CN112887218B

CN112887218B - Message forwarding method and device

Info

Publication number: CN112887218B
Application number: CN202011527610.9A
Authority: CN
Inventors: 齐达; 陈勇
Original assignee: New H3C Technologies Co Ltd
Current assignee: New H3C Technologies Co Ltd
Priority date: 2020-12-22
Filing date: 2020-12-22
Publication date: 2022-10-21
Anticipated expiration: 2040-12-22
Also published as: CN112887218A

Abstract

This specification provides a method and an apparatus for forwarding a packet, where the method includes: when the network equipment is determined to be congested, determining the flow received by a port of the network equipment in an incoming direction and the flow received by a port of the network equipment in an outgoing direction; determining the proportion of the flow received in the incoming direction and the flow received in the outgoing direction; determining and adjusting the threshold value of the queue corresponding to the port according to the proportional value; and when the flow is received again, determining how to forward the message according to the threshold value. The present specification provides a method and an apparatus for dynamically adjusting a queue threshold, where a dynamically set threshold may balance a delay and throughput of a current network environment, so as to achieve an optimal threshold setting.

Description

Message forwarding method and device

Technical Field

The present disclosure relates to the field of communications technologies, and in particular, to a method and an apparatus for forwarding a packet.

Background

The design goal of Explicit Congestion Notification ECN (Explicit Congestion Notification) is defined in RFC (Request For Comments) 3168, and is to sense Congestion of an intermediate path and actively slow down the Transmission rate of TCP by the cooperation of a Transmission Control Protocol (TCP) transmitting end, a receiving end and an intermediate router, so as to avoid packet loss caused by Congestion from an early stage and realize maximum utilization of network performance. The problems that can be solved are as follows:

all TCP sending ends can sense the congestion of the middle path in early stage, actively slow down the sending rate and prevent the congestion from happening.

And on the queue forwarded by the intermediate switch, ECN marking is carried out on the TCP message exceeding the threshold of the queue, forwarding is continued, and the message is not discarded. The discarding of the message and the retransmission of the TCP are avoided.

Because packet loss is reduced, the TCP does not need to start message retransmission by a retransmission timer of several seconds or dozens of seconds, and the user experience of the time delay sensitive application is improved.

Compared with a network without the ECN function, the utilization rate of the network is better, and the situation that the network oscillates back and forth before overload and underload does not exist.

The marking of the ECN needs to be performed by the switch, but for different network congestion conditions, the switch only has a preset high-low threshold to perform marking control of the ECN at present. The upper and lower threshold values cannot be adjusted with the change of the network congestion condition, which may cause the difference of delay and throughput for different network congestion situations.

Disclosure of Invention

In order to overcome the problems in the related art, the present specification provides a method and an apparatus for forwarding a packet.

According to a first aspect of an embodiment of the present specification, a method for forwarding a packet is provided, where the method includes:

when the network equipment is determined to be congested, determining the traffic received by a port of the network equipment in an incoming direction and the traffic received by a port of the network equipment in an outgoing direction;

determining the proportion of the flow received in the incoming direction to the flow received in the outgoing direction;

determining and adjusting the threshold value of the queue corresponding to the port according to the proportional value;

and when the flow is received again, determining how to forward the message according to the threshold value.

Optionally, determining traffic received by a port of the network device in an ingress direction and traffic received by the port of the network device in an egress direction includes:

obtaining the quantity of messages which are received by a first port and carry a first preset value, wherein the first preset value represents that the messages are the messages received by the port in the incoming direction;

and obtaining the quantity of the messages which are received by the second port and carry a second preset value, wherein the second preset value represents that the messages are the messages received by the second port in the outgoing direction.

Optionally, the obtaining the number of the packets received by the first port and carrying the first preset value includes: obtaining the quantity of the messages which are received by the first port and carry the CWR bits as a first preset value;

the obtaining of the number of the packets received by the second port and carrying the second preset value includes: and obtaining the number of the messages with the ECE bit of a second preset value received by the second port.

