WO2006105707A1

WO2006105707A1 - A network performance testing method and the system, the device thereof

Info

Publication number: WO2006105707A1
Application number: PCT/CN2006/000175
Authority: WO
Inventors: Yang Yu; Haitao Zhang
Original assignee: Hangzhou H3C Technologies Co., Ltd.
Priority date: 2005-04-04
Filing date: 2006-01-26
Publication date: 2006-10-12

Abstract

A network performance testing method comprises: emulating one similar test service flow with the service flow required of testing; transmitting the test service flow from the testing origin node to the testing end node in the network; testing the network performance based on the received test service flow at the testing end node; or after processing the received test service flow, looping back it to the testing origin node to perform the backtracking service test. The invention also discloses a testing origin node network device, a testing end node network device and a network performance testing system, which comprises the testing origin node network device, the testing end node network device and the testing management device connected through the network link each other. Using the invention, the test based on the actual wideband network service application is realized. Consequently, the network failure is judged and located roundly, quickly and accurately. The subscriber service quality and the network availability are improved.

Description

Network performance test method, system and network device

The present invention relates to the field of network communication technologies, and in particular, to a network performance testing method, system, and network device. Background technique

OAM (Operation, Management, and Maintenance) functionality is important in public telecommunications networks to simplify network operations, verify network performance, and reduce network operating costs. The OAM function is especially important in networks that provide guaranteed quality of service. According to the actual needs of the network operation of the carrier, the management of the network is generally divided into three categories. Operation and management mainly refer to the analysis, prediction, planning, and configuration of daily network and services. Maintenance is mainly for the network. And its business's daily operations activities such as 'testing and fault management. The OAM mechanism not only prevents network failures, but also needs to quickly diagnose and locate network faults, ultimately improving network availability and quality of service to users.

In order to enhance the telecommunications characteristics of Ethernet, a new Ethernet OAM (IEEE 802.3ah standard) mechanism is disclosed in the prior art. The OAM mechanism under the new IEEE802.3ah standard mainly introduces an OAM sublayer between the traditional MAC (Media Access Control) layer and the PHY (physical) layer, and transmits OAM information between devices. In addition, the network is monitored. The basic principle of the OAM and the packet format are shown in Figure 1.

As shown in Figure 1, the OAM sub-layer of the IEEE802.3ah standard inserts OAM packets in the output direction and parses and distributes OAM packets in the input direction. The OAM packet is the same as the traditional Ethernet frame format, but an OAM type is newly defined. The OAM mechanism can monitor the link through physical PING, loopback, and remote alarms.

The process of monitoring the link through physical PING is shown in Figure 2:

The PE (operator edge) device initiates a PING request, and the CE (Customer Edge) device responds to the request. After receiving the response, the PE device obtains the normal judgment of the subscriber line. In actual use, the PE device is a device of the telecommunication office, and the CE device is a device in the user's home, such as a PTN (IP telecommunication network) terminal in the user's home. It should be noted that the PING here is a PING operation of the physical layer, which is different from the commonly used IP PING operation, and is directed to the MAC address, and only the user device whose destination MAC address is its own will generate a response. The main purpose of physical PING is to check whether the link between the subscriber line and the equipment at both ends is working properly. Line loopback is also a common positioning method. The receiving end is directly connected to the transmitting end, so that the sender can check whether the received data is the same as the one sent by itself, thereby judging whether it is a problem in the internal processing of the receiving device itself. When the user link has lost packets and errors, you can set the loop to determine whether the fault is caused by the device or the line.

The remote alarm is used to notify the PE device in the reverse direction when the fault occurs on the line in one direction. For example, if the user equipment detects a large error, the device will notify the PE device in the reverse direction. The process is shown in Figure 3.

Another prior art related to the present invention is described in U.S. Patent Application Publication No. 200501169182, which discloses an MPLS-based OAM test method for inserting MPLS-OAM test frames into a normal MPLS flow. Check connectivity, packet loss rate, etc., to provide operators with fast network performance testing and optimize their VPN services.

These OAM mechanisms enhance the fault location capability of Ethernet to a certain extent, but the following problems of Ethernet OAM remain unresolved:

Both the traditional OAM test and the above MPLS-OAM are only tested under the third layer, but cannot solve the second layer or above (including IP (Internet Protocol) layer, TCP/UDP (Transmission Control Protocol/User Datagram Protocol). The actual situation of the network at the layer and higher level). In many cases, the continuity of the network at the second or third layer cannot determine the continuity of the higher layer services. In actual applications, there are many devices in the operator's network, and the management is quite difficult. If the administrator is on one of the devices. Improper management configuration may cause business problems. For example, the configuration of the device ACL rules is not well thought out, which may cause business interruption of many VPN users. When a customer complaint is caused, the operator uses traditional means to detect, because the network is Both the second and third floors are unobstructed, so it is impossible to locate the fault and the customer problem is difficult to solve.

