CN108011763B - Communication data network investment construction evaluation method - Google Patents

Abstract

The invention discloses a communication data network investment construction evaluation method, which comprises the following steps: s1: acquiring bottleneck bandwidth, available bandwidth and inherent loss of a communication data network link; s2: analyzing the bandwidth demand and the service flow growth rate of each service type in the region through a flow statistical technique, and solving the bandwidth demand sum of each service type in the region; s3: calculating and obtaining the actual bandwidth demand of the current communication data network link according to a bandwidth demand prediction model; s4: acquiring a traffic convergence ratio of a communication data network link, predicting the traffic demand of a company cross section service in a region according to a national network communication part, and calculating to obtain the actual bandwidth demand of the communication data network link in the coming years by a backtracking method; s5: determining the investment construction direction in the next years according to the service performance of the communication data network; and adjusting the service flow distribution proportion and the service priority of each service type in each year in the future according to the service flow increase rate of each service type.

Description

Evaluation method for investment and construction of communication data network

technical field

The invention relates to a communication data network investment construction evaluation method.

Background technique

Today, communication networks play a pivotal role in all walks of life, and cloud computing and big data have also sounded the horn of the times. Under such an era background, the informatization construction of the power grid is also constantly evolving in this direction, and plays a key role in the guarantee and promotion of its business. At the same time, most of the backbone communication network equipment in the power grid uses traditional communication equipment, which is still a long way from successfully evolving to a flexible, elastic, and easily scalable SDN network architecture. Identify performance bottlenecks and predict service bandwidth requirements. , and used to guide its future network planning and construction.

SUMMARY OF THE INVENTION

The purpose of the present invention is to provide a communication data network investment and construction evaluation method, which can calculate performance bottlenecks, predict service bandwidth requirements, and be used to guide its future network planning and construction.

In order to solve the above-mentioned technical problems, the present invention provides a communication data network investment and construction evaluation method, comprising the following steps:

S1: Obtain the bottleneck bandwidth, available bandwidth and inherent loss of the communication data network link;

S2: Analyze the bandwidth requirements of each service type in the area and the growth rate of service traffic through the traffic statistics technology, and obtain the sum of the bandwidth requirements of each service type in the area;

S3: Calculate and obtain the actual bandwidth demand of the current communication data network link according to the bandwidth demand prediction model;

S4: Obtain the traffic convergence ratio of the communication data network link, and obtain the actual bandwidth demand of the communication data network link in the next few years by the backtracking method according to the forecast of the business flow demand of the company's cross-section in the region by the Ministry of Communications of the State Grid;

S5: Determine the investment and construction direction in the next few years according to the bottleneck bandwidth in the communication data network link, the available bandwidth, the actual bandwidth demand of the current communication data network link, and the actual bandwidth demand of the communication data network link in the next few years; According to the business traffic growth rate of each business type, adjust the business traffic allocation ratio of each business type and the business priority of each business type in the next few years.

Further, the method for obtaining the bottleneck bandwidth of the communication data network link in step S1 adopts the variable-length single-packet measurement method.

Further, the method for calculating the bottleneck bandwidth of the communication data network link using the variable-length single-packet measurement method specifically includes:

According to the principle of variable-length single-packet measurement technology, calculate the delay of data packet k on link i,

为：The time it takes for a packet to arrive at link l from the source

for:

为测量主机接入通信网络所带来的固有时延；where ^sk is the size of the kth data packet, bi is the capacity of link _{i, d i is} the fixed delay of link i,

To measure the inherent delay caused by the host accessing the communication network;

为：The time for packet k to leave link l is

for:

Therefore, the delay t experienced by packet k on link l is:

The above t is equivalent to the end-to-end path delay T, then we have

Among them, S is the size of the data packet, B is the bottleneck bandwidth of the end-to-end path, and D is the transmission delay of the end-to-end path;

That is, the end-to-end path bottleneck bandwidth B is obtained as:

Further, the available bandwidth described in step S1 is obtained by a self-load cycle flow measurement method.

Further, the available bandwidth described in step S1 is measured and obtained by using the Pathload tool.

