Research
Research approach
To tackle the challenges the current development of the Internet poses, our research approach is to collect data from operational networks, analyze them using big-data analytics to identify invariants, revisit assumptions, and detect where current performance bottlenecks in the Internet architecture are. Hereby, we use simulation environments to validate our analysis results and to enable “what-if” studies. These insights are then the foundation for shaping the future Internet via optimizations and alternative designs. Hereby, we consider protocol enhancements, novel network management tools, and how to take advantage of concepts from software-defined networking.
The research group tackles the following topics:
- Development of Defence Methodologies for High-Speed Networks: High-speed networks are inherently fragile constructs but also the backbone of many decision-making processes in economy and politics. Finding ways to stabilize the network performance even under duress without impeding speed and bandwidth therefore is a precondition of further development of network-using applications.
- Digital Infrastructure & Human Factors: Over the past half century, the rise and accelerating evolution of digital technology shaped
- society as much as society shaped technology. We now find ourselves in a world where digital infrastructure brings as much harm as it brings remidies, ranging from security issues in critical infrastructure to public values being threatened by platformization. Yet, we are still far from a true end-to-end understanding of these complex interaction effects in digital infrastructures, ranging from delayering technical complexity to socio-technical questions towards understanding those who ultimately build and maintain digital infrastructure.
- Future Internet: Here the focus is on how to evolve the Internet to overcome its current limitations, e.g., by adding better network management capabilities and enhancing the capabilities of CloudNets.
- Internet Measurement: Based on continuous observations of the Internet the goal of this group is to understand how the Internet evolves. But it also involves understanding the interdependencies between the network elements and the applications as well as the impact of failures.
- Online social networks: In this research area, we aim to analyze activity on online social networks to understand emerging socio-technical issues, develop tools/techniques to detect potentially harmful information (e. g. hatespeech, fake news etc.) and develop and assess the effectiveness of various mitigation strategies.
- Wide area data analytics: The challenge here is to enable wide area data analytics via collaborative data processing pipelines.
Analysis of the Covid-19 pandemic
In January 2020 the COVID-19 pandemic, a coronavirus variant spreading across the entire globe, reached central Europe. It became quickly apparent that this development would have an immense impact on our every day life, including the use of the internet. In our research, we took a closer look at how the pandemic changed our use of the internet.
Benchmarking congestion-control algorithms
Despite four decades of research on congestion-control algorithms, there is yet no “one-size fits-all” solution. There are at least thirty different congestion-control algorithms or schemes as of today, and new ones are still being proposed every so often. Determining how any one scheme falls short in comparison to the rest, and, most importantly, and along what dimensions are quite difficult to answer.
Digital Infrastructure
Over the past half century, the rise and accelerating evolution of digital technology shaped society as much as society shaped technology. We now find ourselves in a world where digital infrastructure brings as much harm as it brings remidies, ranging from security issues in critical infrastructure to public values being threatened by platformization. Yet, we are still far from a true end-to-end understanding of these complex interaction effects in digital infrastructures, ranging from delayering technical complexity to socio-technical questions towards understanding those who ultimately build and maintain digital infrastructure. With our research, we fill this knowledge gap about the technological bedrock of society by attaining an end-to-end understanding of building, maintaining, and ultimately living with digital infrastructure.
Emerging Platforms/Communities on the Web
The Web consists of numerous Web communities, news sources, and services, which are often used by various actors for potentially nefarious purposes. At the same time, it has become increasingly easier to create new platforms and communities, making the Web become a bigger and more complex ecosystem. We aim to study emerging platforms and Web communities that can have a large impact and influence both on the online and offline world and contribute to the research community by making publicly available datasets from social networks.
Flow queries
Many network operations, ranging from attack investigation and mitigation to traffic management, require answering network-wide flow queries in seconds. Although flow records are collected at each router, using available traffic capture utilities, querying the resulting datasets from hundreds of routers across sites and over time remains a significant challenge due to the sheer traffic volume and distributed nature of flow records. In this project, we investigate how to improve the response time for apriori unknown network-widequeries.