Optionally, determining a ratio of the traffic received in the incoming direction to the traffic received in the outgoing direction includes:

and determining the ratio of the flow received in the incoming direction to the flow received in the outgoing direction according to the obtained number of the messages which are received by the first port and carry the CWR bit as the first preset value and the obtained number of the messages which are received by the second port and have the ECE bit as the second preset value.

Optionally, determining to adjust the threshold value of the queue corresponding to the port according to the ratio value includes:

and determining the threshold value of the queue corresponding to the port according to the preset value and the proportional value.

Optionally, determining to adjust the threshold of the queue corresponding to the port according to the ratio, specifically, adjusting the threshold corresponding to the queue to be smaller if the ratio is greater than a preset threshold, and adjusting the threshold corresponding to the queue to be larger if the ratio is less than or equal to the preset threshold.

And adjusting the threshold value according to the inverse proportion relation between the proportional value and the threshold value corresponding to the queue.

According to a second aspect of embodiments of the present specification, there is provided a packet forwarding apparatus, including: the device comprises a judging module, an obtaining module, a determining module, an adjusting module and a forwarding module;

the judging module is used for judging whether the network equipment is congested or not, and if so, the obtaining module obtains the flow received by the port of the network equipment in the inlet direction and the flow received by the port of the network equipment in the outlet direction;

the determining module is used for determining the proportion of the flow received in the incoming direction and the flow received in the outgoing direction;

the adjusting module is used for determining and adjusting the threshold value of the queue corresponding to the port according to the proportional value;

and the forwarding module is used for determining how to forward the message according to the threshold value when the flow is received again.

Optionally, the obtaining module is specifically configured to obtain the number of packets received by the first port and carrying a first preset value, where the first preset value indicates that the packet is a packet received by the port in the ingress direction; and obtaining the quantity of the messages which are received by the second port and carry a second preset value, wherein the second preset value represents that the messages are the messages received by the second port in the outgoing direction.

Optionally, the obtaining module is specifically configured to obtain the number of messages received by the first port and carrying the CWR bit with a first preset value; and obtaining the number of the messages with the ECE bit being a second preset value, which are received by the second port.

Optionally, the determining module is specifically configured to determine, according to the obtained number of the messages with the CWR bit as the first preset value received by the first port and the obtained number of the messages with the ECE bit as the second preset value received by the second port, a ratio between a flow received in an ingress direction and a flow received in an egress direction.

Optionally, the determining module is further configured to determine a threshold value of the queue corresponding to the port according to the preset value and the ratio value.

The technical scheme provided by the embodiment of the specification can have the following beneficial effects: the method provided by the disclosure can dynamically set the threshold value, and further can balance the time delay and the throughput of the current network environment. Specifically, when the ratio value is high, a low threshold value of the queue is set, and low time-delay performance of the queue is guaranteed. And when the proportion value is low, setting a high queue threshold to ensure the high throughput of the queue.

It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the specification.

Drawings

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with this specification and, together with the description, serve to explain the principles of the specification.

Fig. 1 is a schematic diagram of a network architecture provided by the present disclosure;

fig. 2 is a signaling interaction diagram of a message forwarding method provided by the present disclosure;

fig. 3 is a schematic structural diagram of a packet forwarding apparatus provided in the present disclosure.

Detailed Description

Reference will now be made in detail to the exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, like numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary examples do not represent all implementations consistent with this description. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the specification, as detailed in the claims that follow.

In the related art, reserved fields of 7 th and 8 th bits in a TOS (Type of Service) field of an IP header are redefined as an ECN field, where there are four values, which are described in RFC 3168:

(1) 00 represents that the message does not support ECN, so the message is processed as an original non-ECN message, that is: overload packet loss;

(2) The values of 01 and 10 are the same for the router, indicating that the message supports ECN functionality, and if congestion occurs, the two ECN fields are modified to 11 to indicate that the message is congested and continues to be forwarded by the switch. Therefore, the switch forwarding side needs to support the ECN, and the following new functions are required:

when congestion occurs, the original common non-ECN flow is carried out for a message with ECN = 00;

when congestion occurs, for messages with ECN =01 or ECN =10, the messages need to be modified to ECN =11, and the forwarding process is continued.

