RealNeS (Realtime Network Simulator) is the tool of choice to simulate and understand interaction of various protocol layers to derive meaningful conclusions on system-level performance and quality of experience. It is actually not a single simulator for a particular radio access technology, but it covers various technologies, namely LTE(-A), 802.11 and 5G NR. It also features a multi-RAN framework that facilitates the simultaneous simulation of several networks of various RATs each operating in a unique and mutually exclusive frequency band. This facilitates various studies on e.g. traffic steering across networks, mobility and network convergence. The channel model of RealNeS is implemented in CUDA to run very efficiently and in a highly parallelized manner on nVidia graphics cards.

The main features and capabilities of RealNeS are as follows:

Simulation Granularity
  • In terms of communication protocols the various layers of the user plane protocol stacks are implemented in great details handling the transmission of each individual packet in detail.
  • Error modeling is performed based on a computation of effective SINRs in the equivalent complex baseband (ECB) taking into account (codebook and non-codebook based) precoding and receive filtering techniques (MRC, LMMSE, LMMSE-IRC). The SINR computation is done at sufficient—possibly even sub-TTI—time and frequency granularity to track fast fading effects.
  • Simulator speed varies with the size of the scenario ranging from above realtime for small scenarios and significantly below realtime for large scenario with thousands of UEs in the simulation.
Protocol Stack

RealNeS is focused on the user plane. Here, all relevant layers of the protocol stack are implemented from a large set of traffic types configurable per UE and flow over UDP/TCP et al. to the RAT specific layers of PDCP/LLC, RLC, MAC and PHY layer.

Radio Resource Management (RRM)

RealNeS supports various complex scheduling techniques (single- and multi-user MIMO, proportional fair with QoS support, with explicit handling of retransmissions, link-adaptation techniques, etc.)

Radio Access Technologies (RATs)

RealNeS can simulate LTE (incl. eMBMS and various modes of LTE-V2X), 802.11, and 5G networks in frequency bands ranging from 700MHz up to 100GHz.

  • LTE: While RealNeS cannot be claimed to be fully compliant up to a certain release, it does support the basic and most of the relevant advanced features, such as MIMO, HARQ, HetNets, ICIC, eICIC, FeICIC, FDD and TDD, LTE-V2X also via PC5, eMBMS (MBSFN and SC-PTM) and many more. Calibration against simulation results from within 3GPP standardization has been performed.
  • 802.11: RealNeS features a detailed implementation of basic PHY and MAC layer functionality for 802.11a/g/n/p, including CSMA-CA (w/, wo/ RTS-CTS), ARQ, EDCA (802.11e) for QoS, beacon-based mobility, D2D e.g. for vehicular communication and others.
  • 5G: While 5G standardization in 3GPP is still in progress RealNeS implements some early ideas of 5G and is being aligned with 3GPP’s evolving idea of 5G. This includes new numerology and frame structure, mm-wave communication with inband relaying, massive MIMO schemes. In the context of the EU-funded project ‘5G-XCast’ multicast / broadcast functionality is being added, as well.

RealNeS has a mobility module that supports various UE mobility models. These models are:

  • Street Graph Model: Based on an interface to the SUMO™ simulator it is possible to simulate realistic mobility in RealNeS. Using the rich feature set of SUMO™ it is possible to either define own geometries of mobility paths or to import realworld data e.g. from OpenStreetMap.
  • Birth-Death Drop Model: For the purpose of statistical simulations e.g. for capacity analysis UEs can be dropped randomly within the simulated area. In case of session-based traffic models such as FTP traffic, each UE is dropped to a new position after the completion of each session.
Radio Propagation Models

RealNeS supports various propagation models ranging from UMTS 30.03 and manually configured pathloss models, over Winner2/+ to the most recent channel model of 3GPP defined in TR 38.901. Both the classical, purely statistical models are supported as well as partially geometry-based models where channel conditions are explicitly impacted by specific buildings defined in the scenario.


RealNeS provides various KPIs at various protocol layers from throughputs at application, over PDCP packet delays and radio resource allocation at MAC layer down to BLER before and after HARQ and SINRs, e.g. at LTE-PRB resolution.

Network Emulation

If run in realtime live data can be routed through RealNeS to analyse the behavior of a certain application (e.g. video streaming, gaming, augmented / virtual reality) when operated over a particular communication network.