Research Article
ALBLP: Adaptive Load-Balancing Architecture Based on Link-State Prediction in Software-Defined Networking
| Input: | | The network link-state value obtained periodically; | | Output: | | Flow entries used to control the SDN switch; | (01) | Initialize the List of network link sequences List_in = []; | (02) | Initialize the Data prediction flag flag_dp = True; | (03) | Obtain the network link-state value and assign it to R; | (04) | Obtain the length of and assign it to ; | (05) | List_in.append (R); | (06) | if size_of (List_in) <= Tfthen | (07) | Graph = set_graph (R); | (08) | send_flow_entries (DijKstra(Graph)); | (09) | else if size_of (List_in) >= Tf and flag_dp = = True then | (10) | X = standardization (List_in[−Tf, :]); | (11) | Y = de_standardization (LSTM_forecast(X)); | (12) | Y = Y [−int(Tf − Td),:]; | (13) | flag_dp = False; | (14) | else then | (15) | Graph = set_graph (Y[0]); | (16) | Y = delete (Y[0]); | (17) | send_flow_entries (DijKstra(Graph)); | (18) | if size (Y) = = 0 then | (19) | flag_dp = True; | (20) | end if | (21) | end if |
|