Further, the step S2 specifically includes:

S21: Using a traffic statistics technology, collect service traffic of several sites in the current area, and divide the several sites into different service types according to different services;

S22: Count the service traffic of each service type in the area, analyze the bandwidth requirements and service traffic growth rate of each service type in the area, and obtain the sum of the bandwidth requirements of each service type in the area.

Further, the step S3 specifically includes:

Prediction model based on bandwidth demand:

And because there is a bandwidth loss C in the communication data network link, there are

Wherein, the light-load service flow in the above formula is the bandwidth requirement of the current actual communication data network link of the communication data network.

The beneficial effects of the invention are as follows: the present application conducts a comprehensive system analysis on the performance of the communication data network, calculates the network performance bottleneck, measures the end-to-end path bottleneck bandwidth and available bandwidth, determines the network convergence ratio, and based on the current The network load forecasts the bandwidth demand, and uses the retrospective method to forecast the bandwidth demand in the next 5 years, which provides a reliable data foundation for the future communication data network planning and investment and construction direction. rate, and adjust the business traffic allocation ratio and business priority of each business type in the next few years.

Description of drawings

The accompanying drawings described herein are used to provide a further understanding of the application and constitute a part of this application, and the same reference numerals are used in these drawings to refer to the same or similar parts. For the purpose of interpreting this application, it does not constitute an improper limitation to this application. In the attached image:

Figure 1 is the straight line fitting of the bottleneck bandwidth of the end-to-end path from the Qinan New Office to the Tongtong Institute;

Detailed ways

As shown in Figure 1, the evaluation method for investment and construction of communication data network includes the following steps:

S1: Obtain the bottleneck bandwidth, available bandwidth and inherent loss of the communication data network link;

S2: Analyze the bandwidth requirements of each service type in the area and the growth rate of service traffic through the traffic statistics technology, and obtain the sum of the bandwidth requirements of each service type in the area;

S3: Calculate and obtain the actual bandwidth demand of the current communication data network link according to the bandwidth demand prediction model;

S4: Obtain the traffic convergence ratio of the communication data network link, and obtain the actual bandwidth demand of the communication data network link in the next few years by the backtracking method according to the forecast of the business flow demand of the company's cross-section in the region by the Ministry of Communications of the State Grid;

S5: Determine the investment and construction direction in the next few years according to the bottleneck bandwidth in the communication data network link, the available bandwidth, the actual bandwidth demand of the current communication data network link, and the actual bandwidth demand of the communication data network link in the next few years; According to the business traffic growth rate of each business type, adjust the business traffic allocation ratio of each business type and the business priority of each business type in the next few years.

Each component is described in detail below:

According to an embodiment of the present application, the method for obtaining the bottleneck bandwidth of the communication data network link in the above step S1 adopts the variable-length single-packet measurement method.

According to an embodiment of the present application, the method for calculating the bottleneck bandwidth of a communication data network link using the variable-length single-packet measurement method specifically includes:

According to the principle of variable-length single-packet measurement technology, calculate the delay of data packet k on link i,

为：The time it takes for a packet to arrive at link l from the source

for:

为测量主机接入通信网络所带来的固有时延；where ^sk is the size of the kth data packet, bi is the capacity of link _{i, d i is} the fixed delay of link i,

To measure the inherent delay caused by the host accessing the communication network;

为：The time for packet k to leave link l is

for:

Therefore, the delay t experienced by packet k on link l is:

The above t is equivalent to the end-to-end path delay T, then we have

Among them, S is the size of the data packet, B is the bottleneck bandwidth of the end-to-end path, and D is the transmission delay of the end-to-end path;

That is, the end-to-end path bottleneck bandwidth B is obtained as:

When it is actually necessary to measure the bottleneck bandwidth of the end-to-end path, the host sends a series of ICMP ECHO request packets of different sizes to the loopback port of the start node and the loopback port of the leaf node multiple times, and the start node and the leaf node will return the same The size of the ICMP ECHO response packet, respectively, record the minimum RTT value in the round-trip delay sample, the RTT value can be considered to have eliminated the queuing delay, that is, D in the formula has achieved the minimum value. According to the current general Internet standards, when the data packet size exceeds 1500byte, it will be fragmented and transmitted. Therefore, the data packet size in this measurement is controlled between 108byte and 1458byte, and the data packet is incremented in 50byte steps. Send the ICMP ECHO request packet 20 times, record the minimum RTT value respectively, and repeat the above process. This measurement is from 9:00 am to 5:00 pm, and the process is repeated 11 times in total. Each cycle can get a set of bandwidth values. This method eliminates statistical errors. Since the unit of data packet is byte and the unit of RTT value is ms, the minimum RTT value time measured includes the round-trip time of the end-to-end path. Therefore, taking the above factors into consideration, the above-mentioned end-to-end path delay T becomes:

In the above formula, the unit of time T is ms, the unit of data packet S is byte, and the unit of bottleneck bandwidth B is Mbps.

In theory, the relationship between time T and packet size S should be linear, but in practice, due to the interference of many uncertain factors, such as sudden fluctuation of background traffic, inconsistent end-to-end round-trip paths, and the minimum RTT value measured based on statistical methods The delay of packet queuing is not completely eliminated, which may cause a difference between the slope value of the fitted line and the actual bandwidth. Figure 1 shows the fitted line for the bottleneck bandwidth of the end-to-end path from the Qinan New Office to the Tongtong Institute. It can be seen that the Qinan New Office to The slope of the straight line fitted by getting through is 41.4156, that is, the bottleneck bandwidth value is 41.4156Mbps, which is too different from the real value, and the fitting results of other nodes are also the same. Therefore, the method of using straight line fitting to find the bandwidth is not very effective in practice.

To obtain the bottleneck bandwidth using straight line fitting, the measured data needs to have a good linear distribution relationship, otherwise the slope of the fitted straight line cannot represent the real bandwidth value. The analysis shows that the data points are distributed as shown in Figure 1. This is because the interference of various factors (which can be understood as noise) makes the measured corresponding RTT values different when the packet size changes relatively small. Exactly, in such cases even a small amount of noise acting on the RTT value can have a serious effect on the distribution of data points. The noise carried by the RTT value is inherently determined by the network system and cannot be avoided. Therefore, to weaken the influence of noise, it can be achieved by expanding the size of the data packet variation. In extreme cases, the smallest data packet and the largest data packet can be selected. Two points are used for linear fitting, and the slope is calculated, so that the calculated value will be as close to the real value of the bottleneck bandwidth as possible.

Based _on the above ideas, suppose the size of the data packet to be sent is expressed as S ₁ , S ₂ . Corresponding to the minimum RTT value with round-trip delay, it is recorded as RTT _1leaf , RTT _2leaf ..., RTT _28leaf and RTT _1root , RTT _2root ..., RTT _28root , so it is easy to get the following formula:

Combining the above two formulas can be derived:

Among them, leaf represents the end-to-end path leaf node, root represents the end-to-end path root node,

The end-to-end bottleneck bandwidth value calculated by the above method is shown in Table 1. The measured bandwidth value is the end-to-end bottleneck bandwidth from the central computer room of Qinan New Office to each leaf site. Most of the measured bandwidth results are distributed between 400Mbps and 700Mbps. Therefore, this end-to-end bottleneck bandwidth measurement value takes a representative value of 650Mbps. In the actual communication network, the bandwidth of the direct interconnection interface of each router is 1Gbps, and considering the bandwidth loss caused by various factors in the end-to-end path, the end-to-end bottleneck bandwidth measurement result of 650Mbps is in good agreement with the actual situation. The measurement results in Table 1 are all measured during the busy hours of business during the day, and the measurement results are still consistent with the actual situation. Increase the number of times of sending ICMP ECHO request packets (in this paper, due to the time relationship of data collection, the number of times of sending ICMP ECHO request packets is 20 times).

In addition, through this measurement, the inherent loss of bandwidth in the end-to-end path is found. The loss is a fixed constant, and its value is about 350Mbps, that is, the bandwidth that can be effectively used for the network interconnection interface is 1Gbps. The value is about 650Mbps, and the loss is only related to inherent factors such as optical transmission medium loss, equipment installation link docking loss, number of nodes passed, and transmission distance. Therefore, this inherent loss can be taken into account when making bandwidth demand predictions.