When congestion occurs, for the ECN =11 packet, forwarding needs to be continued.

In order to ensure fairness with non-ECN supported messages, when a queue exceeds a certain length, discarding of messages supporting ECN fields needs to be considered.

In the related art, the header modifies the reserved fields of bit8 and bit9 into CWR (Congestion Window Reduced) and ECE (ECN Echo) for modification on the host side. The design in RFC3168 is as follows:

1. the ECN =11 flag in the IP header is received at the TCP receiver and the ECE field is set to 1 when an ACK is replied. And sets the ECE field to 1 in each subsequent ACK.

2. When a TCP sending end receives an ACK message with the ECE field set to be 1, the sending rate of the TCP sending end needs to be halved, and when the next message is sent, the CWR field is set to be 1.

3. When the TCP receiving end receives the message with the CWR field set to 1, the subsequent ECE field is not set to 1. Until the IP header ECN =11 is received again, the above process is repeated.

4. The TCP sender narrows the send window when receiving an ECE =1, and will not narrow the send window again within the RTT (Round-Trip Time).

5. When the TCP receiver replies with an ACK to the sender, if the ACK is a "pure" ACK without data, the IP header ECN =00 must be made, since TCP has no mechanism to respond to a pure ACK and cannot send a congestion notification for a pure ACK.

6. For a host supporting ECN, the TCP layer needs to set the ECN in the IP header to 01 or 10 when sending a message.

Currently, congestion marking ECN is implemented on switch devices, which are used in conjunction with caching.

After the ECN function is started on the device, the congestion management function of the switch processes the packet as follows:

if the queue length is smaller than the lower limit, the message is not discarded, and the ECN domain is not identified and marked.

If the queue length is between the upper limit (HighLimit) and the lower limit (LowLimit), when the device calculates that a certain message needs to be discarded according to the discarding probability, the ECN field of the message is checked. If the ECN field shows that the message is sent by a terminal supporting the ECN, 2 bits are used to identify the ECN in a Differentiated Services Code Point (DSCP) field in the header of the IP message. These 2 bits are respectively: ECT (ECN Cable Transport) and CE (Congeison expert). The equipment marks the ECT bit and the CE bit of the message as 1 and then forwards the message; if the ECN field shows that the message transmission path has already experienced congestion (namely ECT and CE bit are both 1), the equipment directly forwards the message and does not re-mark the ECN domain; if both the ECT bit and the CE bit are 0, the device will discard the packet.

If the queue length exceeds the upper limit, the ECN domains of all the messages in the queue are marked as 11, and when the queue length reaches a queue tail discarding threshold, the messages are discarded.

For ECN marking, which is performed by a switch, and in order to solve a problem of a fixed threshold value of a queue in the related art, the method for adjusting the threshold value of the queue provided by the present disclosure is used, when a network traffic is an Incast model of N-to-one communication, in order to ensure that lossless service is forwarded without packet loss as much as possible, an apparatus dynamically adjusts an ECN threshold of a lossless queue according to an Incast multicast value of N-to-1 network traffic (the larger N indicates the larger the Incast value is, the larger the burst pressure generated to a cache is). The incast value in this disclosure is used to characterize the proportional value of the switch's traffic in the ingress direction and the egress direction.

A typical application of the Incast model of N-to-one communication is a network architecture similar to that shown in fig. 1, when a client sends a data request, and goes through intermediate processing of a switch, a plurality of links compete for an egress buffer of the same switch, which easily causes buffer overflow and generates packet loss. For example, as shown in fig. 1, when the traffic received by the switch is an ultra-large traffic with a specification of 100G, in the outgoing direction, that is, the direction to the receiving end, the outgoing interface of the switch may only send 25G traffic per interface, that is, 75G traffic in total, and the remaining 25G traffic is ready to be discarded. This problem is commonly referred to as Transmission Control Protocol (TCP) incast (incast).

The method comprises the following steps:

step 201, when it is determined that the network device is congested, determining traffic received by a port of the network device in an ingress direction and traffic received by a port of the network device in an egress direction.

In this embodiment, a network device is taken as an example for explanation, but it should be understood that the network device may also be a device for relaying packets, such as a router.

