In computer networks, because different communication subnets and network architectures use different transit control modes, the transit delay in communication subnets is determined by the network state.The delay time caused by the response of the electrical signal is fixed. The smaller the response time, the smaller the delay, the larger the bandwidth, and the higher the transmission rate. Therefore, the larger the bandwidth of the channel, the smaller the delay.The delay time is to get adata packet(packet). The delay time is generally the sum of response delay and transmission delay
In the network, latency, synonymous with delay, is to get a packet from a specific point（packet）Time spent.In some cases, the delay time is the period from sending the packet to returning to the sending end.The reasons for network transmission delay include: the time it takes for transmission and packets to travel at the speed of light between two places.Sending: the media itself (whether optical fiber（fiber optic）, wireless, or other media).The delay caused by large packets is often longer than that caused by small packets.Routers and other processes, eachgateway, Node（node）Taking time to detect may change the packet header (for example, change the lifetime（time-to-live）Hop count within the domain).Other computers and storage delays. At each transmission endpoint in the network, packets may be stored and hard disks will be accessed. The delay occurs in intermediate devices such as switches（switch）And the bridge.
For users, the delay time is also calledNetwork delay, refers to the period from the time when the user sends a request to the time when the remote system responds to the request and returns it to the user.For the Internet based on TCP/IP protocol, each request must be processed as follows: routing processing, ADU (user data unit) transmission on the network, and the server processing the request. These processes will cause delays.
Routing delay
Announce
edit
First, look at the routing delay, which includes domain name request delay, TCP connection establishment delay, TCP connection release delay, and IP routing delay on each gateway.If the user uses the domain name of the opposite host instead of the IP address in the application, the delay caused by the domain name resolution process of the opposite IP address should be resolved before the application communicates, which is called the domain name request delay.The application program hands the domain name to the local parser software, which first looks up the corresponding domain name address linkage in the local cache;If it cannot be found, the local parser constructs a query message and sends it to the initial domain name server (local server). The domain name server answers a response message according to the resolution.Domain Name ServerThe resolution adopts a two-step method: when the initial server cannot find the domain name, it sends the inquiry message to the root server for top-down search (the domain name server is organized into a tree hierarchy).The response latency of the local cache to the domain name request is determined by the CPU, memory, and external memory speed, which is relatively small.The response of the domain name server is related to the network load, server speed and load.When the root server needs to start searching for domain names, the transmission delay becomes the delay on the LAN, which is relatively large.TCP connection establishment delay refers to the time from the time when the transmission service user requests to establish a connection to the time when the connection confirmation is received. It includes the processing work delay of the remote transport layer.Connection release delay refers to the delay between the time when the transport layer user at one end initiates a connection release request and the time when the release at the other end actually occurs.The connection establishment and release delay is related to the network load and server load.In TCP/IP protocol, each IP packet is independently pathfinding.IP routing delay refers to the routing delay of IP packets on each gateway from the source end to the destination end, including the gateway routing table processing delay and address resolution delay.Gateway address resolution is completed by ARP (Address Resolution Protocol) provided by TCP/IP.Because gateway routing is completed in the local machine, and address resolution is also performed in the local network, the IP routing delay is relatively small.
Transmission delay
Announce
edit
The user data unit (ADU) in the application layerApplication layer protocolFormed after calling the service provided by the underlying protocolProtocol Data Unit(PDU), which is transmitted on the network as an independent data unit.The transmission delay of PDU on the network consists of the following parts: packaging delay, transmission delay, propagation delay, queuing delay and processing delay.Packaging delay is unique to real-time streaming applications.Real time streaming applications refer to applications that transmit time-based information (such as video, audio, animation, etc.) in real time.Real time information flows are timed and generated regularly in chronological order. It is necessary to wait for a certain number of information flows to meet the payload of the Protocol Data Unit (PDU) before packaging them into PDUs. This waiting time is called packaging delay.Packaging delay Sp/K, where Sp is the PDU payload length and K is the encoding speed of the application.Therefore, for low-speed encoders, the delay is relatively large.Transmission delay refers to the time delay when all PDU data is transmitted to the line, or the time interval between the first bit of the PDU being transmitted from the endpoint to the line and the last bit leaving the endpoint.The transmission delay is related to the size of the PDU and the transmission rate on the local line. For low-speed links, the transmission delay is considerable.