Future-proof the Internet
The continually evolving nature of the Internet ecosystem routinely introduces new,unforeseen challenges. With the increasing adoption of sensors and Internet of Things (IoT) devices, we are generating an unprecedented volume of data. As if this immense amount of data was not enough, such data now comes from divers endpoints, widely distributed throughout the network (at its edge). Still, this opens new opportunities, e.g., for scalable DDoS analysis. The changes in the Internet ecosystem also provide new opportunities to revisit some long standing networking problems and design new practical, scalable solutions that exploit or leverage these recent changes, e.g., DDoS mitigation, revisiting congestion control, or exploiting new opportunities for video transmissions.
Internet of Things
The number of IoT devices deployed within homes is increasing rapidly. It is estimated that the IoT population will increase to 20 billion by 2025. Such devices include virtual assistants, cameras, TVs, and smart home control devices. While users deploy some IoT devices explicitly, they are often unaware of the security threats and privacy consequences of using such devices.
Internet Security
As the Internet is becoming more and more ubiquitous in people’s everyday lives—pushed even more so by the COVID-19 pandemic—aspects of security in Internet-connected systems are increasing in importance. In addition, the deployment of IPv6-enabled devices in the Internet is steadily increasing, with more than 40% of Google users using IPv6 as of March 2023. To analyze the security of Internet-connected devices,we devise measurement techniques to reach many different devices in the IPv4 and IPv6 Internet, and we conduct Internet measurements and characterize different aspects of security of these devices.
Routing & Network management
Availability, accessibility, and reactivity are the cornerstones of modern web applications. Those characteristics require the underlying network infrastructure to operate not only extremely efficient but also highly robust and flexible. We intensively study the internal and external (routing-related) techniques that networks use to achieve those goals and use our insights to shape the traffic delivery of the future.
Online Content Moderation
Online content moderation is an important aspect of online social networks as it ensures that content posted by users abides by the platform’s guidelines and is appropriate for other users. due to the complexity and black-box nature of content moderation there are several open research avenues, such as the effectiveness of specific moderation strategies or how to maximize the output of content moderation and minimize the mental toll on human moderators from continuous exposure to harmful content.
Online Hate Speech & conspiracy theories
The spread of offensive language, hate speech and conspiracy theories online is an important and timely issue that exists on social networks. The continuous exposure to hateful content can have catastrophic consequences as it can lead to user radicalization and real-world violence (e.g., hate attacks against specific demographic groups). Despite the problem’s importance and societal implications, there are several research gaps in understanding, detecting, and mitigating the spread of hate speech. In this line of work, we aim to study and understand specific forms of hate speech (e.g., Anti Semitism, anti-chinese sentiments, etc.) and develop techniques to automatically detect instances of hate speech.
Understanding, Detecting, and Mitigating Weaponized Information
Weaponized information refers to carefully crafted information that aims to deceive people or information that is presented in such a way to manipulate or attack users. The dissemination of weaponized information at large scale can have devastating effects on the online and offline world. Motivated by the importance and impact that weaponized information can have to society, in this line of research, we aim to understand, design techniques to automatically detect instances of weaponized information, and investigate possible mitigation strategies.
Verification of Programmable Networks
The underlying networks serving the online services and applications of a business must be reliable, secure, and highly available. Outages caused by defective network devices, misconfigurations, or security breaches can lead to expensive losses. Consequently, network verification is a crucial task in the network management and control process.
Video streaming
Video streaming is next to omnipresent today and an essential part of our everyday life. Nevertheless, the conditions under which user stream videos is not always ideal. The result is poor visual quality, as the underlying algorithms try to select a quality of video that is small enough (meaning reduced amount of data that has to be transferred to the user) to be streamed under all current network conditions or the video does not download at all. Our research found a way to mitigate this problem.