Study of the impact of self-similarity on the network node traffic/Srauto savastingumo itakos kompiuteriu tinklo mazgo darbui tyrimas

The article analyses a stochastically bounded the GI/G/m//N circuit switched network model with packet losses, with stochastic input network traffic, stochastic served network node, and deterministic and finite network node buffer capacity. Max- plus algebra instrumentality is used for the network processes analysis. FIFO tail drop or LIFO tail drop buffer is used. We have established that the average waiting time in the queue had increased when the queue service discipline was FIFO as compared with LIFO, while the offered traffic was Poisson and the served in the node traffic was self-similar. The network traffic is served faster in the network node with the buffer queue discipline LIFO, while the offered traffic is Poisson and its intensity exceeds the served in the node traffic 10 times. Ill. 2, bibl. 24 (in English; abstracts in English and Lithuanian). Analizuojamas stochastiskai apribotas GI/G/m//N tinklo modelis su paketu nuostoliais, stochastiniu itekanciuoju tinklo srautu bei determinuota ir baigtine tinklo mazgo buferio talpa. Tinklo proces? analizei panaudotos max-plius algebros priemones. Tinklo mazguose naudota buferine FIFO tail drop ir LIFO tail drop atmintis. Nustatyta, kad vidutine paraiskos laukimo eileje trukme yra ilgesne, kai eiles aptarnavimo disciplina yra FIFO, itekantysis srautas--Puasono, o mazgo aptarnavimo srautas pasizymi savastingumu. Kai itekantysis srautas yra Puasono ir jo intensyvumas daugiau nei 10 kartu virsija tinklo mazgo pralaiduma, geriau tinklo srauta aptarnauja tinklo mazgas su LIFO eiles aptarnavimo disciplina. Il. 2, bibl. 24 (anglu kalba; santraukos anglu ir lietuviu k.).