CN103312607B - A kind of selection of transmission paths method and device - Google Patents

Abstract

The present invention provides a method and device for selecting a transmission path. The method includes: a first device in the network sends a part or all of the transmission paths between the first device and the second device, including routing factors, first detection parameter and quintuple probe packet, and select the routing factor of the transmission path with light load as the routing factor of the data packet to be sent according to the probe response packet fed back by the second device. The method enables data to always avoid transmission via heavily loaded transmission paths, thereby avoiding the problem of data congestion in the network.

Description

A transmission path selection method and device

technical field

The present invention relates to the communication field, in particular to a transmission path selection method and device.

Background technique

With the development of cloud computing and data centers, network equipment is required to provide greater switching capacity. At present, the network equipment switching capacity is mainly realized by cascading multiple switches in a fat tree (FatTree) topology. promote.

A fat tree is a typical multi-level interconnected network structure. For example, in the fat tree network structure shown in Figure 1, each edge device has four equivalent routes to other edge devices, and the data sent by the edge device The peer edge device can be reached through any route. In order to avoid the occurrence of network congestion as much as possible, it is necessary to ensure the load balance of each equal-cost route. In order to achieve load balance, the data path selection method based on the hash (HASH) algorithm is usually used in the fat tree network, that is: assigning Packets with the same number of equivalent routes. However, since the size of each data packet is different, the existing data path selection method cannot guarantee that the data traffic transmitted by each equivalent route in the fat tree network structure is balanced.

Contents of the invention

In view of this, the present invention provides a transmission path selection method and device, aiming to solve the problem that the existing path selection method for fat tree network structure cannot guarantee the load balance of each data transmission path in the network.

To achieve the above object, the present invention provides the following technical solutions:

A transmission path selection method, comprising:

The first device sends a detection packet _k to the N transmission paths between the first device and the second device according to the preset sending rules, and the detection packet _k includes the routing factor _k , the first detection parameter and five elements group, the routing factors in the detection packets of different transmission paths are different, wherein, N is a positive integer greater than 1, k is a positive integer and k=1, 2...M, M is greater than or equal to N positive integer;

The first device respectively receives a probe response packet _k from the second device in response to the probe packet _k from the N transmission paths, and the probe response packet _k includes a routing factor _k , a second probe parameter, and the said five-tuple;

The first device selects a routing factor of the data packet to be sent according to the M probe response packets, and determines a transmission path of the data packet to be sent according to the routing factor of the data packet to be sent.

A routing device, comprising:

A detection packet sending unit, configured to send a detection packet _k to the N transmission paths between the first device and the second device according to a preset sending rule, and the detection packet _k includes a routing factor _k , a first detection Parameters and quintuples, the routing factors in the detection packets of different transmission paths are different, where N is a positive integer greater than 1, k is a positive integer and k=1, 2...M, M is greater than or equal to N positive integer;

A detection packet receiving unit, configured to receive a detection response packet _k from the second device in response to the detection packet _k from the N transmission paths, and the detection response packet _k includes a routing factor _k , a second detection parameter and said quintuple;

The path selection unit is configured to select a routing factor of the data packet to be sent according to the M probe response packets, and determine a transmission path of the data packet to be sent according to the routing factor of the data packet to be sent.

In the transmission path selection method and device described in the embodiments of the present invention, by sending detection packets to each data transmission path respectively, the load condition on each transmission path is known, and the path selection factor corresponding to the light-loaded transmission path is written into In the data packet to be sent, this transmission path is selected for transmission of the data packet, so that the data packet can always select the transmission path with a lighter load through the routing factor and avoid transmission through the transmission path with a heavy load, thereby realizing transmission in the network Path load balancing.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only These are some embodiments of the present invention. Those skilled in the art can also obtain other drawings based on these drawings without creative work.

