Trimming the Multipath for Efficient Dynamic Routing

Multipath routing is a trivial way to exploit the path diversity to leverage the network throughput. Technologies such as OSPF ECMP use all the available paths in the network to forward traffic, however, we argue that is not necessary to do so to load balance the network. In this paper, we consider multipath routing with only a limited number of end-to-end paths for each source and destination, and found that this can still load balance the traffic. We devised an algorithm to select a few paths for each source-destination pair so that when all traffic are forwarded over these paths, we can achieve a balanced load in the sense that the maximum link utilization is comparable to that of ECMP forwarding. When the constraint of only shortest paths (i.e. equal paths) are relaxed, we can even outperform ECMP in certain cases. As a result, we can use a few end-to-end tunnels between each source and destination nodes to achieve the load balancing of traffic.

💡 Research Summary

The paper addresses the problem of efficient multipath routing in modern data‑center networks, where the traditional approach—OSPF Equal‑Cost Multipath (ECMP)—uses all available equal‑cost shortest paths to spread traffic. While ECMP is widely deployed because of its simplicity and distributed nature, the authors argue that using every possible shortest path is unnecessary for achieving good load balance and can even cause congestion on overlapping links. To explore this, they propose a model that limits the number of end‑to‑end paths used for any source‑destination pair to at most k (where k can be a fixed parameter or determined adaptively). The objective is to minimize the maximum link utilization across the network, which directly translates to reducing the worst‑case congestion.

Key contributions

1. Path‑selection cost function – They define a cost for a candidate path p as
   \