As can be seen from the above description, the Ethernet OAM is located at the second layer in the network hierarchy, and the tested forwarding path is symmetric. However, many of the actual services of users on the network are applied in the high-level protocols of the network hierarchy, and in most cases, the paths are asymmetric, and the traffic of the uplink and downlink is also used for most of the access in terms of traffic. It is asymmetrical. In this case, how to provide business security for users, there is currently no effective solution. Summary of the invention

The object of the present invention is to provide a network performance testing method and system, which can overcome the failure of the prior art to provide a test based on the service itself, and realize rapid diagnosis of network failures. Positioning, improving the availability of the network and the quality of service to users.

Another object of the present invention is to provide a test start network device and a test end network device to provide a test based on the service itself to the network and to quickly locate the network fault.

To this end, the present invention provides the following technical solutions: A network performance testing method, including the steps:

A. Test the test starting device and the test end device;

B. At the test starting point, simulate the test service flow carrying the specific service application according to the test setting;

c. Sending the test service flow from the test starting point of the network to the test end point;

D. Obtain a test service flow at the test end point for analysis use; or loop the obtained test service flow back to the test starting point to perform a return service test.

Preferably, the step B includes:

The test service flow is inserted into the normal service flow in a predetermined proportion, or the test service flow is inserted after the normal service flow is aborted.

Preferably, the test setting includes an analog configuration for generating a test service flow, and identification information of the test service flow; the simulation in the step B is performed based on the simulation configuration; and the insertion identification information is further included.

Preferably, the simulation configuration includes test end point information and a test service flow generation manner; the identification information inserted in the step B includes at least one of a test identifier, a time stamp, or a data packet identifier; the identifier information is inserted in the data. The specified location of the package.

Preferably, the test setting further includes a loopback setting; the loopback of the step D is performed based on a loopback setting.

Preferably, the loopback is set to an asymmetric loopback setting, and the step D includes setting a packet return rate or a packet size in the test service flow.

Preferably, the test setting further includes an access control table; and the obtaining the test service flow in the step D is performed based on the access control list.

The network performance testing system provided by the present invention comprises: a test starting point device and a test end point device connected through a network, and a test management device for managing a test process; the test starting point device is simulated and inserted according to the test setting. Testing the service flow of the specific service application, and sending the test service flow to the test end point device; the test end point device extracts the test service flow according to the configured test configuration, and provides the test service flow according to the extracted test service flow The service quality of the network is analyzed, or the obtained test service flow is looped back to the test starting point to perform a return service test.

Preferably, the test setting is performed by the test management device to the test starting device and the test The endpoint device configuration; or the test endpoint device address information and test service type in the test management device configuration test settings, the test start device and the test endpoint device negotiate the remaining facility settings or statically configure the test settings in advance.

Preferably, the test starting device comprises:

a test setting storage unit for storing test settings;

a test stream insertion unit coupled to the test setting storage unit, configured to generate a test service flow according to the stored test set simulation, and insert the test service flow into a normal service flow sent by the test start device;

The test endpoint device includes:

a test setting storage unit for storing test settings;

And an end point test stream extracting unit, configured to extract a corresponding test service flow from the service flow received by the test end point device according to the test setting;

The end point test processing unit is coupled to the test setting storage unit and the test stream extraction unit, respectively, for performing analysis processing on the test service flow according to the test setting.

Preferably, the analyzing process performed by the key test processing unit includes looping back the test service flow to the test starting device; the test starting device further includes a starting test flow extracting unit, configured to receive normal services from the test starting device. The test traffic flow returned from the test endpoint is extracted from the flow.

Preferably, the test starting point device and the test end point device are respectively network devices having a central processing unit and a storage function.

Preferably, the test set stored by the test endpoint device includes an asymmetric loopback setting.

Another network performance testing system provided by the present invention includes a network management device, a test starting device, and a test end device.

The network management device is configured to send test settings to the specified test start device and the end device;

The test starting device generates a test service flow according to the test setting; the test end device identifies the test service flow sent by the test starting device according to the test setting, and generates a return test service flow to the test starting point according to the test setting. The device is sent in the direction.

Preferably, the test setting stored by the test end point device includes a test setting of the asymmetric flow, and the data rate or the size of the return test service flow generated by the test end device according to the above setting is different from the test service flow sent by the test device. .

The test start network device provided by the present invention has a network management interface, and is configured to receive test settings sent by the test management device, where the network device further includes: a test setting storage unit for storing test settings received from the test management device;

The test stream insertion unit simulates a specific application-based test service flow according to the test setting, and sends the test service flow to the test end point network device specified by the test setting.