Figure 1 is a schematic diagram of a fat tree structure;

FIG. 2 is a schematic diagram of a simplified network structure applied in an embodiment of the present invention;

FIG. 3 is a flow chart of a method for selecting a transmission path disclosed in an embodiment of the present invention;

Fig. 4 is a diagram of a method in which the first device sends detection packets _k to the N transmission paths between the first device and the second device according to preset sending rules in a method for selecting a transmission path disclosed in an embodiment of the present invention flow chart;

FIG. 5 is a schematic structural diagram of a detection packet in a method for selecting a transmission path disclosed in an embodiment of the present invention;

FIG. 6 is a schematic structural diagram of a transmission path selection device disclosed in an embodiment of the present invention;

7 is a schematic structural diagram of a detection packet sending unit in a transmission path selection device disclosed in an embodiment of the present invention;

Fig. 8 is a schematic structural diagram of a path selection unit in a transmission path selection device disclosed in an embodiment of the present invention.

detailed description

The invention discloses a transmission path selection method and device to solve the problem in the prior art that the load of each transmission path is unbalanced when selecting the transmission path. The general idea of the present invention is that the first device in the network respectively sends a detection packet including routing factor, first detection parameter and quintuple to some or all transmission paths between the first device and the second device, and according to The probe response packet fed back by the second device selects the routing factor of the transmission path with a lighter load as the routing factor of the data packet to be sent, so that the data packet to be sent is transmitted on the transmission path with a lighter load.

The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some, not all, embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present invention.

The present invention discloses a method and device for selecting a transmission path. For the convenience of description, the technical solutions provided by the embodiments of the present invention will be introduced below in conjunction with the fat tree network shown in FIG. 2 . It should be noted that the technical solutions disclosed in the present invention are applicable to any fat tree network, and the network scale is not limited. As shown in FIG. 2 , it includes terminals PC1 and PC2, edge devices S1 and S4, and intermediate devices S2 and S3. Data transmission paths for terminals PC1 and PC2 include S1-S2-S4 and S1-S3-S4.

The invention discloses a transmission path selection method, as shown in Figure 3, comprising:

301: The first device sends a detection packet _k to the N transmission paths between the first device and the second device according to a preset sending rule, and the detection packet _k includes a routing factor _k and a first detection parameter And five-tuple, the routing factor in the detection packet on different transmission paths is different;

Wherein, N is a positive integer greater than 1, k is a positive integer and k=1, 2...M, and M is a natural number greater than or equal to N.

Wherein, the first device refers to a network device that executes the path selection method, including routers, switches, and the like. The second device refers to the destination network device indicated in the data packet, usually an edge switch or an edge router. For example, in Figure 2, in the transmission path from PC1 to PC2, the first device can be S1, S2 or S3, and the second device can be for S4.

As shown in Fig. 2, in the present embodiment, there are two transmission paths from PC1 to PC2, and the detection packets are sent to these two transmission paths respectively, that is, S1 sends a detection packet to the two transmission paths between S1 and S4 respectively. Send detection packet ₁ and detection packet ₂ , detection packet ₁ includes routing factor ₁ , the first detection parameter and quintuple, and detection packet ₂ includes routing factor ₂ , the first detection parameter and quintuple. It should be noted that the routing factors in the probe packets sent to different transmission paths are different, that is to say, probe packet ₁ and probe packet ₂ can only be sent to different paths respectively, and cannot be sent to on the same path.

In practical applications, there may be multiple transmission paths between the first device and the second device, and detection packets may be sent to all transmission paths, or a part of the transmission paths may be selected to send detection packets. The function of the detection packet is to detect the load condition of each device on the passing transmission path, and one or more detection packets can be sent to a transmission path.

As shown in Figure 4, the specific process of the first device sending the detection packet _k to the N transmission paths between the first device and the second device according to the preset sending rule may include:

401: The first device constructs the detection packet _k according to a preset sending rule, which includes a routing factor _k , a first detection parameter, and a five-tuple;

FIG. 5 is a schematic diagram of the message structure of the detection packet _k . in:

The routing factor can be set to any value. For example, in FIG. 2, there are two transmission paths between S1 and S2, and the routing factors of the two transmission paths can be set to 1 and 2 respectively.

The first detection parameter may be port load information, or port memory margin, or port load information and port memory margin; the first detection parameter may be obtained from a port of the first device, or may be a preset fixed value .

Five-tuple, including source IP address, source port, destination IP address, destination port, and transport layer protocol number. For example: 192.168.0.110000TCP 121.14.88.7680 constitutes a quintuple, which means that a terminal with an IP address of 192.168.1.1 uses the TCP protocol through port 10000, and a terminal with an IP address of 121.14.88.76 and a port of 80 to connect. In this embodiment, the constructed detection packet may only include 5-tuples of large flow data packets.