Table 1 End-to-end bottleneck bandwidth measurement results

According to an embodiment of the present application, the available bandwidth in step S1 is obtained by a self-load periodic flow measurement method, which specifically adopts the Pathload tool to measure and obtain the available bandwidth of the communication data network.

The Pathload tool is a technology developed based on Self-Load Periodic Flow Measurement (SLPS) to measure the available bandwidth of the end-to-end path. The core idea is that when the probe packet flow is sent from the source to the sink, and the rate of the probe packet flow is greater than the available bandwidth of the measurement path, the one-way delay of the probe packet flow shows an upward trend, otherwise the delay does not change significantly. At the source end, a binary search algorithm (Pathload tool) or linear algorithm (IGI tool) is used to adjust the rate of the detection packet flow. By monitoring the one-way delay at the sink end, the inflection point of the curve can be considered as the source end detection packet The flow rate is equal to the available bandwidth of the measurement path. Therefore, the Pathload application can be used to determine the theoretically available bandwidth of a network path between two points, even when each network device is under load. Pathload consists of two main programs, pathload_snd and pathload_rcv. The pathload_snd program runs on the server side to receive connection requests from remote hosts (clients). The pathload_rcv program runs on the client side. The pathload_rcv program uses the remote host name or IP address on the command line. Specify the Pathload server and initiate a data connection.

For example, use the Pathload tool to measure the available bandwidth of the Longsheng Institute, the Triangle Institute, the Triangle Station, the Wansheng Customer Center, the Maintenance Base and the Qinan New Bureau Data Center.

Execute ./pathload_snd-i on the server of Qinan New Office, and execute the ./pathload_rcv-s 172.28.253.130 command on the client of the leaf node that needs to be measured, and the result information of the available bandwidth test will be displayed. Take the test result of the triangle station as an example, As follows:

Receiver helong-virtual-machine starts measurements at sender172.28.253.130on Thu Jun 15 10:08:05 2017

Interrupt coalescion detected

Receiving Fleet 0,Rate 83.59Mbps

Receiving Fleet 1,Rate 155.53Mbps

Receiving Fleet 2,Rate 103.97Mbps

Receiving Fleet 3,Rate 111.83Mbps

Receiving Fleet 4,Rate 115.33Mbps

Receiving Fleet 5,Rate 105.15Mbps

Receiving Fleet 6,Rate 105.15Mbps

Receiving Fleet 7,Rate 95.64Mbps

Receiving Fleet 8,Rate 105.15Mbps

Receiving Fleet 9,Rate 102.72Mbps

Receiving Fleet 10,Rate 103.23Mbps

Receiving Fleet 11,Rate 105.15Mbps

Receiving Fleet 12,Rate 105.15Mbps

Receiving Fleet 13,Rate 105.15Mbps

Receiving Fleet 14,Rate 96.74Mbps

Receiving Fleet 15,Rate 105.15Mbps

Receiving Fleet 16,Rate 103.27Mbps

Receiving Fleet 17,Rate 103.27Mbps

Receiving Fleet 18,Rate 103.27Mbps

Receiving Fleet 19,Rate 105.15Mbps

Receiving Fleet 20,Rate 105.15Mbps

Receiving Fleet 21,Rate 105.15Mbps

Receiving Fleet 22,Rate 96.74Mbps

Receiving Fleet 23,Rate 105.15Mbps

Receiving Fleet 24,Rate 105.60Mbps

Receiving Fleet 25,Rate 105.80Mbps

*****RESULT*****

Available bandwidth range: 101.78-109.81(Mbps)

Measurements finished at Thu Jun 15 10:08:31 2017

Measurement latency is 26.02sec

Through this method, 5 stations are randomly selected, including Longsheng Institute, Sanjiao Institute, Sanjiao Station, Wansheng Customer Center, Maintenance Base and Qinan New Bureau Data Center. The measurement results of available bandwidth are shown in Table 2.

