On the Packet Delay Distribution in Power-Law Networks

2009 ◽  
Author(s):  
Takahiro Hirayama ◽  
Shin'ichi Arakawa ◽  
Ken-ichi Arai ◽  
Masayuki Murata
Keyword(s):  
Power Law ◽  
Packet Delay ◽  
Delay Distribution

A study of on/off timing channel based on packet delay distribution

2009 ◽  
Vol 28 (8) ◽  
pp. 785-794 ◽  
Author(s):  
Lihong Yao ◽  
Xiaochao Zi ◽  
Li Pan ◽  
Jianhua Li
Keyword(s):  
Packet Delay ◽  
Delay Distribution ◽  
Timing Channel

scholarly journals Asymptotic Delay Analysis for Cross-Layer Delay-Based Routing in Ad Hoc Networks

2007 ◽  
Vol 2007 ◽  
pp. 1-11 ◽  
Author(s):  
Philippe Jacquet ◽  
Amina Meraihi Naimi ◽  
Georgios Rodolakis
Keyword(s):  
Ad Hoc Networks ◽  
Wireless Ad Hoc Networks ◽  
Power Law ◽  
Ad Hoc ◽  
Buffer Size ◽  
Cross Layer ◽  
Power Law Distribution ◽  
Delay Distribution ◽  
Np Hard Problem ◽  
Hoc Networks

This paper addresses the problem of the evaluation of the delay distribution via analytical means in IEEE 802.11 wireless ad hoc networks. We show that the asymptotic delay distribution can be expressed as a power law. Based on the latter result, we present a cross-layer delay estimation protocol and we derive new delay-distribution-based routing algorithms, which are well adapted to the QoS requirements of real-time multimedia applications. In fact, multimedia services are not sensitive to average delays, but rather to the asymptotic delay distributions. Indeed, video streaming applications drop frames when they are received beyond a delay threshold, determined by the buffer size. Although delay-distribution-based routing is an NP-hard problem, we show that it can be solved in polynomial time when the delay threshold is large, because of the asymptotic power law distribution of the link delays.

Dynamics of Feedback-Induced Packet Delay in Power-Law Networks

2010 ◽  
Author(s):  
Takahiro Hirayama ◽  
Shin'ichi Arakawa ◽  
Ken-ichi Arai ◽  
Masayuki Murata
Keyword(s):  
Power Law ◽  
Packet Delay

scholarly journals Heavy Tails in Queueing Systems: Impact of Parallelism on Tail Performance

2013 ◽  
Vol 50 (1) ◽  
pp. 127-150 ◽  
Author(s):  
Bo Jiang ◽  
Jian Tan ◽  
Wei Wei ◽  
Ness Shroff ◽  
Don Towsley
Keyword(s):  
Power Law ◽  
Heavy Tails ◽  
Queueing Systems ◽  
Splitting Scheme ◽  
Exact Asymptotics ◽  
Delay Distribution ◽  
Power Law Exponent ◽  
Distribution Tail

In this paper we quantify the efficiency of parallelism in systems that are prone to failures and exhibit power law processing delays. We characterize the performance of two prototype schemes of parallelism, redundant and split, in terms of both the power law exponent and exact asymptotics of the delay distribution tail. We also develop the optimal splitting scheme which ensures that split always outperforms redundant.

scholarly journals A Mathematical Model of Shifted Gamma Distribution for Internet Traffic Packet Delay over a Next Generation Core Internetwork

2013 ◽  
Author(s):  
Dao Ngoc Lam ◽  
Le Huu Lap ◽  
Le Nhat Thang
Keyword(s):  
Mathematical Model ◽  
Gamma Distribution ◽  
Traffic Engineering ◽  
Packet Delay ◽  
Internet Traffic ◽  
Core Network ◽  
Next Generation ◽  
Broadband Internet ◽  
Delay Distribution ◽  
Ip Packet

Broadband  Internet  traffic  is transported over the  next generation  core internetworks, which are composed of several IP/MPLS/GE network sections and transport  multi-services. In practice, IP packet delay is normally  measured  in  each  separated  network  section but  not  over  a  whole  internetwork.  It  is  proved  in  the paper  that  packet  delay  distribution  of  Internet  traffic component  in  core  network  sections  can  be approximately  expressed  as  a  shifted  gamma distribution.  Moreover  a  new  explicit  mathematical model  based  on  shifted  gamma  distribution  has  also been proposed to compose delay distribution of Internet traffic  packet  transported  over  a  core  internetwork from  component  ones  in  each  network  section.  It  is resulted from this model that Internet packet delay over an  internetwork  inherits  distribution  properties  from that  over  component  networks.  Other  properties  and parameters  relationship  of  the  model  such  as  additive property of shape and location parameters, the relation between  distribution  lower  moments  and  parameters, the dependence  of distribution  on parameters variation are also exposed in the paper. The proposed model of IP packet  delay  distributions  has  a  certain  scientific significance  and  plays  an  important  role  in  practical performance analysis, network  planning, designing and traffic engineering for improving the quality of service

Sign in / Sign up

Export Citation Format

Share Document