402: Perform a hash operation on the routing factor _k and the quintuple in the detection packet _k to obtain the port address corresponding to the detection packet _k ;

Which transmission path the probe packet is sent to is determined by the routing factor and the five-tuple. Specifically, a hash operation is performed on the routing factor _k and the five-tuple in the detection packet _k to obtain the port address corresponding to the detection packet _k . After the port address corresponding to the detection packet is determined, the corresponding relationship between the routing factor and the transmission path is also determined. For probe packets with the same quintuple, one routing factor corresponds to only one transmission path, and one transmission path may correspond to multiple different routing factors.

403: Send the detection packet _k to the transmission path corresponding to the port address.

That is to say, different detection packets have established correspondences with different transmission paths, and different detection packets will detect their corresponding transmission paths.

The routing factor _k and the port address may be written into the forwarding header (not marked in FIG. 5 ) of the detection packet _k .

Optionally, the preset sending rule in 301 may be that the transmission path between the first device and the second device satisfies a preset overload condition, that is, only the transmission between the first device and the second device The first device sends a detection packet to the second device only when the path satisfies a preset overload condition.

The preset overload condition refers to a preset condition that a certain transmission path is about to be overloaded. Because congestion is easy to occur after each transmission path is overloaded, and the excessive data volume load on the transmission path is the fundamental condition that causes congestion, so the overload condition described in this embodiment can also be regarded as a condition that the network is about to be congested. According to the preset overload condition, the possible congestion can be judged, so that the network can avoid the congestion.

Further, in this embodiment, whether the transmission path in the network satisfies the preset overload condition can be judged by the following method:

(1), the first device judges whether the size of the data packet received per unit time is greater than a preset value;

If the size of the data packets received per unit time is greater than a preset value, it is determined that the transmission path selected by the data packets satisfies a preset overload condition.

For example, if the flow of data packets received by the first device is greater than a preset value, such as a flow value of 100M per second, it is considered that the data packets may cause an overload of the transmission path during transmission, so it is determined to transmit the data The path of the package satisfies the overload condition.

In this embodiment, the anti-overload path selection can be performed only for data packets with a large flow rate, so as to save network resources.

(2) Judging by the information sent by the equipment on the transmission path, including:

receiving a quantized congestion notification (quantized congestion notification, QCN) data packet on the transmission path;

Among them, the QCN data packet is used to report its own congestion situation to the device at the data sending end, so as to facilitate the device at the data sending end to adjust its own transmission capability, so as to achieve the effect of no packet loss.

If the first device receives the QCN data packet, it may indicate that the transmission path has been overloaded or is about to be overloaded, and then it is determined that the transmission path satisfies the overload condition.

The above-mentioned first method judges from the perspective of data packets, and the second method judges from the perspective of network equipment. In practical applications, one of the above judgment methods can be selected, or both methods can be used to judge at the same time. Satisfying one of them means meeting the preset overload condition, or satisfying both of them means meeting the preset overload condition.

Or, optionally, the preset detection rule may also be sent periodically. Then 301 specifically, the first device periodically sends the detection packet _k to the N transmission paths between the first device and the second device according to a preset sending period.

302: The first device respectively receives a probe response packet _k from the second device in response to the probe packet _k from the N transmission paths, and the probe response packet _k includes a routing factor _k and a second probe parameter and said quintuple;

The probe response packet _k mentioned here is returned by the second device through the original transmission path after the probe packet _k passes through the transmission path.

The first detection parameter and the second detection parameter may be port load information, or port memory remaining capacity, or port load information and port memory remaining capacity.

The second detection parameter may be the same as the first detection parameter, or the second detection parameter may be obtained by updating the first detection parameter by a device on the transmission path according to a preset update rule.

All devices between the first device and the second device on a transmission path, such as intermediate switches and core switches, can update the detection parameters in the detection packet _k . In this embodiment, the following update methods are preferred:

When the first detection parameter in the detection packet _k , such as port load information, is smaller than the local port load information in the device on the transmission path, update the port load information in the detection packet with the local port load information to obtain the first Two detection parameters. For example, if the pre-stored port load information is 60% of the rated load, and the current port load information of the switch is 80% of the rated load, then the port load information in the detection packet is updated to be 80% of the rated load.

