From Multimedia Communications to the Future Internet

In the 1980s multimedia became one of the most prominent research areas in systems and communication. Early work around these topics were carried out at the European Networking Center (ENC) in Heidelberg. The so called HeiProjects formed a project suite where the most relevant research topics in this space were addressed. It was, for instance, investigated what the specific requirements of multimedia are and how they can be supported through networks and operating systems (OS). The concept of streaming to provide a good user experience was also introduced. This was initially associated with Quality of Service (QoS). MHEG (Multimedia and Hypermedia information coding Expert Group) was another relevant topic to support multimedia throughout the networked system context and Quality of Experience (QoE) became a topic of interest. Ralf Steinmetz and the team were also some of the first to address synchronisation in all the various forms within a multimedia system. In this chapter, some of this early research is recalled and the lessons learned are discussed.

Digital multimedia systems would not have been successful without understanding synchronization and they would not have been able to compete with analog television and radio quality if digital media such as audio and video would not have been properly synchronized. When first digital multimedia systems emerged in 1980^th, it was not clear (a) what the delay thresholds should be to lip-synchronize the various media, (b) what the hardware-software frameworks should be to achieve synchronization goals, known from analog TV and radio, (c) what new synchronization primitives and protocols we need besides mutexes and semaphores to synchronize multiple modalities in multimedia systems.

The in-depth answers to these questions were formulated by Ralf Steinmetz and his collaborators in the early 1990s. In his early synchronization work we find many fundamental solutions on which current multimedia systems relied in the past and rely even today. In this paper, we will first explain the early synchronization results from Steinmetz’ work and then show examples of the impact that his synchronization work had on the multimedia community.

Adaptivity is a cornerstone concept in video streaming. Equipped with the concept of Transitions, we review in this paper adaptivity mechanisms known from classical video streaming scenarios. We specifically highlight how these mechanisms emerge in a specific context, such that their performance finally depends on the deployment conditions. Using multiple examples we highlight the strength of the concept of adaptivity at runtime for video streaming.

In this paper, we review the development of fine-grained quality of service (QoS) models for the Internet protocol stack. We discuss the major developments and milestones arising from substantial research work in both signalling and packet scheduling. These approaches were often coupled with the aim of providing effective multicast group communication in the network layer. However, despite significant intellectual contributions and technological advancements, production networks evolved in a different direction. The paper reflects on the driving forces for this development. The aim is to summarize these developments, provide longer-term insight that might help in assessing current research, as well as highlight the work that originated in Ralf Steinmetz’s group at TU Darmstadt.

This chapter explores the evolution of computer networks from closed data transmission systems to complex, programmable infrastructures critical to modern life and industry. It highlights the growing demand for flexibility and adaptability in networks due to diverse and unpredictable application requirements. Key enablers of this are the concepts Software Defined Networking (SDN), Network Functions Virtualization (NFV), and programming concepts for processors as well as reconfigurable hardware. The chapter delves into the importance of network programmability in meeting future network demands, providing a foundation for understanding the terminology and applications of softwarized networks.

This review builds upon the paper “Europe’s Mission in Next-Generation Networking,” originally published in “The GENI Book” in 2016 [1]. While the earlier publication focused into European initiatives such as Framework Program 7 and the German Initiative German-Lab, this iteration broadens its scope to encompass the Future Internet Architecture (FIA) by DARPA, alongside the FIND and GENI projects by the National Science Foundation (NSF) in the US. Additionally, it includes insights from the MAKI project funded by the German Research Foundation (DFG).

The true value of research artifacts only becomes clear once they have been confirmed to exhibit their expected behavior in practice. Taking research results beyond analytical and theoretical studies alone has been in the DNA of the Multimedia Communications Lab (KOM) at TU Darmstadt ever since its inception. Be it small-scale prototyping, the assembly of exhibits, experimentation on building level, or Internet-scale experiments, KOM has traditionally been an environment that fostered practical innovation and encouraged academic staff and students alike to experiment. Besides a strong focus on the evaluation of technological aspects, experimentation also frequently involved human subjects in order to assess the usability and practical relevance of research artifacts.

In this article, we have curated a set of noteworthy highlights of almost thirty years of laboratory-level experimentation on topics related to multimedia communications. This way, we showcase the diversity of topics as well as the diligence and dedication that has gone into ensuring their appealing demonstration. By highlighting the breadth of activities performed by more than a hundred researchers over three decades, we create a memento of the laboratory work performed at KOM.

Today, Service-oriented Computing (SOC) is a major design and implementation approach for a wide range of applications, ranging from Everything-as-a-Service (XaaS) in the Cloud to service-based machine learning approaches like federated learning. The Multimedia Communications Lab (KOM) at TU Darmstadt has been one of the early contributors to this research field.

With the retirement of Ralf Steinmetz, it is time to look back at the areas where people from KOM have contributed to SOC, and which impact this research has generated. Therefore, this paper presents 10 years of SOC research at KOM, roughly covering the years 2006 to 2016.

Machine Learning (ML) and its application is a research area that has become increasingly important, especially in the last decade. ML approaches in the field of supervised learning depend on labeled data sets for their training process. But, in some problem domains, the collection and generation of labeled training data can be hard because, for example, the underlying problem is in general hard to solve algorithmically. This paper proposes a conceptual framework for the generation of labeled training data sets for supervised learning approaches based on the GIPS framework. For this purpose, GIPS combines Graph Transformation (GT) with Integer Linear Programming (ILP) techniques to solve graph-based optimization problems to obtain labels for data points. A prototypical implementation is used to demonstrate the functionality of our solution in the context of a small-scale Virtual Network Embedding (VNE) example. The evaluation of our prototype shows promising results when compared to an optimal ILP-based implementation.

