Tactical Traffic Engineering with Segment Routing Midpoint Optimization
Alexander Brundiers, Timmy Schüller, Nils Aschenbruck · 2023
Tactical Traffic Engineering (TE) plays a crucial role in the operation of modern backbone networks as it enables operators to quickly react to failures or unforeseen traffic changes. In this paper, we propose a new Segment Routing (SR)-based optimization algorithm called MOLS. It is the first algorithm that applies the recent concept of Midpoint Optimization (MO) for SR to the use case of fast, tactical TE. In an extensive evaluation, based on various real-world topologies, we show that our algorithm performs virtually on par or better than comparable state-of-the-art tactical TE approaches that rely on conventional SR. However, especially for larger networks, the number of SR policies required by our algorithm is substantially lower. This not only reduces the introduced overhead but also allows for faster deployment of the computed configurations since less changes have to be applied to the network. Furthermore, we show that MOLS is able to remove congestion in sub-second fashion for multiple TE use cases. Lastly, MOLS also is the first algorithm in literature to fully utilize the capabilities of MO without any artificial limitations. This enables it to find similar or even better solutions than the only other MO-capable algorithm in literature in just a fraction of the time.