Forecasting in neural network

I have data of hourly usage of a lan. I have to predict how much bandwidth is it using hourly.

x-axis = 1,2,3..... 24 hours of a day y axis = percentage usage among all other lans in that hour.

I have to use neural network for forecasting the usage. what nn to use-- back propagation or time series. I am new to this. So please help. a sample example can be a help.

[Greg Heath]

Insufficient explanation.

How much data do you have?

How much do you want to predict ahead?

Is the total number of lans constant?

[Meghna]

i am collecting data... i have around 3 months data- per day-per hour.

I want to predict the next day ahead.

yes total number of lans are constant.

I have to predict for each one of them and then decide which uses the maximum in that particular hour of the day.

I am confused between time series of fitnet?? I used narxnet but its not giving good result.

i read an example from file exchange of electricity load forecasting. Will that be helpful??