Table 2 Measurement results of available bandwidth from five representative sites to Qinan New Bureau

Longsheng Institute Triangle Triangle Station Wansheng Customer Center Maintenance base available bandwidth 87.44-99.74(Mbps) 84.04-96.80(Mbps) 91.68-109.81(Mbps) 75.32-168.11(Mbps) 94.30-123.90(Mbps)

Since more than 90% of the upstream traffic of each site in the overall architecture of Qinan Power Grid goes through Qijiang Station to Qinan New Bureau, the upstream traffic of the five sites in Table 2 also passes through Qijiang Station, and their remaining available bandwidth is 100Mbps. Between -150Mbps, take the typical value of 150Mbps. The bottleneck bandwidth has been measured as 650Mbps, so the used bandwidth value is close to 500Mbps. In addition, there is a certain error whether it is the measurement result using the statistical method or the currently measured bottleneck bandwidth and available bandwidth. Therefore, a typical compromise value can be taken. The used bandwidth is 450Mbps, which is close to 80% of the bottleneck bandwidth value. Therefore, the entire network is running under heavy load. Once there is a burst of traffic, the network will be congested and stuck. .

According to an embodiment of the present application, the step S2 specifically includes:

S21: Using a traffic statistics technology, collect service traffic of several sites in the current area, and divide the several sites into different service types according to different services;

S22: Count the service traffic of each service type in the area, analyze the bandwidth requirements and service traffic growth rate of each service type in the area, and obtain the sum of the bandwidth requirements of each service type in the area.

For example, in this embodiment, the applicant conducts statistics on the service traffic of 11 randomly selected sites in the area, and the results are shown in Table 3. All the results in the table are peak data at the measurement time. It can be seen from the table that the business office occupies a very high bandwidth, and the substation generates very little business traffic. The top three services in terms of bandwidth occupation are management information services, information intranet (office), and unified video monitoring. The total bandwidth of the seven offices is 272.52Mbps, and the bandwidth occupied by the four substations is 1.5608Mbps. The entire Qijiang power grid has 19 business offices and 46 substations. Assuming that the service bandwidth occupation distribution of the remaining unmeasured sites is consistent with Table 3, the total bandwidth requirements are:

Table 3 Statistics on business traffic of 11 random sites

According to an embodiment of the present application, step S3 specifically includes

Prediction model based on bandwidth demand:

In addition, due to the inherent loss of bandwidth C in the communication data network link, and the value of C is a constant C≈350Mbps, there are

Wherein, the bandwidth value calculated by the above formula is the actual bandwidth requirement of the current communication data network link constructed by the communication data network.

For example, considering that the peak aggregated traffic in the current estimated network can reach 450Mbps, the actual link bandwidth requirements are calculated based on the current service bandwidth occupancy as follows:

Considering that the actual real link bandwidth is 1Gbps, the link needs to be appropriately expanded.

According to an embodiment of the present application, step S4 is to obtain the traffic convergence ratio of the communication data network link. According to the forecast of the business flow demand of the company's cross-section in the region by the State Grid Communications Department, the backtracking method is used to calculate and obtain the communication data network link in the next few years. actual bandwidth requirements.

For example, according to the forecast of the business traffic demand of the county company's cross-section by the State Grid Headquarters, it is expected that the traffic demand will reach 1603.52Mbps in the next five years (the concurrency ratio is calculated based on 100%). According to the actual calculation results, the actual concurrency ratio (flow convergence ratio) of Qijiang power grid is 1:2. Therefore, through retrospective calculation, the actual bandwidth demand in the next five years is:

Re-examination of the inherent bandwidth loss in the communication data network link, the actual link bandwidth demand in the next five years can be obtained as follows:

In general, to ensure that the current network operates under light load, the current network link needs to be expanded to 1475Mbps or more. To ensure light-load operation of the network in the next five years, the network link capacity needs to be expanded to 2354.4Mbps or more.