Preferably, the test starting network device further includes:

The starting point test stream extracting unit is configured to identify the test service flow received from the test end network device, and extract the returned test service flow for use in management analysis.

Preferably, the test setting includes an analog configuration for generating a test service flow, a ratio setting of a test service flow to a normal service flow, and at least one of a test identifier, a time stamp, or a packet identifier.

Preferably, the specific application is a service higher than a network layer.

Preferably, the test traffic flow is simulated by writing corresponding parameters to the device hardware registers according to the analog configuration included in the test settings to enable it to quickly generate test traffic flows.

The test destination network device provided by the present invention has a network management interface, which is used for receiving test settings sent by the test management device, and the network device further includes:

a test setting storage unit for storing test settings received from the test management device;

And an endpoint test stream extracting unit, configured to extract a corresponding test service flow from the service flow received by the test end network device according to the test setting.

Preferably, the test end network device further comprises: an end point test processing unit, configured to loop back the test service flow to the test end network device according to the test setting.

It can be seen from the technical solution provided by the present invention that the present invention simulates the test service flow with the same format as the specific service application in the input direction of the service inflow port of the network test starting device, and inserts it into the normal service flow and sends it to the normal service flow. Testing the end point device; the test service flow is the same as the normal service flow, and the normal forwarding process of the actual network device between the test starting point and the test end point reaches the test end point device; the test service flow is extracted at the out port of the test end point device, It performs OAM statistics to implement end-to-end service OAM testing, such as connectivity testing, packet loss testing, and bandwidth testing. Since the test service flow is the same as the service flow format of the actual application, the OAM function of the network above the MAC layer can be detected; the address information and the data frame size and/or the return data frame in the extracted test service flow are replaced by the test destination device. The number of the test service flows is returned to the test starting device through a symmetric path or an asymmetric path to perform asymmetric service testing, and the backhaul link is further detected. Therefore, with the present invention, It can provide real-time, end-to-end O AM testing for users or network operators to diagnose and locate network faults comprehensively, quickly, and accurately, thereby improving network availability and quality of service to users. DRAWINGS

Figure 1 is a schematic diagram of the operation, maintenance, and management mechanism of the Ethernet described in the 802.3ah standard;

Figure 2 is a schematic diagram of the link monitoring process through physical PING or line loopback in the 802.3ah standard;

3 is a schematic diagram of a link monitoring process by a remote alarm in the 802.3ah standard; FIG. 4 is a flowchart of an implementation of an embodiment of the method of the present invention;

Figure 5 is a flow chart showing an implementation of another embodiment of the method of the present invention;

6 is a schematic structural diagram of a VPN networking of the same L2TP;

7 is a schematic diagram of networking of an embodiment of the system of the present invention;

Figure 8 is a block diagram of another embodiment of the system of the present invention. detailed description

The core of the present invention is to simulate a test service flow carrying a specific service application in the input direction of the service inflow port of the test starting device, and carry the test service flow carrying the test identifier or time stamp and the like, and then at the test starting point The test end device is inserted into the normal service flow and sent to the test service flow; the test service flow is normally forwarded to the test end point device; and the test end point device extracts the test service flow according to the test identifier, and the service-based analysis can be directly performed. The test service flow can also be looped back to the test start device, and the test start flow device analyzes the test service flow returned from the test end point.

In order to make those skilled in the art better understand the present invention, the present invention will be further described in detail below in conjunction with the drawings and embodiments.

Those skilled in the art know that on the one hand, as the demand of users continues to increase and the development of Ethernet bandwidth technology and the QoS (Quality of Service) support capability of the entire network are improved, the network provides end-to-end (SLA) service level agreements to users. The guaranteed business is already a requirement and trend. On the other hand, although the network of operators is constantly expanding and improving, the network throughput is quite large, but this does not mean that all kinds of practical application services of access users can be transmitted smoothly on the carrier network, as mentioned above. The quality of the network cannot determine the quality of the business. However, how to verify whether the services provided by the network meet the requirements of the SLA, there is currently no way to implement end-to-end service-based testing of the quality of service provided by the network, and the present invention is adapted to meet such needs. Before the test starts, it is first necessary to test the starting device and the terminal device that initiate the test. At least the test starting device needs to know the correct address information of the test terminal device and how to generate the service flow, that is, the simulation configuration, and the simulation configuration can be passed. Write a byte of a specific length that can characterize the service application to be tested to the register of the starting device, and then generate a test service flow, and then insert it through hardware; then both parties need to know how to identify and identify the received test service flow, ie Set the test ID or timestamp or packet ID, and whether to set the loopback test. The identification and loopback settings can be implemented by issuing ACL rules. If the test is separate from the normal service of the user, you can also set it. Test identification, because only the test traffic flow does not need to be identified, you can set the test service flow timestamp or packet ID to match the test; if you only need one-way test, you do not need to perform loopback setup; If you perform an in-band test, you can set the insertion rate. ; If you want to test asymmetric traffic, the rate may be set or loopback packet size to be amended; described above shows that the present invention has a considerable flexibility in the test set. The above test settings are generally configured by the test management center or the network management system through the test management channel. The test settings themselves can also be negotiated by the device. For example, the management device only needs to address the terminal device address and needs. The basic information such as the type of service to be tested is informed to the starting device, and other test settings can be determined by mutual agreement between the two parties or static configuration in advance. The specific implementation of the invention is further described below:

