Research Area:  Wireless Sensor Networks
Modern data center networks are usually constructed in multi-rooted tree topologies, which require the highly efficient multi-path load balancing to achieve high link utilization. Recent packet-level load balancer obtains high throughput by spraying packets to all paths, but it easily leads to the packet reordering under network asymmetry. The flow-level or flowlet-level load balancer avoids the packet reordering, while reducing the link utilization due to their inflexibility. To solve these problems, we design a Queueing Delay Aware Packet Spraying (QDAPS), that effectively mitigates the packet reordering for packet-level load balancer. QDAPS selects paths for packets according to the queueing delay of output buffer, and lets the packet arriving earlier be forwarded before the later packets to avoid packet reordering. Moreover, we adopt the power-of- n-choices paradigm on QDAPS to alleviate the impact of herd behavior under multiple forwarding engines. We compare QDAPS with ECMP, LetFlow and RPS through NS2 simulation and Mininet implementation. The test results show that QDAPS reduces flow completion time (FCT) by ~30%-50% over the state-of-the-art load balancing mechanism.
Author(s) Name:  Jiawei Huang; Wenjun Lyu; Weihe Li; Jianxin Wang; Tian He
Journal name:   IEEE/ACM Transactions on Networking
Publisher name:  IEEE
Volume Information:  ( Volume: 29, Issue: 3, June 2021) Page(s): 1183 - 1196
Paper Link:   https://ieeexplore.ieee.org/abstract/document/9352527