The Delay Measurement and Analysis of Unreachable Hosts of Internet


INTERNATIONAL ARAB JOURNAL OF INFORMATION TECHNOLOGY, vol.19, no.1, pp.63-71, 2022 (SCI-Expanded) identifier identifier identifier

  • Publication Type: Article / Article
  • Volume: 19 Issue: 1
  • Publication Date: 2022
  • Doi Number: 10.34028/iajit/19/1/8
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Applied Science & Technology Source, Arab World Research Source, Computer & Applied Sciences
  • Page Numbers: pp.63-71
  • Keywords: RTT distributions, ICMP-T3, ICMP, self-similarity, worm propagation, Hurst parameter, TIME
  • Kayseri University Affiliated: No


Delay related metrics are significant quality of service criteria for the performance evaluation of networks. Almost all delay related measurement and analysis studies take into consideration the reachable sources of Internet. However, unreachable sources might also shed light upon some problems such as worm propagation. In this study, we carry out a delay measurement study of unreachable destinations and analyse the delay dynamics of unreachable nodes. 2. Internet Control Message Protocol (ICMP) destination unreachable Internet Control Message Protocol-Destination Unreachable (ICMP T3) packets are considered for the delay measurement according to their code types which shows network un reach ability, host un reach ability, port un reach ability, etc., Measurement results show that unreachable sources exhibit totally different delay behaviour compared to reachable IP hosts. A significant part of the unreachable hosts experiences extra 3 seconds Round Trip Time (RTT) delay compared to accessible hosts mostly due to host un reach ability. It is also seen that, approximately 79% of destination un reach ability causes from host un reach ability. Obtained Hurst parameter estimation results reveal that unreachable host RTTs show lower Hurst degree compared to reachable hosts which is approximately a random behaviour. Unreachable sources exhibit totally different distributional characteristic compared to accessible ones which is best fitted with Phased Bi-Exponential distribution.