Referring to the flowchart of the method of the present invention shown in Fig. 4, the method of the present invention comprises the following steps: First, after the test setup is completed, in step 401: a test service flow is generated. In the ingress port of the service flow of the test starting device, a test service flow based on the service application is generated according to the test setting, so that the test service flow may include information indicating the application, that is, the MAC layer or more (including the IP layer, Some information about the TCP/UDP layer and higher level, for example, when the service is FTP (File Transfer Protocol), the message format of the test service flow is generated according to the FTP protocol and the address information, and the port is generally 21.

Step 402: Insert a test service flow at the ingress port of the test start point.

Considering the advantages of in-band testing, the test service flow can be inserted into the normal service flow according to a certain proportion before the normal forwarding process of the service data packet, for example, the test service flow and the normal service flow each account for 50% of the traffic. . Of course, it is also possible to suspend the normal service flow and insert the test service flow.

The insertion of the test service flow can be done by hardware cooperation (of course, it can also be implemented by software). The specific process is as follows:

According to the type of business application to be simulated and the test settings, the data packet characteristic byte content of the generated test service flow, such as the content of a certain byte before the packet header: Destination MAC address The address, source MAC address, destination IP address, source IP address, length/type, subtype, version number, port number, encoding, etc. are written into the test traffic stream feature register, the hardware reads these feature registers, and the payload is automatically inserted. And a padding word (the payload and padding settings can be included in the test settings), and automatically inserting at least one of a test flag, a timestamp, or a packet ID according to the test settings, the information can be inserted in the payload, or Is the MAC header of the packet or any specified location such as the IP header.

The payload data required for testing the stream data packet may be pre-written in the memory. When inserted, the payload data of the corresponding length is read from the memory according to the packet length information in the test traffic stream feature register, and the test stream data may also be randomly generated. The payload data required for the package.

The insertion speed of the test service flow can be adjusted according to the traffic volume of the test service flow. The insertion of the test service flow can be mutually exclusive with the normal service flow. That is to say, according to the actual needs, the normal service flow needs to be stopped during the test; or it can be performed at a certain ratio, for example, the test service flow and the normal service flow. 50% of the traffic, etc.

In step 403: the test service flow is forwarded between nodes of the network according to the normal forwarding process of the service.

These test service flows simulate the service flow of a specific business application, so in the process of sending the test start device to the test end device, each network device has the same decision for the test service flow forwarding, and it seems to be normal in the network node device. Like the service flow, the normal service flow forwarding process is performed according to the destination address it contains.

Once the test traffic is forwarded to the test endpoint, these services can be extracted and analyzed to determine network health.

Step 404: Detect network performance according to the obtained test service flow at the test end point. As described above, the test endpoint device can use the ACL to identify the test identifier of the test service flow, or the timestamp or the packet ID according to the test setting, thereby extracting the corresponding test service flow and sending it to the upper layer software for line analysis, such as When the test service flow includes a timestamp, the connection speed and time delay of the starting device and the destination device may be determined according to the number of received data packets per unit time, which will be described in detail later.

In other words, the test setting includes an access control list of the test destination port pre-configured by the test management device, and the address information of the sender and the receiver, the test stream ID, and the like are included in the table. At the outbound port of the test device, the test device extracts the corresponding test data stream according to the information, and then performs operation management and maintenance statistics on the extracted test data stream, for example, network connectivity, error statistics, and packet loss rate statistics. (Depending on the number of packets or the ID order corresponding to the time stamp).