If the local port load information of the device is smaller than the port load information stored in the detection packet _k , it will not be updated. In this way, the final port load information in the detection packet _k reflects the port load of the device with the heaviest load on this transmission path Condition.

It should be emphasized that during the transmission of the detection packet _k , only the first detection parameter is changed, while other parts are not changed. The detection packet _k passes from the first device through the current transmission path between the first device and the second device. After reaching the second device, the second device returns the updated probe packet _k , that is, the probe response packet _k , to the first device through the current path.

303: The first device selects a routing factor of the data packet to be sent according to the M probe response packets, and determines a transmission path of the data packet to be sent according to the routing factor of the data packet to be sent.

In this embodiment, as shown in Figure 2, S1 that initiates the probe will receive two probe response packets, for example, the probe response packet ₁ on the path S1-S2-S4 and the probe response packet on the path S1-S3-S4 Response packet ₂ . After receiving the probe response packet ₁ and the probe response packet ₂ , S1 compares the second probe parameters in the two probe response packets, for example, compares the port load information in the probe response packet ₁ and the probe response packet ₂ , Because the load of the transmission path passed by the probe response packet with the smallest port load information is the lightest, the probe response packet on the transmission path with the lightest load can be determined through comparison, and the selection in the probe response packet on the transmission path with the lightest load can be obtained. The routing factor is used as the routing factor of the data packet to be sent, and the selected routing factor is written into the data packet to be sent.

In this embodiment, writing the selected routing factor into the data packet to be sent can be realized by any one of the following two methods:

First, when the original data packet sent by the client is received, it is constructed as a data packet to be sent including the default routing factor, where the default routing factor can be any fixed value, for example, it can be 0; After the routing factor of the data packet to be sent is obtained, the default routing factor in the data packet to be sent may be updated to the routing factor of the data packet to be sent.

For example, in Figure 2, after comparing the second detection parameters in the probe response packet ₁ fed back by the transmission path S1-S2-S4 and the probe response packet ₂ fed back by S1-S3-S4, the load of the transmission path of the probe response packet ₂ is determined If it is light, the routing factor 2 in the probe response packet ₂ is selected as the routing factor of the data packet to be sent, and the routing factor 0 in the data packet to be sent is updated to 2.

Second, after the routing factor of the data to be sent is selected, a new data packet is constructed, that is, the routing factor is added to the original data packet sent by the client terminal, and a new data packet including the selected routing factor of the data packet to be sent is constructed. new packets.

After the first device writes the routing factor of the data packet to be sent into the data packet to be sent, it performs hash operation on the routing factor of the data packet to be sent and the quintuple of the data packet to be sent to obtain the data to be sent the sending port address of the packet, and determine the transmission path corresponding to the sending port address of the data packet to be sent as the transmission path of the data packet to be sent.

It should be emphasized that if the final transmission path determined by the method described in this embodiment is still the original transmission path indicated in the data packet sent by the client, this path can still be used for data packet transmission, or, The transmission of data can be suspended, and the network overload prompt information can be issued to the user.

The network data path selection method described in this embodiment, by sending a detection packet to the transmission path, detects the load of the transmission path, and finally selects a transmission path with a lighter load for data transmission, thereby realizing the load of each transmission path in the network. Balanced, further avoiding the blockage of fat tree network data transmission.

The present invention also discloses a transmission path selection device, as shown in Figure 6, comprising:

A detection packet sending unit 601, configured to send a detection packet _k to each of the N transmission paths between the first device and the second device according to a preset sending rule, and the detection packet _k includes a routing factor _k , the first A detection parameter and a quintuple, the routing factors in the detection packets of different transmission paths are different, wherein, N is a positive integer greater than 1, k is a positive integer and k=1, 2...M, M is a positive integer greater than or equal to N;

A probe packet receiving unit 602, configured to respectively receive a probe response packet _k from the second device in response to the probe packet _k from the N transmission paths, the probe response packet _k includes a routing factor _k , a second probe parameters and said quintuple;

The path selection unit 603 is configured to select a routing factor of the data packet to be sent according to the M probe response packets, and determine a transmission path of the data packet to be sent according to the routing factor of the data packet to be sent.

