ATM call admission control using sparse distributed memory

We have proposed a neural network call admission control (CAC) method based on sparse distributed memory (SDM). CAC is a key technology of ATM network traffic control. It should be adaptable to the rapid and various changes of the ATM network environment. Conventional approaches to the ATM CAC require network analysis in detail in all cases. The optimal implementation is said to be very difficult. Therefore, neural approaches have recently been employed. However, it does not meet the adaptability requirements. We, thus, have proposed a method which is based on SDM as the neural network controller. Since SDM is a RAM-like associative memory, it has the property of good adaptability. It provides CAC with good adaptability to manage changes. Experimental results are as good as those of the previous neural approaches without additional analytical data, and without relearning from initial state