Referring to FIG. 5 together, in addition to the direct processing, in consideration of the fact that the path of the user's uplink and downlink is asymmetrical in most cases, it is also possible to proceed to step 405. Step 405: Perform a loopback test. In this step, the terminal device processes the extracted test service flow and returns it to the test starting point to perform a return service test. The return process is based on the test setting, and the ACL rule is used to perform the packet header. And the source address and so on are modified accordingly and sent back to the starting device. When traffic asymmetry testing is required, for example, when the user uses ADSL access or uplink and downlink asymmetric VPN access, the bandwidth of the uplink and downlink is asymmetric, and the terminal device can change according to the analog configuration in the test setting. The rate or the length of the data packet is simulated to test the rate of the user's uplink and downlink, so that the whole test will be closer to the user's actual practical situation, and the location of the network fault will be more accurate and complete.

The test end point and the test starting point can mainly perform the following network performance tests.

Statistics of connectivity and packet loss rate:

It is obtained by comparing the number of data packets received within a certain period of time with the number of transmitted data packets. The number of data packets received in a certain period of time is counted by the hardware, and the packet loss rate is calculated by the software according to the number of hardware statistics.

(2) Calculation of actual use bandwidth:

Calculated based on the traffic received in the unit time.

(3) Test of network transmission delay:

Calculate the time delay of network transmission based on the time I carry information in the test packet. Can be divided into two-way delay test and round-trip delay test.

When performing the one-way delay test, the test start device and the test end device synchronize the network time through the network time protocol. Each device then configures the time into its own time register, which is automatically incremented by hardware. When the test starting point device sends the test service flow, the test data frame is marked with a local timestamp. During the process of passing the time-stamped data packet through the network, the timestamp information does not change. After the test end device is reached, the test end device extracts the time stamp from the test service flow and compares it with the current local time to obtain the time interval between the two, that is, the network transmission delay of the data packet. .

If the one-way transmission delay test is performed, since the transmission and reception are performed between two different devices, and the network time protocol itself has a certain time error, the clocks of the two remote devices are not accurately synchronized, so one-way operation is adopted. When the transmission is tested for delay, there will be a certain time error in the test results.

In order to obtain network transmission delay more accurately, round-trip delay testing can be performed. After receiving the test data packet, the test endpoint device does not change the timestamp information flag marked on the test start device, and loops the test data packet according to the original path. When the test starting device receives the test packet with the time stamp marked by itself from the backhaul, the timestamp is extracted, and according to the connection The current time information when the test packet is received, and the accurate round-trip transmission delay is obtained. Since the test traffic flow is sent and received during the round-trip delay test, the reference time of one device used is obtained, so that an accurate test value can be obtained.

(4) Return business test:

According to the configuration of the test endpoint access control list, the received test data packet is processed accordingly, and then looped back to the test starting point, and the asymmetric path, asymmetric traffic flow, and the like can be tested.

The present invention performs end-to-end testing of any service based on actual network by simulating the same test service flow as the service flow for different service tests, and provides better monitoring means for network QoS.

The following describes the test service flow generated in the present invention by taking the encapsulation format of the Layer 2 tunneling protocol L2TP as an example.

Those skilled in the art know that there are two types of L2TP encapsulation formats, one is encapsulated in IP data packets, and the other is encapsulated in IP UDP (User Datagram Protocol) data packets. The following is an example of L2TP encapsulation in IP data.

The L2TP encapsulation format is as shown in Table 1. It is a format in which an Ethernet data frame is encapsulated in an IP tunnel, and includes a two-layer MAC header and two layers of IP headers.

Table 1: Destination MAC address 48 bits Source MAC address 48 bits Ethernet encapsulation header format Fixed 8100 16 bits

(18 bytes) VLAN ID and priority 16 bits

EtherType 16 bit version number 4 bits

IP header length 4 bits Service type 8 bits

IP header encapsulation format IP packet total length 16 bits (20 bytes) IP packet flag 16 bits

IP segmentation flag 3 bits IP segment offset 13 bits

TTL survival time 8 bits

IP protocol domain 8 bits

IP source address 32 bits

IP destination address 32 bits

L2TP header control 12 bits

L2TP header version 4 bits

L2TP header encapsulation format reserved field 16 bits (16 bytes) Session ID 32 bits

Cookie (optional) Max 64 Bit Ethernet Over PPP Head Fixed 00-31 16 bits (4 bytes) Flag and padding 8 bits

Type (fixed to 01) 8 bits Destination MAC address 48 bits Source MAC address 48 bits User Ethernet encapsulation header format (18 fixed 8100 16-bit bytes) VLAN ID and priority 16 bits

EtherType 16 bit version number 4 bits

IP header length 4-bit service type 8-bit user IP header encapsulation format (20 words IP packet total length 16 bits) Section) IP packet flag 16 bits

IP segmentation flag 3 bits

IP segment offset 13 bits

TTL survival time 8 bits

IP protocol domain 8 bits

IP source address 32 bits

IP destination address 32-bit user static load user static load length variable user IP packet check (4 Byte) IP packet check 32-bit user Ethernet data frame check Ethernet data frame check 32-bit (4Byte) Encapsulation tunnel IP packet check (4B te) IP packet check 32-bit encapsulated tunnel Ethernet data frame check Ethernet data frame check 32 bits (4Byte) Due to standard Ethernet Over PPP The CRC (Cyclic Redundancy Check) of the (bearing point-to-point protocol) is optional, so the last check in the above table is 4 parts, which are the MAC header and IP header for the encapsulated tunnel, and the MAC header of the user service. And the verification of the IP header. The format of the L2TP encapsulation header is 92 bytes.