The transmission path selection device described in this embodiment can be set on an edge device in the network, such as an edge switch, or can be set on an intermediate device, such as an intermediate switch or a core server. In order to describe the path of the transmission path selection device in detail Select the function, here the edge device is preferred.

For example, in Figure 2, all edge devices can be provided with the transmission path selection device described in this embodiment. When PC1 transmits data to PC2, there are two optional paths, wherein S1 can be the first device, and S4 can be As the second device, after S1 receives the data packet sent by PC1, before S1 sends the data packet to the intermediate device S2, if it is judged that the transmission paths S1-S2-S4 and S1-S3-S4 meet the overload condition, Then construct detection packet ₁ and detection packet ₂ , detection packet ₁ includes routing factor ₁ , the first detection parameter and quintuple, detection packet ₂ includes routing factor ₂ , the first detection parameter and the quintuple, The routing factor _{1 in detection packet 1 is} 1, and the routing factor ₂ in detection packet ₂ is 2. Through hash calculation, the transmission path corresponding to detection packet ₁ is S1-S2-S4, and the transmission path corresponding to detection packet ₂ is For S1-S3-S4, send the detection packet ₁ and the detection packet ₂ to the corresponding transmission path respectively, and receive the detection response packet ₁ and the detection response packet ₂ of the second device S4 in response to the detection packet ₁ and the detection packet, and the detection response Packet ₁ includes the routing factor ₁ , the second probe parameter and the quintuple, and the probe response packet ₂ includes the routing factor ₂ , the second probe parameter and the quintuple S1 by comparing the probe response packet ₁ with the probe response packet ₂ , it is found that the load of the transmission path for transmitting the probe response packet ₁ is relatively light, then the routing factor 1 in the probe response packet ₁ is written into the data packet to be sent, so that S1 passes through the transmission path S1-S2-S4 send the data packets to be sent to PC2.

The device described in this embodiment, by sending different detection packets to different transmission paths, selects the transmission path with lighter load as the data transmission path, and realizes the load balancing of each path in the network, thereby avoiding network Congestion problem.

Further, as shown in FIG. 7, the detection packet sending unit described in this embodiment includes:

A detection packet construction subunit 701, configured to construct the detection packet _k according to preset sending rules;

The calculation subunit 702 is configured to perform a hash operation on the routing factor _k in the detection packet _k and the quintuple to obtain the port address corresponding to the detection packet _k ;

The sending subunit 703 is configured to send the detection packet _k to the transmission path corresponding to the port address.

Further, as shown in FIG. 8, the path selection unit in this embodiment includes:

A comparing subunit 801, configured to compare the second probe parameters in the M probe response packets to determine the probe response packet of the path with the lightest load;

The acquiring subunit 802 is configured to acquire the routing factor in the probe response packet of the path with the lightest load, and use the routing factor in the probe response packet of the path with the lightest load as the routing factor of the data packet to be sent ;

A writing subunit 803, configured to write the routing factor of the data packet to be sent into the data packet to be sent;

The determining subunit 804 is configured to perform a hash operation on the routing factor of the data packet to be sent and the quintuple to obtain the sending port address of the data packet to be sent, and determine the address corresponding to the sending port address The transmission path is the transmission path of the data packet to be sent.

Each embodiment in this specification is described in a progressive manner, each embodiment focuses on the difference from other embodiments, and the same or similar parts of each embodiment can be referred to each other.

The above description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention will not be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims (8)

