Title :
RTTometer: measuring path minimum RTT with confidence
Author :
Zeitoun, Amgad ; Wang, Zhiheng ; Jamin, Sugih
Author_Institution :
Dept. of Electr. Eng. & Comput. Sci., Michigan Univ., Ann Arbor, MI, USA
Abstract :
Internet path delay is a substantial metric in determining path quality. Therefore, it is not surprising that round-trip time (RTT) plays a tangible role in several protocols and applications, such as overlay network construction protocol, peer-to-peer services, and proximity-based server redirection. Unfortunately, current RTT measurement tools report delay without any further insight about path condition. Therefore, applications usually estimate minimum RTT by sending a large number of probes to gain confidence in the measured RTT. Nevertheless, a large number of probes does not directly translate to better confidence. Usually, minimum RTT can be measured using few probes. Based on observations of path RTT presented by Z. Wang et al. (see Proc. Passive & Active Measurement Workshop - PAM´03, 2003), we develop a set of techniques, not only to measure minimum path RTT, but also to associate it with a confidence level that reveals the condition on the path during measurement. Our tool, called RTTometer, is able to provide a confidence measure associated with path RTT. Besides, given a required confidence level, RTTometer dynamically adjusts the number of probes based on a path´s condition. We describe our techniques implemented in RTTometer and present our preliminary experiences using RTTometer to estimate RTT on various representative paths of the Internet.
Keywords :
Internet; delays; parameter estimation; time measurement; Internet path delay; RTT measurement tools; RTTometer; confidence level; minimum RTT estimation; overlay network construction protocol; peer-to-peer services; protocols; proximity-based server redirection; round-trip time; Application software; Current measurement; Delay estimation; Gain measurement; Internet; Peer to peer computing; Probes; Protocols; Statistical distributions; Statistics;
Conference_Titel :
IP Operations & Management, 2003. (IPOM 2003). 3rd IEEE Workshop on
Print_ISBN :
0-7803-8199-8
DOI :
10.1109/IPOM.2003.1251234