According to an embodiment of the present application, step S5 specifically includes

Determine the investment and construction direction in the next few years according to the bottleneck bandwidth, available bandwidth in the communication data network link, the actual bandwidth demand of the current communication data network link and the actual bandwidth demand of the communication data network link in the next few years; The business traffic growth rate of the business type, adjust the business traffic allocation ratio of each business type and the business priority of each business type in the next few years. When allocating the business traffic of each business type in the future, the required proportion of the business traffic of each business type in the next few years can be predicted according to the business traffic growth rate of each business type, and then the proportion of the business traffic of each business type in the next few years can be predicted according to the forecast. Rather than re-plan the bandwidth occupancy ratio and priority of each service type, allocate more service traffic to the service type with more service traffic, and set its priority to give priority to the service traffic demand of this service type, instead of using the current average allocation The method can improve the utilization rate of the communication data network.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be Modifications or equivalent substitutions without departing from the spirit and scope of the technical solutions of the present invention should be included in the scope of the claims of the present invention.

Claims (7)

1. a communication data network investment construction evaluation method, is characterized in that, comprises the following steps: S1: Obtain the bottleneck bandwidth, available bandwidth and inherent loss of the communication data network link; S2: Analyze the bandwidth requirements of each service type in the area and the growth rate of service traffic through the traffic statistics technology, and obtain the sum of the bandwidth requirements of each service type in the area; S3: Calculate and obtain the actual bandwidth demand of the current communication data network link according to the bandwidth demand prediction model; S4: Obtain the traffic convergence ratio of the communication data network link, and obtain the actual bandwidth demand of the communication data network link in the next few years by the backtracking method according to the forecast of the business flow demand of the company's cross-section in the region by the Ministry of Communications of the State Grid; S5: Determine the investment and construction direction in the next few years according to the bottleneck bandwidth in the communication data network link, the available bandwidth, the actual bandwidth demand of the current communication data network link, and the actual bandwidth demand of the communication data network link in the next few years; According to the business traffic growth rate of each business type, adjust the business traffic allocation ratio of each business type and the business priority of each business type in the next few years. 2 . The communication data network investment and construction evaluation method according to claim 1 , wherein the method for obtaining the bottleneck bandwidth of the communication data network link described in step S1 is a variable-length single-packet measurement method. 3 . 3. communication data network investment construction evaluation method according to claim 2 is characterized in that, the method that adopts described variable length single packet measurement method to calculate the bottleneck bandwidth of communication data network link specifically comprises: According to the principle of variable-length single-packet measurement technology, calculate the delay of data packet k on link i, The time it takes for a packet to arrive at link l from the source

for:

where ^sk is the size of the kth data packet, bi is the capacity of link _{i, d i is} the fixed delay of link i,

To measure the inherent delay caused by the host accessing the communication network; The time for packet k to leave link l is

for:

Therefore, the delay t experienced by packet k on link l is:

The above t is equivalent to the end-to-end path delay T, then we have

Among them, S is the size of the data packet, B is the bottleneck bandwidth of the end-to-end path, and D is the transmission delay of the end-to-end path; That is, the end-to-end path bottleneck bandwidth B is obtained as:

4 . The method for evaluating investment and construction of a communication data network according to claim 1 , wherein the available bandwidth in step S1 is obtained by a self-load periodic flow measurement method. 5 . 5 . The method for evaluating the investment and construction of a communication data network according to claim 4 , wherein the available bandwidth described in step S1 is obtained by measuring with a Pathload tool. 6 . 6. The communication data network investment construction evaluation method according to claim 1, is characterized in that, described step S2 specifically comprises: S21: Using a traffic statistics technology, collect service traffic of several sites in the current area, and divide the several sites into different service types according to different services; S22: Count the service traffic of each service type in the area, analyze the bandwidth requirements and service traffic growth rate of each service type in the area, and obtain the sum of the bandwidth requirements of each service type in the area. 7. The communication data network investment construction evaluation method according to claim 1, is characterized in that, described step S3 specifically comprises: S31: Predict the model according to the bandwidth demand:

And because there is a bandwidth loss C in the communication data network link, there are

Wherein, the light-load service flow in the above formula is the bandwidth requirement of the current actual communication data network link of the communication data network.

Images