For example, user A and user B are in the same L2TP VPN (Virtual Private Network), and the two communicate with each other through the network. The networking structure is shown in Figure 6. User A accesses the network through gateway A. User B accesses the network through gateway B. When user A sends a service to user B, user A's MAC address and IP address are used as the source MAC address and source IP address of the inner MAC header and IP header, and user B's MAC address and IP address are used as the inner MAC header and IP address. The destination MAC address and IP address of the header.

The MAC address and IP address of the outer layer are determined by the addresses of gateway A and gateway B of the outer tunnel. When user A sends traffic to user B, the MAC address of gateway A The address and IP address are used as the source MAC address and source IP address of the outer package, and the MAC address and IP address of the gateway B are used as the destination MAC address of the outer package and the IP address of the D.

Before conducting the test, gateway A should negotiate with the gateway B for test rules, or be configured by the network administrator or manually for test setup. For example, the test identifier can be represented by a reserved field of 16 bits of the L2PT header, such as the value 20; by negotiation or by configuring the destination device ACL, the loopback is selected.

When performing the routine test, the terminal device replaces the 92 bytes in front of the packet header in the test service flow with the contents of the register according to the configured ACL action, and modifies or maintains the original time stamp content from the test endpoint network. The egress port of the device loops back to the upstream processing, so that the test service flow automatically returns to the test starting point.

When performing asymmetric traffic test, according to the action of the ACL (access control list), all 92 bytes of the header in the test service flow are replaced with the contents of the register, and the original time stamp content is modified or maintained. At the same time, depending on the configuration, the length of the test packet can be changed, or the number of transmitted packets can be changed. Then, from the test terminal network device's egress port loop back to the upstream processing, the test service flow is automatically returned to the test starting point.

Assume that user A now finds that it cannot perform normal service connection with user B; it complains to the operator about this situation; if the operator initiates the connectivity test from gateway A to gateway B by using the traditional method, the network does not find that the network has Any exceptions cannot resolve customer complaints. The reason for this problem is that the operator's test is based on the network itself that does not carry the application; and the customer's problem is that the network connection carrying the specific application is not normal. For example, a network device between gateway A and gateway B blocks user A's private IP address, which cannot be found in traditional tests.

By using the above specific test method of the present invention, as long as the test service flow encapsulated with the user data packet is simulated and tested, the problem device can be quickly found.

Further, it is assumed that the FTP service between User A and User B is unreachable, and other services are forwarded normally. At this point, the test setup needs to be re-executed, and then the starting device generates a test service flow based on the FTP protocol according to the simulation configuration, so that the test service flow is based on the FTP application. Next, the operator can choose to start the test one by one from the gateway A, or use the faster dichotomy, that is, if there are 4N devices between the gateways AB, the test between the starting device and the 2N device is performed first, if the test After passing, the test is performed with the 3N device, and if it is not passed, the test is performed with the Nth device..... The test is repeated until the network device blocking the FTP service is found. Generally speaking, the operator administrator checks the configuration of the device, such as the ACL rule, and finds that the FTP service is blocked. If the ACL rule of the network device is incorrectly configured, the FTP service is blocked. The administrator can restore the FTP service of the user by making the corresponding modification. Assume that the user FTP service is not blocked, but the transmission rate is quite slow. After the test settings can be modified, use the above method to find the bottleneck device that causes the transmission rate to be slow. Modify the network device configuration or restore the device hardware and software. The normal transmission rate of the user.

Referring to Figure 7, there is shown a block diagram of an embodiment of a system of the present invention: The system includes: a test management device 600, a test origin device 601, and a test endpoint device 602, which are connected by a network link, for example, Coaxial wires, optical cables, etc.

The interconnection of different networks requires different network node devices to implement, and the service data may need to be forwarded by multiple node devices from the source node to the destination node. Depending on the network topology, there may be one or more network node devices (not shown) between the test origin network device 601 and the test endpoint network device 602, such as repeaters, bridges, routers, Gateways, network switches, etc.

The test management device 600 completes the parameter configuration and test process management of the test start network device 601 and the test end network device 602.