Propagation delay is the time taken by PDU to propagate in the medium, which depends on the speed of signal passing through the medium.In other words, the delay is determined by the physical characteristics of the link and is independent of the traffic volume of the link.The propagation delay increases linearly with distance. The propagation delay in LAN and MAN is only 50~200us, which is not an important component of delay.The propagation delay in WAN becomes more important because the propagation delay can reach tens of milliseconds due to long-distance transmission.Queue delay isPacket switched networkThe main delay in, which refers to the set of buffer delays caused by each exchange of PDU on the transmission path.If packet switching is temporarily overloaded, there may be many packets queued on the destination output port of each PDU.Each packet in front of the PDU in the queue will generate an additional delay equal to the transmission delay.In the exchange of first in first out (FIFO) queuing mechanism, the queuing delay of newly arrived packets is equal to the sum of all packet transmission delays queued on the output port.Therefore, the queuing delay is related to both the number of packets in front of the queue and the transmission speed of the output port.As the main delay, the queuing delay is affected by the current network load, and it is also the main factor of delay variation in packet switching networks.Because every router on the Internet backbone has a large number of packets queued, queuing delay has also become the main delay on the Internet.If you pass through 10 routers, each router has 10 IP packets queued on average, and the queuing delay on this path can reach hundreds of milliseconds.The processing delay is the sum of the time spent on each switch on the path and on the source and destination ends to process the packets. It does not include the queuing delay.In the process of packet switching, the processing delay is negligible compared with the queuing delay, but it is still considerable at the endpoint, especially for applications such as video decompression[1]。
measuring method
Announce
edit
Network performance measurement methods can be divided into active measurement and passive measurement according to the measurement mode.Active measurement is to inject the probe packet into the network under test. The probe packet records the transmission in the network. The terminal obtains network performance data by parsing the probe packet.The passive measurement method uses monitoring equipment or tools to analyze network characteristics by capturing the data packets transmitted in the tested network.Both methods have advantages and disadvantages.The active measurement method is flexible and controllable, and can set parameters according to the intention of the measurer, such as the type and size of the package.However, the measured data flow injected into the network will increase the network load, thus affecting the network performance. The measured data cannot truly reflect the actual network performance.The advantage of passive measurement is that it will not increase the traffic on the network, and the measured data can truly reflect the network performance.The disadvantage is poor controllability, and the confidentiality and security of the collected data will be affected.Therefore, sometimes these two methods are often combined for network measurement.According to the definition of network delay measurement, the delay of any layer can be attributed to one-way delay or round trip delay, but the protocol is different.Therefore, the following test methods can be applied to delay measurement above the network layer.
One way delay measurement method
First, select the protocols used for measurement, such as TCP, UDP, ICMP (Internet Control Message Protocol), OWDP (One Way Delay Protocol), IPMP (Internet Protocol Measurement Protocol, IP layer measurement protocol), etc., and determine the more detailed problems of the selected protocols, such as TCPUDP port number, timeout waiting time, measurement packet size, etc.After these parameters are determined, proceed as follows:
(1) Negotiate source and destination synchronization: that is, the clocks at the source and destination must be synchronized accurately.
(2) On the source side, determine the IP address of the destination side, and fill in the test package according to the structure of the selected measurement protocol.The vacant part should be filled with random bits to avoid the measured delay being lower than the delay after the compression technology is used in the path.Set the time stamp in the test package, and then send the filled test package to the destination host.
(3) At the destination end, prepare to receive the packet (use the passive measurement method, that is, use the monitoring tool to capture the packet).If the packet arrives within a reasonable time, the estimated one-way delay can be calculated by subtracting the sending time from the receiving time.This value is valid only when the source and destination clocks are synchronized.If the packet cannot arrive in a reasonable time, the one-way delay value should be set to undefined.There are two methods to solve the clock synchronization problem: software synchronization and hardware synchronization.GPS (Global Position System) based synchronization belongs to hardware synchronization;Synchronization based on Network Time Protocol and synchronization based on algorithm estimation belong to software synchronization.Multiple use of time synchronization on the InternetNTP serverThe NTP server usesGPS receiverTo provide a reference clock, the host obtains the time of the NTP server through the network time protocol NTP.The algorithms to eliminate clock asynchrony include linear regression algorithm, median line adaptation algorithm, piecewise minimum algorithm, Convex Hull Approach algorithm and linear programming algorithm.
Round trip delay measurement method
The preparation work is the same as the one-way delay measurement method. The measurement is carried out as follows:
(1) On the source side, determine the IP address of the destination side, and fill in the test package according to the structure of the selected measurement protocol.In order to avoid the measured delay being lower than the expected result due to the compression technology on the path, the vacant part should be filled with random bits;In addition, the ID number of the test package must be set so that the source host can judge the response of the test package sent by itself after receiving the response package.
(2) On the source side, set the sending timestamp in the test package, and then send the filled test package to the destination host.The timestamp can be placed inside or outside the packet, as long as it contains a suitable identifier so that the received timestamp can be compared with the sent timestamp.
(3) At the destination end, prepare to receive the test package.If the test packet reaches the destination host, the response packet should be generated quickly and sent to the source.
(4) At the source end, prepare to receive the corresponding response packet.Determine whether it is the response of your own test package according to the ID number.If the response packet arrives within the timeout waiting time, the round-trip delay value can be calculated by subtracting the sending time from the receiving time.If the response packet cannot arrive within the timeout waiting time, the round-trip delay value is set to undefined[2]。
