Networking technologies make up the oft invisible rump of the Internet. This covers essentially everything below layer 7 (the application layer) — from transport to the actual PHY interface. Recent key developments range from changes to the control plane paradigms (SDN, NFV) to innovations in the layer 1 and 2 networks, including wireless networks (5G and LPWAN) and Industrial networks (TSN), as well as the manifestation of these technologies in new classes of devices (IoT).
Our current research focuses on the following topics.
The ever-increasing traffic demands and application variety in the Internet complicates network management tasks. The over-reliance on middleboxes --- think, e.g., application-layer firewalls or load balancers -- contributes to the lack of flexibility in typical networks. These ---oft proprietary --- intermediary nodes accomplish a single, fixed task and cannot be adopted to suit specific needs.
In order to remedy such circumstances and offer tailor-made networking solutions, Software Defined Networking (SDN) and Network Function Virtualization (NFV) is nowadays being employed in Data Centers and Cloud Computing. Where in the past Control and Data Plane have been somewhat coupled and control was distributed, SDN provides a distinct separation of these. This allows the creation of dumb packet forwarders under the direction of a logically centralized controller, easily implemented and customized in software with well-defined APIs. NFV is intended to replace middleboxes with smaller, portable software functions, thus enabling rapid development and simpler ways to scale out to the current load.
And yet, these technologies also pose new challenges, including performance-related ones. Our research on these topics applies methods of analytical performance evaluation, discrete event simulations and testbeds to examine the solution space. Our tasks encompass exploring solutions for an effective resource and function placement, evaluating distributed approaches to reach consensus in these dynamic systems, and programmable network hardware that implements P4. Finally, we are also looking at security-related network test and validation methods.
In contrast to IETF-specified Internet architectures, mobile network architectures --- as defined by 3GPP --- are deeply vertically integrated, with a strong emphasis on the control plane, state and signaling interactions. Each generation of mobile network exhibits different properties through an updated architecture and signaling. The 5G network architecture of the current 3GPP release 16 even defines an entirely new architecture.
These mechanisms allow the mobile network --- both radio access network (RAN) as well as the core (CN) --- to be more easily managed and optimized for specific tasks and traffic characteristics. But it makes it also less future proof and predictable in its behavior towards other, unknown usage patterns.
Thus, we strife to understand and model signaling behavior and the resulting control plane load. This is crucial in order to scale mobile networks to environments like massive IoT and Smart Cities, which typically have been more suitably deployed as LPWA (Low-Power Wide-Area) networks.
From the Internet and modern software we are used to a certain kind of flexibility. Connect your device to any network, and every App auto-magically simply works. Industry 4.0, Industrial IoT, and Smart Factories aim to bring such plug & play concepts to the factory floor. In a similar notion, vehicles are being equipped with increasing amounts of assistive tools in order to some day accomplish SAE level 5 autonomous driving.
These tools and devices need to be connected to the vehicle's control unit and can produce large amounts of data. Previous approaches, including CAN bus, Profibus, or EtherCAT, either do not reach the bandwidth demands, do not give real-time guarantees, or are not flexible enough.
In order to achieve these goals at scale, standardized Ethernet-based networking solutions are being developed, that still fulfill strict real-time requirements under the term Time Sensitive Networking (TSN). These includes synchronized variants, where a central controller ensures that all connected devices operate in a (time-)synchronous fashion (e.g. IEEE 802.1Qbv), but also asynchronous variants that enforce proper transmissions through their shapers and schedulers (see e.g. 802.1Qcr) with the help of admission control procedures.
Our research revolves around simulative and analytical examinations of such standards. Currently in development are packet-level simulations, mathematical bounds for certain shapers, but we are also evaluating tools and approaches for network planning and dimensioning. Of particular interest --- yet tremendously more challenging to achieve --- are time sensitive mobile and wireless networks, advertised as URLLC (Ultra-Reliable Low Latency Communication) in 5G and RAW (Reliable and Available Wireless) by the IETF.
Modeling of IoT systems, architectures and IoT traffic is one of the key research activities of the Chair. In particular, traffic modeling of the large number of devices, sensors and actuators are in focus. Furthermore, research is being done in the area of simulation, dimensioning and scaling of IoT architectures within several projects.
Traffic Aggregation Models An IoT architecture usually consists of a number of small individual components. These components are grouped according to their spatial or thematic orientation. A group uses a so-called gateway to send data. The aggregated traffic patterns are the subject of research in this area. With the traffic patterns one can estimate emerging traffic, scale and optimize architecture components.
Low Power Wide Area Network (LPWANs) Low Power Wide Area Networks describe a class of network protocols for connecting low power devices such as battery powered sensors to a network server. At the chair, research is being conducted to optimize and enhance such protocols.
Smart City is a collective term for holistic development concepts that aim to make cities more efficient, technologically advanced, green and socially inclusive. In addition to a number of projects with this objective, research focuses on IoT in the field of Smart City and on the investigation of specific application cases.
Scalability of IoT systems With the large number of sensors, actuators and devices, the scalability of the entire IoT architecture is an open question in research. It is unclear how and if one should scale individual components so that both resources and costs increase according to the effort. The chair conducts research on mobile communication architectures that connect IoT devices.
5CALE - Massive scaling of fully virtualized 5G mobile core networks in the context of IoT
(Februray 2019 - January 2022)
Within the project, new scaling methods and resource management approaches are being developed for the next generation 5G mobile communication network with IoT traffic. It is funded within the 5G call "Digitale Offensive" of the Bavarian Ministry of Economic Affairs with a total budget of one million euros and a term of 3 years.
Grigorjew, A., Metzger, F., Hoßfeld, T., Specht, J.: A Simulation of Asynchronous Traffic Shapers in Switched Ethernet Networks. Workshop for Advanced Communication Networks for Industrial Applications (2019).
Metzger, F., Hoßfeld, T., Bauer, A., Kounev, S., Heegaard, P.E.: Modeling of Aggregated IoT Traffic and Its Application to an IoT Cloud. Proceedings of the IEEE. 107, 679-694 (2019).
Geißler, S., Lange, S., Wamser, F., Zinner, T., Hoßfeld, T.: KOMon - Kernel-based Online Monitoring of VNF Packet Processing Times. 2019 International Conference on Networked Systems (NetSys), Best Paper Award (2019).
Lange, S., Linguaglossa, L., Geißler, S., Rossi, D., Zinner, T.: Discrete-Time Modeling of NFV Accelerators that Exploit Batched Processing. IEEE Conference on Computer Communications (Infocom) (2019).
Grigorjew, A., Gray, N., Hoßfeld, T., Shukla, A., Zinner, T.: Bridging the Gap Between Programming and Implementation of Networking Devices. Student Workshop of the 14th International Conference on emerging Networking EXperiments and Technologies (2018).