When performing the network service quality test, the test management device 600 first sends the test setting information to the test start device 601 and the test end device 602 through a dedicated test management channel or a physical channel used by the network management. For example, the simulation configuration used to generate the test traffic, the ratio of test flows to normal traffic, the test data flow that needs to be inserted, the performance and statistics that need to be detected, the test ID or timestamp, or the packet ID. The test setup of the endpoint device 602 includes an asymmetric loopback setting, i.e., the packet of the test traffic stream sent by the endpoint test processing unit differs in rate or magnitude from the received test traffic.

At the ingress port of the test origin device 601, the test service flow is inserted according to the configured parameter information, and the test service flow is application-based, that is, the test service flow data packet is content containing more than three layers. The test traffic is then sent along with the service flow to the test endpoint. These test traffic flows, like the traffic flow, are forwarded by the network node device between the test origin device 601 and the test endpoint device 602 to the test endpoint device 602.

After receiving the data, the test endpoint device 602 extracts the test service flow according to the configured parameter information, and detects the service quality of the network according to the extracted test service flow. Generally, the test management device 600 can send the outbound port ACL to the test end point device 602 before the test, and the test end point device 602 identifies the characteristics of the test service flow according to the egress port ACL and performs corresponding test processing, for example, according to the configuration. The address information of the sender and the receiver, the test service stream ID information, and the like perform operations such as error statistics and packet loss rate statistics. Since both the test start device and the test end device require some software or hardware processing, in general, both the test start device and the test end device need to have a CPU (Central Processing Unit) and storage functions, for example, a network with a CPU. Switch or other network device. At the same time, the starting point or destination device needs to have a network management interface for accepting test settings issued by the test management device or other network management devices.

As shown in FIG. 7, the test starting point device 601 includes: a test setting storage unit 611, a test flow inserting unit 612, and a starting point business flow extracting unit 613.

After receiving the test setting sent by the test management device 600, the test start device 601 stores it in the test setting storage unit 611, and when it needs to be inserted, the test stream insertion unit 612 reads the data at its out port, and Insert it into a normal business flow in a certain proportion. During the insertion of the service flow, the test service flow can be inserted into the normal service flow according to the traffic volume of the test service flow, or the test service flow can be inserted after the normal service flow is aborted.

The test endpoint device 602 includes: a test setup storage unit 621, a test traffic flow extraction unit 622, and an endpoint test processing unit 623.

After receiving the test settings sent by the test management device 600, the test destination device 602 stores them in the test setting storage unit 621. After receiving the service flow forwarded by the network node device, the test service flow extraction unit 622 stores the data according to the storage. The test setup extracts the corresponding test traffic from these traffic flows. The test processing unit 623 performs direct analysis or loopback processing on the extracted test traffic according to the test flow feature information.

If loopback processing is required for the test service flow, the test setting including the loopback setting (available by the ACL) may be sent by the test management device 600 to the test endpoint device 602 through the egress port, and the test endpoint device 602 extracts the corresponding After testing the service flow, according to the action of the ACL, perform corresponding packet processing, such as replacing some bytes in the packet header according to the action indication, and putting the corresponding time stamp content, and returning from the exit port of the test destination network device. Test the starting network device; or if the test setting includes an asymmetric traffic test setting, modify the rate or size of the returned packet so that the returned test traffic is different from the original test traffic, simulating the asymmetric traffic ratio setting set.

After the test start device 601 receives the return test service flow forwarded by the network node device, the start test flow extraction unit 613 extracts the corresponding return test service flow from the service flows, and performs performance analysis on the flowback, so that the backhaul can be known. The health of the link.

The test setting storage unit 611, the test flow insertion unit 612, the start service flow extraction unit 613, the start point test processing unit 614, the test setting storage unit 621, the test service flow extraction unit 622, and the end point test included in the above-described start point and destination devices are included. The processing unit 623 is based on logical division, which is variable in the implementation form of software and hardware. The technical personnel of the domain can have different software and hardware implementations through the introduction of the above, but these methods all belong to the form changes without departing from the spirit of the present invention. In particular, the static configuration of the test settings mentioned in the previous section. In addition to the address information of the test destination device and the service type of the test, the simulation configuration and loopback of various applications can be completely configured in advance. The setting, insertion rate setting, etc. are stored in the test setting storage unit, and multiple sets of test settings can be made to implement any of the settings for multiple tests, such as setting one-way test tasks, loopback test tasks, and asymmetric flow tests. task.

The system of the present invention can be applied to a variety of networks, such as Ethernet, RPR (Flexible Packet Ring) networks, and the like. The aforementioned identification technology is not limited to the use of ACL rules, but can also be implemented by other technologies, such as Cisco's NBAR (application-based identification) technology; analog configuration can be implemented not only in hardware, but also in software.

