A detailed survey on heavy hitter detection in cloud data centre network

Heavy hitter detection is an essential function in cloud gateways, serving as the access points for network heavy hitters entering the cloud. Heavy hitters denote a limited group of flows or packets that utilise an excessive share of bandwidth, thereby impacting network performance. Real-time detection of heavy hitters is crucial for optimising network resource utilisation, mitigating network congestion, and ensuring quality of service. This survey presents a detailed examination of heavy hitter detection techniques utilised in cloud gateways. The techniques covered include sampling-based methods, sketch-based methods, hash-based methods, and learning-based methods. This document addresses the challenges related to heavy hitter detection in cloud gateways, including high-speed networks, dynamic heavy hitter patterns, limited resources, and the necessity for real-time detection. In conclusion, this document delineates prospective research avenues and discoveries, emphasising the necessity for enhanced algorithms for heavy hitter detection, the incorporation of learning and artificial intelligence methodologies, and the evaluation of novel network architectures.