Servicenavigation

Project Details

Back

Projekt Print View

Design and Performance Evaluation of Future Internet Routing Architectures

Subject Area Security and Dependability, Operating-, Communication- and Distributed Systems
Term from 2012 to 2018
Project identifier Deutsche Forschungsgemeinschaft (DFG) - Project number 214843191
 
Final Report Year 2023

Final Report Abstract

State of the art routing technologies struggle to meet the modern demands of emerging and established services. The scope of this project was to investigate novel routing architectures for the Internet. Focus were resilient and scalable forwarding structures, multicast, and LoC/ID split to increase the flexibility and capabilities of routing in the future. Fast reroute (FRR) mechanisms are state of the art technologies to quickly protect traffic against component failures. However, existing mechanisms cannot protect all destinations or require extensive overhead, e.g., in terms of large headers or many additional forwarding rules. In this project we enhanced loop-free alternates (LFAs). LFAs are one of the most prominent unicast FRR mechanisms due to their simplicity. That is, when a node cannot forward traffic on the default path, traffic is send to a suitable other neighbor instead. This neighbor has to have a working shortest path towards the destination despite the failure. However, LFAs cannot protect all destinations and may create loops under severe failure conditions. We increase the protection capabilities of LFAs by adding explicit protection tunnels. Furthermore, we introduce advanced loop detection (ALD) to detect and prevent all loops. We implement all mechanisms on high-performance hardware and perform extensive evaluations to investigate the properties of the proposed mechanisms and compare them to state of the art technologies. IP Multicast (IPMC) is an important technology to enable scalable one-to-many traffic. However, traditional IPMC requires stateful core routers to implement forwarding structures. This causes serious scalability issues. The IETF proposed bit index explicit replication (BIER) as an efficient transport mechanism for IPMC without the beforementioned downsides. We extend BIER with FRR capabilities (BIER-FRR), implement BIER and BIER-FRR on high-performance hardware, and perform significant evaluations to better understand the scalability and capabilities of BIER. Thereby, we significantly contributed to the development of BIER in the IETF. We developed xRAC to enable fine granular steering of application traffic. For example, this is highly important in high-security environments, where the user may need to use an internet browser but the machine itself and other applications should not be connected to the internet. xRAC is currently in the progress of being patented. During the research project we came in contact with the novel technology P4 which stands for programming protocol-independent packet processors. It is a programing language and architecture to describe data planes on a high level. Device manufacturers provide target devices and specific compilers that map P4 programs to the pipeline of their devices. This enables user with fast and independent evolution of their networking infrastructure. We recognized the growing importance of P4 and performed a comprehensive literature study to provide a suitable entry point into the P4 ecosystem for novel users and reviewed hundreds of P4-related work to analyze the capabilities of P4 and identify potential future research areas.

Publications

 
 

Additional Information

© 2026 DFG Disclaimer / Copyright / Privacy Policy

Textvergrößerung und Kontrastanpassung