For the principle of the test starting point network device and the test end point network device of the present invention, reference may be made to the test starting point device 601 and the test end point device 602 in FIG. 7 , and details are not described herein again.

Please refer to FIG. 8, which is a block diagram of another embodiment of the system of the present invention. The embodiment is substantially the same as the embodiment shown in FIG. 7. The difference is that the test starting point device in the embodiment does not have a starting point test stream extracting unit, which is not described here.

While the invention has been described by the embodiments of the invention, it will be understood that

Claims

Rights request

A network performance testing method, characterized in that the method comprises the steps of: A: testing and setting a test starting device and a test end device;

C. Sending the test service flow from a test starting point of the network to a test end point;

2. The network performance testing method according to claim 1, wherein the step B comprises:

3. The network performance testing method according to claim 1, wherein the test setting comprises: an analog configuration for generating a test service flow, and identification information of a test service flow; and the simulation in the step B is based on The simulation configuration is performed; it also includes inserting identification information.

The network performance testing method according to claim 3, wherein the simulation configuration includes test end point information and a test service flow generation manner; and the identification information inserted in the step B includes a test identifier, a time stamp, or a data packet. Identifying at least one of the three; the identification information is inserted at a specified location of the data packet.

The network performance testing method according to claim 4, wherein the test setting further comprises a loopback setting; and the loopback of the step D is performed based on a loopback setting.

The network performance testing method according to claim 5, wherein the loopback is set to an asymmetric loopback setting, and the step D includes returning a data packet rate or data in the test service flow. The package size is set.

The network performance testing method according to claim 1, wherein the test setting further comprises an access control table; and the obtaining the test service flow in the step D is performed based on the access control list.

8. A network performance test system, comprising: a test start device and a test end device connected through a network, and a test management device for managing a test process; wherein the test start device simulates and inserts according to a test setting Carrying a test service flow for a specific service symptom, and sending the test service flow to the test end point device; The test end point device extracts the test service flow according to the configured test configuration, and analyzes the service quality of the network according to the extracted test service flow, or loops the obtained test service flow back to the test starting point, and performs a return journey. Business testing.

9. The network performance testing system according to claim 8, wherein the test setting is configured by the test management device for the test start device and the test end device; or by the test management device configuration test setting. The test endpoint address information and the test service type, the test start device and the test endpoint device negotiate the remaining facility settings or statically configure the test settings in advance.

The network performance testing system according to claim 8 or 9, wherein the test starting device comprises:

a test setting storage unit for storing test settings;

a test stream insertion unit coupled to the test setting storage unit, configured to generate a test service flow according to the stored test set simulation, and insert the test service flow into a normal service flow sent by the test starting device;

The test endpoint device includes:

a test setting storage unit for storing test settings;

The network performance testing system according to claim 10, wherein the analyzing process performed by the key test processing unit comprises looping the test service flow back to the test starting device; the test starting device further includes a starting test flow And an extracting unit, configured to extract, from the normal service flow received by the test starting device, a test service flow returned from the test end point.

12. The network performance testing system according to claim 9, wherein the test starting point device and the test end point device are respectively network devices having a central processing unit and a storage function.

13. The network performance testing system of claim 9, wherein the test settings stored by the test endpoint device comprise asymmetric loopback settings.

14. A network performance test system, comprising a network management device, a test start device, and a test end device, wherein:

The test starting device generates a test service flow according to the test setting described above; The test end point device identifies the test service flow sent by the test start device according to the test setting, and generates a return test service flow to send to the test start device according to the test setting.

The network performance testing system according to claim 14, wherein the test setting stored by the test end point device comprises a test setting of asymmetric traffic, and the test end service device generates a return test service flow according to the foregoing setting. The packet rate or size is different from the test traffic sent by the test equipment.

A test start network device, having a network management interface, configured to receive test settings sent by the test management device, wherein the network device further includes:

The test starting point network device according to claim 16, wherein the test starting point network device further comprises:

18. The test starting point network device according to claim 16, wherein the test setting comprises an analog configuration for generating a test service flow, a ratio setting of a test service flow to a normal service flow, and a test identifier, The timestamp or packet identifies at least one of three.

The test starting point network device according to claim 16, wherein the specific application is a service higher than a network layer.

20. The test starting point network device according to claim 16, wherein the simulation of the test service flow is performed by writing a corresponding parameter to a device hardware register according to an analog configuration included in the test setting. Generate test traffic.

A test end network device, having a network management interface, configured to receive test settings sent by the test management device, wherein the network device further includes:

The test end point network device according to claim 21, wherein the test end point network device further comprises: The end point test processing unit is configured to loop back the test service flow to the test end network device according to the test setting.