Abstract

A very large amount of data must be used to reasonably measure the available network bandwidth in a Grid by simply checking the time that it takes to send it across the network with TCP. The Network Weather Service (NWS) is the most common tool for obtaining transfer delay predictions from network measurements in Grids. We show that, in simple tests in a real Grid, the results that it obtains are not good enough or require heavily loading the network. The point of this study is to illustrate the need for more sophisticated and appropriately designed network measurement tools.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call