1. A method for selecting a transmission path, comprising: The first device sends a detection packet _k to the N transmission paths between the first device and the second device according to the preset sending rules, and the detection packet _k includes the routing factor _k , the first detection parameter and five elements The routing factors in the detection packets of different transmission paths are different, and the routing factor _k and the quintuple are used to determine the transmission path of the detection packet _k from the N transmission paths, where N is a positive integer greater than 1, k is a positive integer and k=1, 2...M, M is a positive integer greater than or equal to N; The first device respectively receives a probe response packet _k from the second device in response to the probe packet _k from the N transmission paths, and the probe response packet _k includes a routing factor _k , a second probe parameter, and the said five-tuple; The first device obtains, according to the M probe response packets, the routing factor in the probe response packet on the transmission path with the lightest load, and uses it as the routing factor of the data packet to be sent, and according to the The routing factor of the data packet determines the transmission path of the data packet to be sent; The sending of the detection packet _k by the first device to the N transmission paths between the first device and the second device according to a preset sending rule specifically includes: The first device constructs the detection packet _k according to a preset sending rule; Performing a hash operation on the routing factor _k in the detection packet _k and the five-tuple to obtain the port address corresponding to the detection packet _k ; Send the detection packet _k to the transmission path corresponding to the port address. 2. The method according to claim 1, wherein the preset sending rules include: The transmission path between the first device and the second device satisfies a preset overload condition. 3. The method according to claim 2, wherein the preset overload condition comprises: The data packets received per unit time are greater than the preset value; or, receiving a quantized congestion notification data packet on the transmission path; or, The data packets received per unit time are greater than a preset value, and the quantized congestion notification data packets on the transmission path are received. 4. The method according to claim 1, wherein the first detection parameter and the second detection parameter are specifically port load information, or port memory margin, or port load information and port memory margin; The second detection parameter is the same as the first detection parameter, or the second detection parameter is obtained by updating the first detection parameter by a device on the transmission path according to a preset update rule. 5. The method according to claim 1, wherein the first device obtains the routing factor in the probe response packet on the transmission path with the lightest load according to the M probe response packets, and uses it as The routing factor of the data packet to be sent, including: comparing the second probe parameters in the M probe response packets to determine the probe response packets of the path with the lightest load; Obtaining the routing factor in the probe response packet of the path with the lightest load; The routing factor in the probe response packet of the path with the lightest load is used as the routing factor of the data packet to be sent. 6. The method according to any one of claims 1 to 5, wherein the determining the transmission path of the data packet to be sent according to the routing factor of the data packet to be sent comprises: Writing the routing factor of the data packet to be sent into the data packet to be sent; Performing a hash operation on the routing factor of the data packet to be sent and the quintuple to obtain the sending port address of the data packet to be sent; Determine the transmission path corresponding to the sending port address as the transmission path of the data packet to be sent. 7. A path selection device, characterized in that it comprises: A detection packet sending unit, configured to send a detection packet _k to the N transmission paths between the first device and the second device according to a preset sending rule, and the detection packet _k includes a routing factor _k , a first detection parameter and A five-tuple, the routing factors in the detection packets of different transmission paths are different, and the routing factor _k and the five-tuple are used to determine the transmission path of the detection packet _k from the N transmission paths, wherein , N is a positive integer greater than 1, k is a positive integer and k=1, 2...M, M is a positive integer greater than or equal to N; A detection packet receiving unit, configured to receive a detection response packet _k from the second device in response to the detection packet _k from the N transmission paths, and the detection response packet _k includes a routing factor _k , a second detection parameter and said quintuple; The path selection unit is configured to obtain the routing factor in the probe response packet on the transmission path with the lightest load according to the M probe response packets, use it as the routing factor of the data packet to be sent, and obtain the routing factor according to the pending A routing factor for sending data packets to determine the transmission path of the data packets to be sent; The detection packet sending unit includes: A detection packet construction subunit, configured to construct the detection packet _k according to preset sending rules; A calculation subunit, configured to perform a hash operation on the routing factor _k in the detection packet _k and the five-tuple to obtain the port address corresponding to the detection packet _k ; The sending subunit is configured to send the detection packet _k to the transmission path corresponding to the port address. 8. The device according to claim 7, wherein the path selection unit comprises: A comparing subunit, configured to compare the second probe parameters in the M probe response packets to determine the probe response packet of the path with the lightest load; The acquiring subunit is configured to acquire the routing factor in the probe response packet of the lightest-loaded path, and use the routing factor in the probe response packet of the lightest-loaded path as the routing factor of the data packet to be sent; A write subunit, configured to write the routing factor of the data packet to be sent into the data packet to be sent; A determining subunit, configured to perform a hash operation on the routing factor of the data packet to be sent and the quintuple to obtain the sending port address of the data packet to be sent, and determine the transmission port address corresponding to the sending port address The path is the transmission path of the data packet to be sent.