HULA

TLDR

Datacenter networks use multi‑rooted topologies such as Leaf‑Spine and Fat‑Tree to deliver high bisection bandwidth, relying on multipathing and requiring load‑balancing mechanisms like ECMP to distribute traffic evenly across paths. The study proposes congestion‑aware load‑balancing techniques, exemplified by CONGA, to address ECMP’s shortcomings. These techniques suffer from limited scalability due to constrained switch memory that restricts congestion‑tracking state at edge switches, and from inflexibility because they are implemented in custom hardware that cannot be modified in the field.

Abstract

Datacenter networks employ multi-rooted topologies (e.g., Leaf-Spine, Fat-Tree) to provide large bisection bandwidth. These topologies use a large degree of multipathing, and need a data-plane load-balancing mechanism to effectively utilize their bisection bandwidth. The canonical load-balancing mechanism is equal-cost multi-path routing (ECMP), which spreads traffic uniformly across multiple paths. Motivated by ECMP's shortcomings, congestion-aware load-balancing techniques such as CONGA have been developed. These techniques have two limitations. First, because switch memory is limited, they can only maintain a small amount of congestion-tracking state at the edge switches, and do not scale to large topologies. Second, because they are implemented in custom hardware, they cannot be modified in the field.