Aim and Objectives Introduction Simulations and Results Conclusion References Affects of Queuing Mechanisms on RTP Traffic Comparative Analysis of Jitter, End-to- End Delay and Packet Loss Gregory Epiphaniou 1 Carsten Maple 1 Paul Sant 1 Matthew Reeves 2 1 Institute for Research in Applicable Computing University of Bedfordshire 2 Modern Networks Hitchin, United Kingdom ARES Conference, 2010 Epiphaniou,Maple,Sant,Reeves Affects of Queuing Mechanisms on RTP Traffic
Aim and Objectives Introduction Simulations and Results Conclusion References Outline Aim and Objectives 1 Introduction 2 Some facts on VoIP implementation VoIP Impairments Queuing Mechanims Simulations and Results 3 First-In-First Out Random Early Detection Differentiated Services Conclusion 4 References 5 Epiphaniou,Maple,Sant,Reeves Affects of Queuing Mechanisms on RTP Traffic
Aim and Objectives Introduction Simulations and Results Conclusion References Aim and Objectives Investigate VoIP traffic behaviour under burst traffic conditions Illustrate precise effects of queuing mechanisms to VoIP and compare results Find the most appropriate mechanism to be used in the case of unelastic traffic DropTail (FIFO), RED and DiffServ, and their effects on real-time voice traffic Extract measurements on e2e delay, jitter and packet loss rates for each mechanism examined The NS-2 simulator has been used Epiphaniou,Maple,Sant,Reeves Affects of Queuing Mechanisms on RTP Traffic
Aim and Objectives Introduction Some facts on VoIP implementation Simulations and Results VoIP Impairments Conclusion Queuing Mechanims References Some facts on VoIP implementation IP telephony (IPTel) refers to the technology to transport real-time media over an IP network Converging voice and data into a best-effort service network demands preferential handling of traffic QoS assurance has increased due to the enormous growth of users accessing networks Voice coders contribute to end-to-end delay due to sample accumulation delays and look-ahead delays VoIP must reach the QoS that a Public Switched Telephone Network (PSTN) provides Best-effort service networks has resulted in many technical challenges regarding traffic engineering and shaping Epiphaniou,Maple,Sant,Reeves Affects of Queuing Mechanisms on RTP Traffic
Aim and Objectives Introduction Some facts on VoIP implementation Simulations and Results VoIP Impairments Conclusion Queuing Mechanims References VoIP Impairements – Jitter and E2E Delay Jitter can be effectively described as the unwanted variation in the inter-arrival packet time Jitter = ( T i − T i − 1 ) − ( T si − T si − 1 ) (1) Epiphaniou,Maple,Sant,Reeves Affects of Queuing Mechanisms on RTP Traffic
Aim and Objectives Introduction Some facts on VoIP implementation Simulations and Results VoIP Impairments Conclusion Queuing Mechanims References VoIP Impairements – Jitter and E2E Delay Jitter can be effectively described as the unwanted variation in the inter-arrival packet time Jitter = ( T i − T i − 1 ) − ( T si − T si − 1 ) (1) End-to-end delay is one of the most important threats to perceived QoS � L H H − q k � L D k = � � i + d k = = (2) i C i C i C i i =1 i =1 Epiphaniou,Maple,Sant,Reeves Affects of Queuing Mechanisms on RTP Traffic
Aim and Objectives Introduction Some facts on VoIP implementation Simulations and Results VoIP Impairments Conclusion Queuing Mechanims References VoIP Impairements – Packet Loss Rate and NIST Recommendations Packet loss may occur at any stage of a network transmission pkts lost = pkts sent i − pkts received j Plr ij = (3) pkts received pkts received j PLR ij ≤ 0 . 05 ∗ Pkts senti (4) The ITU-T (International Telecommunication Union) recommends that for one way transmission the actual end-to-end delay limits are: 0 to 150ms: Acceptable for most network hosted applications. 150 to 400ms: International or satellite connections. > 400ms: Unacceptable for general network purposes. Epiphaniou,Maple,Sant,Reeves Affects of Queuing Mechanisms on RTP Traffic
Aim and Objectives Introduction Some facts on VoIP implementation Simulations and Results VoIP Impairments Conclusion Queuing Mechanims References VoIP Impairements – Media Path Additional Delays Packetisation Delay T DP = P s (5) C bw Serialisation Delay D s = P s + H L (6) L s Propagation Delay Figure: VoIP media path [ ? ] D pr = L (7) u Epiphaniou,Maple,Sant,Reeves Affects of Queuing Mechanisms on RTP Traffic
