insight on how OpenFlow can improve routing efficiency in large mobile core IP network compared to classical routing approach. This study further recommends testing of OpenFlow in production networks.
12:00 - 12:20 Joint discussion - generic topics
12:20 - 13:00 Deep discussion of presentations (open-ended)
Presentation 4 in room A, Presentation 5 in room B, Presentation 6 in room C
13:00 - 14:00 Lunch break
14:00 - 14:20 Presentation 7: Jan Zavřel, Vladimír Veselý (slides)
Enhanced Interior Gateway Routing Protocol Enhanced Interior Gateway Routing Protocol was published as RFC with the blessing of Cisco a few years ago. EIGRP is just like Babel hybrid routing protocol that leverages the best features of both distance-vector and link-state routing protocols. Properties like adjacency detection, multi-address family support, and the famous DUAL algorithm make EIGRP enterprise solution (that offers “loop-free routing” even during network reconvergence). We have successfully implemented EIGRP supporting IPv4 and IPv6 for ANSA; and now is the excellent time to incorporate EIGRP directly into INET.
14:20 - 14:40 Presentation 8: Konrad Fuger (slides)
Evaluation of Avionic Routing Protocols using a Multiscale Simulation in OMNeT++ One focus at the Institute of Communication Networks at the Hamburg University of Technology is routing for avionic applications. Here, especially the case of aircraft crossing the North Atlantic Corridor (NAC) between Europe and North America is an interesting scenario as it is the busiest airspace of the world without any ground infrastructure. Obtaining performance indicators for routing protocols on this scenario from simulations creates a challenging task: On the one hand, mobility has a strong effect on the routing performance as it defines the topology of the network. On the other hand, realistic link and PHY layer implementations are necessary to obtain meaningful end-2-end delays. While mobility over the NAC is a slow moving process, requiring the simulation to cover five to six hours of mobility with topology changes on the scale of minutes, a realistic lower layer implementation will create events every few milliseconds.

A simulation covering both the mobility as well as the lower layer communication protocols at the same time is computationally infeasible. Therefore the concept of Multiscale Simulations emerge: The core idea is to split the simulation effort into two, where a fast running "Macro Scale" simulation evaluates the mobility with simplified lower layers and creates snapshots of the simulation to be investigated in more detail. Those snapshots are then picked up by a "Micro Scale" simulation which starts a simulation of the created snapshot with a full lower layer implementation.
14:40 - 15:00 Presentation 9: Yevhenii Shudrenko(slides)
Routing Protocol for Low-Power and Lossy Networks (RPL) in INET Framework We will shortly introduce the RPL protocol and the RFC 6550 features implemented in our simulation model. We will provide and overview of our implementation structure and present example scenarios and showcases. In the future, we intend to integrate our work with the 6TiSCH stack.
15:00 - 15:40 Deep discussion of presentations
Presentation 7 in room A, Presentation 8 in room B, Presentation 9 in room C
15:40 - 16:00 Wrap up and farewell

ORGANIZERS

Founding Chair
  • Andras Varga (OpenSim Kft.)

Virtual Summit Organizers
Technical Program Organizers


In case you want to contact the organizers, feel free to write an e-mail to: info [at] omnetpp [dot] org.