ARTICLE
TITLE

Adaptive Algorithm for Virtual Connection on Feedback-Based Flow Control in WAN-ATM Network

SUMMARY

Algoritma pengendalian aliran yang bersifat adaptif dapat diterapkan pada jaringan ATM wilayah luas guna mengalokasikan kebutuhan lebar pita pada hubungan virtual. Pengendalian aliaran menggunakan mekanisme control lup tertutup, yaitu suatu mekanisme kontrol yang bersifat reaktif dan secara dinamis, mengatur kecepatan transmisi sel yang diizinkan untuk setiap hubungan virtual dengan menggunakan informasi umpan balik dari jaringan. Informasi umpan balik berupa bit yang dapat mengidentifikasi apakah panjang antrian di buffer berada di atas atau di bawah batas ambang yang diberikan. Dengan bit indikasi ini kecepatan sumber pengiriman dapat dikendalikan. Penelitian yang telah dilakukan sebelumnya terbatas pada jumlah hubungan sebanyak 24VC dan kecepatan transmisi 155 Mbps dengan batas ambang buffer 6.5 sel. Karena itu, dilakukan penelitian jika jumlah hubungan bertambah hingga 100 VC dan kecepatan transmisi 600 Mbps dengan batas ambang buffer dalam rentang yang lebar. Hasil simulasi algoritma adaptif menunjukkan bahwa dengan menggunakan kriteria kestabilan terhadap hubungan virtual tunggal akan diketahui rentang penggunaan parameter rasio penguatan terhadap peredaman yang tepat sehingga untuk hubungan virtual banyak akan lebih mudah ditunjukkan sifat keadilan terhadap setiap hubungan. Sifat kekokohan ditunjukkan melalui simulasi, yaitu dengan menambah jumlah hubungan yang mungkin dapat dilayani. Dari respon kecepatan aggregate terlihat bahwa terjadi penurunan pemanfaatan lebar pita jaringan (<90%). Keadaan ini dapat diatasi dengan memperbesar kecepatan transmisi, menambah waktu update, dan menambah panjang ambang batas antrian di buffer. Adaptive flow control Algorithm can be applied into wide area ATM network to allocate lebar pita to virtual connection. Flow control uses closed loop control mechanism, i.e, a reactive control mechanism which dynamically controls permissible transmission speed for any virtual connection by using feedback information of the network. Information of the network is in the form of bit that can identify whether a queue length in buffer is above or below a given threshold. By this indication bit the source of cell sending can be controlled. The earlier investigation was limited to 24 VC connection and in 155 Mbps transmission speed with 6.5 cells buffer threshold. Considering that it was not enough, a further investigation was carried out with increasing connection up to 100 VC, 600 Mbps transmission speed, and buffer threshold in a wide range. Adaptive algorithm simulation results showed that by using stability criterion on single virtual connection we found the range of gain ratio parameter used on proper damping, hence, fairness can be made easier in multiple virtual connections. There are two parameters - each is important for every virtual connection, i.e minimum bandwith parameter and weight factor to gain bandwith share that provided by the network. Proper selection on the two parameters gives different service. Aggregate speed response showed that there is degradation of network bandwith use (<90%). This situation can be overcome by increasing the transmission speed, lengthening update time, and enlarging the buffer threshold.

 Articles related

Shaobo Li,Chenglong Zhang,Jinglei Qu    

The production process of modern manufacturing industry is complex and changeable, manufacturing resources have extensive dynamic characteristics. For effectively managing and controlling manufacturing resources, realizing real-time location data collect... see more


Lieping Zhang,Fei Peng,Peng Cao,Wenjun Ji    

Aiming at the low accuracy of DV-Hop localization algorithm in three-dimensional localization of wireless sensor network, a DV-Hop localization algorithm optimized by adaptive cuckoo search algorithm was proposed in this paper. Firstly, an improved DV-Ho... see more


Ming Li,Huanyan Qian    

The deployment of mobile sensor networks is a hotspot issue. This paper is dedicated to proposing a distributed, adaptive, and scalable mobile sensor network algorithm based on the theory in electrostatic field. In the algorithm, the obstacles and nodes ... see more


Tran Vu TU,Kazushi SANO    

This paper firstly proposes an improved genetic algorithm (GA) for optimization in adaptive bus signal priority control at signalized intersections. Unlike conventional genetic algorithms with slow convergence speed, this algorithm can increase the conve... see more


Amalina Abdullah, Channarong Banmongkol, Naebboon Hoonchareong, Hidaka Kunihiko    

Issues on detecting the occurrence of a fault, justifying the type, and estimating the exact location of the fault should be resolved to eliminate faults promptly and restore power supply with minimum interruption. Conventional approaches have contribute... see more