On the switch, access control list ACL matching can be periodically performed on the messages received by all the interfaces, and the traffic of the port of the network device in the incoming direction and the traffic of the port of the network device in the outgoing direction can be determined.

For the network device, it needs to determine which messages are in-direction messages and which messages are out-direction messages.

Take two ports a, B of the switch in fig. 3 as an example. The port A receives a message sent by a sending end, forwards the message to a receiving end through the port B, and receives a response message sent by the receiving end through the port B. That is, at this time, the switch needs to distinguish which traffic is received from the ingress direction and which traffic is received from the egress direction.

The incoming direction refers to the direction in which a message is received from a transmitting end, and the outgoing direction refers to the direction in which a message is transmitted to a receiving end.

An optional implementation manner of the switch for distinguishing the traffic in the ingress direction from the traffic in the egress direction is as follows: different identities are carried for traffic in different directions, so that it is possible to distinguish whether received traffic is received from an ingress direction or an egress direction by identifying the identity. For example, the first preset value may be carried in a message sent by the sending end, and the second preset value may be carried in a message sent by the receiving end.

Correspondingly, the switch can determine the quantity of the traffic received by the port of the network equipment in the incoming direction by obtaining the quantity of the messages which are received by the first port and carry the first preset value; the switch may determine the amount of traffic received in the outgoing direction by obtaining the amount of packets received by the second port and carrying the second preset value.

In the case of ECN support of the TCP/IP packet described above. In a more specific implementation manner, the switch may determine the number of the messages in the receiving direction of the switch through the message carrying the CWR bit of 1 sent by the sending end, and the switch may determine the number of the messages received by the switch in the sending direction through the message carrying the ECE bit of 1 sent by the receiving end.

Step 203, determining the ratio of the flow received in the incoming direction and the flow received in the outgoing direction.

Step 205, determining and adjusting the threshold value of the queue corresponding to the port according to the proportional value.

After determining the ratio between the traffic received in the ingress and egress direction and the traffic received in the egress direction (in this embodiment, the ratio value may also be referred to as an incast value), the determined incast value may be, for example, 1:1,2:1,4:1,8:1, etc., so that the threshold value of the queue in the network device can be adjusted according to the incast value.

And adjusting the threshold value of the queue according to the inverse relation between the proportional value and the threshold value corresponding to the queue.

Specifically, if the ratio is greater than a preset threshold, the threshold corresponding to the queue is decreased, and if the ratio is less than or equal to the preset threshold, the threshold corresponding to the queue is increased.

And when the Incast value is high, setting a low ECN threshold value to ensure low ductility of the queue. And when the Incast value is low, setting a high ECN threshold to ensure high throughput of the queue.

Step 207, when the traffic is received again, how to forward the message is determined according to the threshold value.

In the method provided by the present disclosure, a method for dynamically adjusting a queue threshold is provided, and a dynamically set threshold value can balance the delay and throughput of the current network environment, so as to achieve an optimal threshold setting.

In order to better explain the message forwarding method provided by the present disclosure, this embodiment explains a complete implementation process of the present disclosure by taking the signaling interaction diagram shown in fig. 3 as an example.

As shown in fig. 3, the method includes:

step 301, a sending terminal sends a message carrying an ECN identifier;

step 302, if the exchanger identifies that the current message is congested during forwarding, the exchanger changes the ECN mark of the message to 11 and sends the message to a receiving end;

step 303, the receiving end sets ECE to 1 when replying ACK (response message).

And step 304, the exchanger sends the ACK message to the sending end.

Step 305, when receiving the ACK with ECE set to 1, the sending end reduces the sending rate by half, and sends the CWR position of 1 to the switch when sending the next message;

step 306, the switch forwards the message with CWR 1 to the receiving end.

Step 307, the receiving end sets the ECE to 0 until receiving the message with the ECN identifier 11 again.

The method of the present disclosure may identify, when receiving the messages in step 303 and step 305, from which direction the message is received according to the corresponding identification bits.

And determining the threshold of the queue according to the obtained incast value, and further determining whether the message needs to be forwarded or discarded.

