Detection Collision Flows in SDN Based 5G Using Machine Learning Algorithms

The rapid advancement of wireless communication is forming a hyper-connected 5G network in which billions linked devices generate massive amounts data. traffic control and data forwarding functions are decoupled software-defined networking (SDN) allow the to be programmable. Each switch SDN keeps track information flow table. switches must search table for rules that match packets handle incoming packets. Due obvious vast quantity centres, capacity restricts plane’s capabilities. So, from across whole network. depends on Ternary Content Addressable Memorable Memory (TCAM) storing quick regulations; it restricted owing its elevated cost energy consumption. Whenever abused overflowing, usual regulations cannot executed quickly. In this case, we consider low-rate overflowing causes collision installed consumes excessive existing by delivering don’t fit at low rate. This study introduces machine learning techniques detecting categorizing flows SDN, using Feed Forward Neural Network (FFNN), K-Means, Decision Tree (DT). We two topologies, Fat Simple Topologies, with Mininet simulator coupled OpenDayLight (ODL) controller. efficiency efficacy suggested algorithms assessed several assessment indicators such as success rate query, propagation delay, overall dropped packets, consumption, bandwidth usage, latency rate, throughput. findings showed technique tackle congestion problem minimizes number while retaining statistical consistency By putting proposed method checking whether packet may move point A B without breaking certain regulations, evaluation tool examines every against set criteria. FFNN DT K-means obtain accuracies 96.29% 97.51%, respectively, identification flows, according experimental outcome when associated methods literature.