The proposed knowledge of the internet traffic pattern measured on the actual network is very useful for the intelligent TE, e.g.,QoS routing and load balancing because the traffic congestion of Internet service providers (ISP) in Thailand is different in various periods proved in the experimental result in the section IV. This enables congested paths to be avoided as well as distribute traffic load over the multiple ADSL connections so that efficiency can be realized appropriately. From overview, the outcomes of three ADSL operators comparison are in the same direction, which are proved that TRUE ADSL operators in Thailand should provide the best performance compared to others.
Since there are several ADSL Internet providers in Thailand and appearing seasonal proved in the experiment result in the section V, an intelligent load balancer with neuro-fuzzy approach will be implemented using the proposed information about traffic characteristics including seasonality in this