Applications sensitive to latency, such as video web meetings, are particularly vulnerable to end-to-end delays. Network congestion greatly contributes to end-to-end delay, and it can lead to packet loss and jitter. Congestion control mechanisms regulate data flow to mitigate these delays, often resulting in queues at the sender’s end. For applications prioritizing timeliness over reliability, discarding outdated video frames from the queue reduce queue size and subsequent delays. However, this frame-dropping practice can introduce visual artifacts. This study proposes a deadline-centric approach to video transmission that strategically removes frames exceeding a certain waiting time from the queue, aiming to strike a good balance between latency reduction and video quality preservation.

The success of both the Internet and wireless communication is exceptional. The combination of these technological achievements made today’s digital world—that relies on wireless networks and is deeply entrenched into everyday live—possible. Within the Collaborative Research Center (CRC) MAKI, several research challenges in the area of wireless networking have been addressed and solved. We present a selection of these challenges, the corresponding solutions, and how MAKI made the solutions possible.

Peer-to-peer (P2P) overlays have strongly influenced scalable communication infrastructures. In this paper we study design decisions in the context of publish/subscribe, a central communication paradigm for building distributed applications at scale. In particular, we highlight an important aspect of adaptively managing decentralized peer-to-peer systems called fairness. Fairness accounts for a fair contribution of peers dependent on the benefit of using a service. For this, we introduce f-fairness, a fairness metric accounting for unbalanced workloads in publish/subscribe systems. To support f-fairness, we propose and review adaptation mechanisms that support building membership and event distribution for publish/subscribe.

Over the last three decades, computer systems completely changed from large-scale and mainly isolated systems to interactive, often mobile devices. Such devices are equipped with various sensors to gather information about their environment. Further, pervasive communication technology is required, i.e., changes to the dynamics of the environment, such as changing the quality or availability of communication channels. Self-adaptive and self-organizing (SASO) systems can adjust their behavior to environmental changes to maintain their functionality and performance. Those systems provide adaptive behavior by integrating an adaptation control system, often relying on the so-called MAPE functionality, standing for the basic functionalities of monitoring the system and the environment, analyzing if adaptation is required, and planning and executing the adaptation. However, being a well-known concept that emerged around 20 years ago, the MAPE functionality misses an inherent integration of learning. Further, it focuses on reactive, central, coordinated adaptation reasoning, which is not always beneficial and contradicts decentralized, local decision-making. This article analyzes those shortcomings and presents a new system model for integrating learning into the MAPE cycle, focusing on proactive adaptation in distributed systems. The paper concludes with a discussion of research challenges using the running example of a communication system.

We present major results from our research on user-generated video streaming for transition-enabled networks, i.e., networks with mechanisms for smooth transitions under changing conditions. We concentrate on the application layer with multiple real-time videos as our source. Algorithms are presented to select the best video source at any given time and to transmit only that video to a set of receivers. A novel video composition component is proposed that receives the input videos and sends only one output video to the viewers. An example for an application would be a concert with an audience where many persons transmit video at the same time, from different perspectives. The transitions in the network are performed in real-time in order to always stream the highest-quality video to the receivers.

Exploratory testing (ET) has been introduced in the academic world as an agile testing methodology. It has so far been considered in the context of several types of software systems, but there is a lack of scientific foundation, when ET is applied to safety-critical large-scale embedded systems. We show that safety-critical large-scale embedded systems possess properties that are quite different to pure software systems - with regard to among others the user interface, involved hardware, environmental influences and the role of software safety standards. ET relies on the notion of tours. Some of the existing tours are described and their relevance for safety-critical large-scale embedded systems is discussed. The main contribution of this paper is the theoretical preparation of empirical work on ET tours applied to safety-critical large-scale embedded systems. Throughout this paper spacecrafts serve as the example of choice.

The Internet fosters democratic principles, bringing information, services, and social contacts to everyone’s fingertips. Conversely, rampant monopolization, exploitation, and manipulation trends cause doubts about the democratization effects of the Internet. We argue that key functionality atop the core “Transport Internet” shall be made publicly available and governed as meta-services. Thereupon, innovators shall be able to build real services easily for a competitive market, while the underlying Meta-services ensure interoperability and are publicly governed to counter undemocratic and antisocial mechanisms. The proposed meta-services represent rather well-known functionality, and the proposed list is not comprehensive. What is new is their joint consideration as pillars of “democratic computing”. We introduce our vision as a basis for discussions about the feasibility, potential benefits, and threats, and about the most pressing meta-services to start with. We propose a public ledger (largely equivalent to an existing blockchain yet publicly governed) as underlying technology along with the Web and the core Transport Internet, with the following meta-services atop: neutral edge computing for the provision of storage and computing; trust management; AI serving; persistent metaverse; and mission-critical provisioning. For each meta-service, we will discuss smaller or larger steps that were already made towards their conceptualization.

In this paper we discuss the basis for future 6G mobile architectures ranging from the physical layer up to the application. Building on advances form radio technologies, the core of our architecture lies on the functionality that allows collecting and analyzing data from the network. Leveraging this data analytics functionality, we envisage two applications. One lies in using the data to improve different aspects of network operation comprising control, management and orchestration. The other takes advantage of the collected data for applications outside the network, setting the foundation for a data-based economy.