Introspecting Effect of Packet Size on End-to-End Network Delay
A message in computer networks is often divided into frames/packets for various reasons. For a store-and-forward network, a suitable packet-size can drastically reduce delay. In this paper, I have investigated the impact of packet size on delivery time. I
- PDF / 377,994 Bytes
- 9 Pages / 439.37 x 666.142 pts Page_size
- 2 Downloads / 184 Views
Abstract A message in computer networks is often divided into frames/packets for various reasons. For a store-and-forward network, a suitable packet-size can drastically reduce delay. In this paper, I have investigated the impact of packet size on delivery time. I have shown that delay is a non-linear function of (i) number of hops the packet traverses, (ii) message size and (ii) the number the message is divided into packets. Since, I can’t customize the former two; the last one can suitably be chosen to minimize delay. I found an optimal number of packets that minimizes the delay. Analytical and simulation results show the correctness of the proposed scheme. Keywords Network
⋅
End-to-end delay
⋅
Packet size
⋅
Hop
1 Introduction In computer networks, a message is often divided into smaller chunks by various protocol layers. This is done for various reasons such as regulating data flow, error detection, correction and control, link multiplexing, switching etc. However, for a store-and-forward network, the size of packet also has direct impact on end-to-end packet delay hence network performance. By choosing a suitable packet size, it is possible to improve the network performance especially delay. In this paper, I have investigated the impact of packet size on delivery time with rigorous theoretic derivation. I have shown that delay is a non-linear function of various parameters such as (i) number of hops the packet traverses, (ii) size of the message being broken into and (ii) the number the message is divided into packets. Since, former two are out of our reach; the last one can suitably be chosen to minimize delay. I found an optimal number of packets that minimizes the delay. Analytical and simulation results show the correctness of the proposed scheme. U.K. Roy (✉) Department of Information Technology, Jadavpur University, Kolkata, India e-mail: [email protected] © Springer Science+Business Media Singapore 2017 J.K. Mandal et al. (eds.), Proceedings of the First International Conference on Intelligent Computing and Communication, Advances in Intelligent Systems and Computing 458, DOI 10.1007/978-981-10-2035-3_14
127
128
U.K. Roy
2 Related Work There is relatively scant literature on the impact of number of packets on packet delay. In [1], authors presented design guidelines for WNCS over MANET using the NS2 simulator. In [2, 3], authors estimated the medium access control layer packet delay distribution for IEEE 802.11 considering the differences between busy probability and collision probability. In [4], Anyaegbu and et al. characterized the delay profile of an Ethernet cross-traffic network statically loaded with one of the ITU-T network models and a larger Ethernet inline traffic loaded with uniformly sized packets, showing how the average time interval between consecutive minimum-delayed packets increases with increased network load. In [5] a TCP control mechanism based on the character of delay distribution and wireless packet loss was proposed. In [6], authors developed an analytic model for c
Data Loading...
