Hasil Pencarian  ::  Simpan CSV :: Kembali

"The purpose of this book is to give the reader two things, to paraphraseMark Twain:Roots to know the basics of modeling networks and Wings to fly away and attemptmodeling other proposed systems of interest.The Internet phenomenon is affecting us all in the way we communicate, conduct business, and access information and entertainment. More unforeseen applications are still to come. All of this is due to the existence of an efficient global highperformance network that connects millions of users and moves information at ahigh rate with small delay."

"ABSTRAKThe aim of this work is to define a procedure for the modelling of a computer network for the simulation of the network traffic induced by the execution of web applications. The work consists in describing the objects of a library defined for modelling both the hardware and the software of the system. These are specific objects with an architecture based on complex queuing service structures. The objects of the library are used to both build the computer network model and describe the operation of a web application executed on the network in a distributed fashion."

"Pada penelitian ini diusulkan peramalan trafik jaringan menggunakan Artificial Neural Network dengan model Nonlinear Autoregressive. Model prediksi beban trafik dilakukan dalam tiga skenario yaitu tanpa input eksogen, dengan input eksogen jumlah pelanggan, dan dengan input eksogen jumlah pelanggan dan inflasi. Hasil penelitian dengan nilai MAPE dan MSE terkecil terdapat pada prediksi beban trafik dengan input eksogen jumlah pelanggan. Pada penelitian diprediksi beban trafik hingga l tahun kedepan untuk dapat merencanakan pembangunan dan peningkatan kapasitas node-b/ BTS 3G. Diharapkan dengan melakukan peramalan penggunaan-jaringan-oleh-pelanggan akan menghasilkan estimasi akurat permintaan kebutuhan pelanggan di masa mendatang sehingga organisasi dapat melakukan strategi yang tepat dalam merencanakan peningkatan kapasitas demi menjaga 4aality ofservice.

---

This research proposed network traffic forecasting using Artificial Neural Network with Nonlinear Autoregressive models. The traffic load prediction model is done in three scenarios: without exogenous input, with the input of exogenous number of customers, and with exogenous inputs the number of subscribers and inflation. The smallest MAPE and MSE values are in the traffrc load prediction with subscribers as exogenous inputs. The traffic load is predicted up to 1 year ahead in order to plan the development and improvement of the capacity of the node-b / 3G base stations. By forecasting the network usage generate by the customer, we expect to have an accurate estimated demand of customer needs in the future so that the organization can perform the right strategy for planning the capacity to maintain the quality of service."