Packet Transfer Delay Sample Clauses

Packet Transfer Delay. ‌ The IP Packet Transfer Delay (IPTD) of the IP service is defined as time difference between the time the first bit of a packet passing the ingress measurement point, and the time the last bit passes the egress measurement point at the other end. It is determined by the IP packet size, queueing delays inside the networks, the raw medium transmission speeds at the sending and receiving end, and of course the (geographical) distance. Note that although the shortest path between any two points is a straight line, network cables tend to be put in different places from that line, and are hence longer. Also the route that packets take through networks are determined by factors like routing tables, the topology of the various subnetworks, peering agreements between various networks and so on. Because routing in IP networks is dynamic down to the individual packet level latency can vary con- siderably from one packet to the next. If fragmentation of packets occurs, the individual packet frag- ments can travel along different routes before getting reassembled, and cause considerable differences in delay. Also, when sending a larger number of packets, the nature of the routing and forwarding functions in IP networks is such that these packets can travel different routes, and thus experience different packet delays. Packet delay is of interest for anything interactive, like video conferencing, audio/telephony over IP, real time transaction systems (banking), web-browsing, etc. Applications can be sensitive to delay even when one might not expect this. Transport layer protocols like TCP for example have difficulty achieving high throughput values if the transfer delay is too high. The sensitivity of an application to packet delay can be asymmetrical. For example, an application that uses the TCP transport protocol to receive bulk data is sensitive to a high delay in the direction of the actual data flow, but less sensitive in the direction of the flow of TCP acknowledgements. For this reason it is useful to not only specify in the SLA values for the round trip delay, but also for the one-way delay. For a Voice over IP connection to be usable, the one-way delay should be below 150 ms [G.114]. [ITU-R M.1079] probably has some info on acceptable delay and delay variation for voice traffic, but I haven’t been able to get hold of this document yet.