Aim and Objectives Introduction Some facts on VoIP implementation Simulations and Results VoIP Impairments Conclusion Queuing Mechanims References Queuing Mechanisms - FIFO In FIFO all packets are treated Computationaly inexpensive equally and served at the same order in which they were placed in Predictable queue behaviour the queue since no packet re-ordering takes place Minimum queuing delay at the intermediate hops Does not support traffic clasification Linear relationship between aggregation of incoming traffic and mean queuing delay Benefits UDP flows Figure: FIFO Queue Epiphaniou,Maple,Sant,Reeves Affects of Queuing Mechanisms on RTP Traffic
Aim and Objectives Introduction Some facts on VoIP implementation Simulations and Results VoIP Impairments Conclusion Queuing Mechanims References Queuing Mechanisms - RED Counts the dropping probability of a packet Decisions are made based on Max/Min queue length threshold If the avg queue length is greater than max threshold the packets are marked The average queue size does not exceed the maximum threshold High throughput and low avg delay for high speed networks with TCP connections Figure: RED Queue Epiphaniou,Maple,Sant,Reeves Affects of Queuing Mechanisms on RTP Traffic
Aim and Objectives Introduction Some facts on VoIP implementation Simulations and Results VoIP Impairments Conclusion Queuing Mechanims References DiffServ Figure: DiffServ code point field Figure: IPv4 ToS byte Achieves division of packets into classes by using a 6-bit DSCP Define the policies for the level of service for each flow Performs packet marking and traffic shaping based on the policies Core routers forward packets based on their marking Epiphaniou,Maple,Sant,Reeves Affects of Queuing Mechanisms on RTP Traffic
Aim and Objectives Introduction First-In-First Out Simulations and Results Random Early Detection Conclusion Differentiated Services References Simulation Parameters Elastic and unelastic traffic examined (RTP and TCP) Pareto ON/OFF model has been used to mimick VoIP traffic G.711 (64kbps), burst time 20ms, idle time 10ms and rate 87.4kbps Default payload size of 160 bytes Total simulation time 62sec No cRTP/Multiplexing mechanisms Table: Link characteristics nodes Propagation Delay [ms] Link capacity [Mb] n 0 - n 2 10 1 n 1 - n 2 10 1.5 n 2 - n 3 20 2 n 3 - n 4 10 1 n 3 - n 5 10 1 Epiphaniou,Maple,Sant,Reeves Affects of Queuing Mechanisms on RTP Traffic
Aim and Objectives Introduction First-In-First Out Simulations and Results Random Early Detection Conclusion Differentiated Services References FIFO Results and Discussion 0.09 Delay variation for RTP traffic with FIFO [sec] 0.08 0.07 0.06 0.05 0.04 0.03 0.02 Figure: Simulation Model FIFO 0.01 0 0 10 20 30 40 50 60 Simulation time [sec] Traffic sent Dropped Figure: Jitter for RTP traffic with FIFO TCP 10484 5 RTP 101771 1262 Epiphaniou,Maple,Sant,Reeves Affects of Queuing Mechanisms on RTP Traffic
Aim and Objectives Introduction First-In-First Out Simulations and Results Random Early Detection Conclusion Differentiated Services References FIFO e2e delay 0.2 0.16 RTP OWD for RTP traffic with 50pkts buffer size [sec] RTP OWD for RTP traffic with 25pkts buffer size [sec] 0.18 0.14 0.16 0.12 0.14 0.1 0.12 0.08 0.1 0.08 0.06 0.06 0.04 0.04 0.02 0.02 0 0 0 10 20 30 40 50 60 0 10 20 30 40 50 60 Simulation time [sec] Simulation time [sec] Figure: Simulation Model FIFO: RTP One Figure: Simulation Model FIFO: RTP One Way Delay with buffer size of 50pkts and Way Delay with buffer size of 25pkts and DropTail DropTail Epiphaniou,Maple,Sant,Reeves Affects of Queuing Mechanisms on RTP Traffic
Aim and Objectives Introduction First-In-First Out Simulations and Results Random Early Detection Conclusion Differentiated Services References RED Results and Discussion 0.08 Delay Variation for RTP traffic with RED [sec] 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0 10 20 30 40 50 60 Simulation time [sec] Figure: Simulation Model RED Figure: Simulation Model RED: Delay Variation for RTP traffic against simulation time Epiphaniou,Maple,Sant,Reeves Affects of Queuing Mechanisms on RTP Traffic
Aim and Objectives Introduction First-In-First Out Simulations and Results Random Early Detection Conclusion Differentiated Services References RED e2e delay and packet loss 0.25 600 0.2 Total RTP packets dropped with RED 500 RTP OWD with RED [sec] 0.15 400 300 0.1 200 0.05 100 0 0 10 20 30 40 50 60 Simulation time [sec] 0 0 10 20 30 40 50 60 Simulation time [sec] Figure: Simulation Model RED: RTP One Figure: Simulation Model RED: RTP Way Delay with buffer size of 50pkts packet drop rate against simulation time against simulation time Epiphaniou,Maple,Sant,Reeves Affects of Queuing Mechanisms on RTP Traffic
Recommend
More recommend