On the basis of the foregoing embodiment, the present disclosure further provides a message forwarding apparatus for executing the message forwarding method executed by the switch, and fig. 3 shows a schematic structural diagram of the message forwarding apparatus, and as shown in fig. 3, the apparatus includes: the system comprises a judging module 401, an obtaining module 402, a determining module 403, an adjusting module 404 and a forwarding module 405;

the determining module 401 is configured to determine whether the network device is congested, and if the network device is congested, the obtaining module 402 obtains traffic received by a port of the network device in an ingress direction and traffic received by a port of the network device in an egress direction;

the determining module 403 is configured to determine a ratio between the traffic received in the incoming direction and the traffic received in the outgoing direction;

the adjusting module 404 is configured to determine to adjust a threshold value of a queue corresponding to the port according to the ratio;

the forwarding module 405 is configured to determine how to forward the packet according to the threshold when receiving the traffic again.

Optionally, the obtaining module 402 is specifically configured to obtain the number of packets received by a first port and carrying a first preset value, where the first preset value indicates that the packet is a packet received by the port in the ingress direction; and obtaining the quantity of the messages which are received by the second port and carry a second preset value, wherein the second preset value represents that the messages are the messages received by the second port in the outgoing direction.

Optionally, the obtaining module 402 is specifically configured to obtain the number of the packets, which are received by the first port and carry the CWR bit as the first preset value; and obtaining the number of the messages with the ECE bit being a second preset value, which are received by the second port.

Optionally, the determining module 403 is specifically configured to determine, according to the obtained number of packets with CWR bits as a first preset value received by the first port and the obtained number of packets with ECE bits as a second preset value received by the second port, a ratio between a flow received in an ingress direction and a flow received in an egress direction.

The determining module is further configured to determine a threshold value of the queue corresponding to the port according to the preset value and the proportional value.

The apparatus provided in this embodiment is configured to execute the method performed by the switch in the foregoing embodiment, and specific detailed execution processes are not described herein again, and reference may be made to the embodiment of the method.

In the embodiments provided in the present application, it should be understood that the disclosed apparatus and method may be implemented in other manners. The apparatus embodiments described above are merely illustrative, and for example, the flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of apparatus, methods and computer program products according to embodiments of the present application. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems that perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.

In addition, functional modules in the embodiments of the present application may be integrated together to form an independent part, or each module may exist separately, or two or more modules may be integrated to form an independent part.

The functions, if implemented in the form of software functional modules and sold or used as a stand-alone product, may be stored in a readable storage medium. Based on such understanding, the technical solutions of the present application or portions thereof that contribute to the prior art may be embodied in the form of a software product, which is stored in a readable storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the methods described in the embodiments of the present application. And the aforementioned readable storage medium comprises: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.

It will be understood that the present description is not limited to the precise arrangements described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present description is limited only by the appended claims.

The above description is only a preferred embodiment of the present disclosure, and should not be taken as limiting the present disclosure, and any modifications, equivalents, improvements and the like made within the spirit and principle of the present disclosure should be included in the protection scope of the present disclosure.

The above description is only a preferred embodiment of the present disclosure, and should not be taken as limiting the present disclosure, and any modifications, equivalents, improvements, etc. made within the spirit and principle of the present disclosure should be included in the scope of the present disclosure.

Claims

1. A message forwarding method is applied to network equipment and is characterized by comprising the following steps:

determining the proportion of the flow received in the incoming direction and the flow received in the outgoing direction;

determining and adjusting the threshold value of the queue corresponding to the port according to the proportion;

when the flow is received again, determining how to forward the message according to the threshold value;

the step of determining and adjusting the threshold value of the queue corresponding to the port according to the proportion comprises the following steps:

if the ratio is larger than a preset threshold, the threshold corresponding to the queue is adjusted to be smaller, and if the ratio is smaller than or equal to the preset threshold, the threshold corresponding to the queue is adjusted to be larger.

2. The method of claim 1, wherein determining traffic received by a port of a network device in an ingress direction and traffic received in an egress direction comprises:

and acquiring the quantity of the messages which are received by the second port and carry a second preset value, wherein the second preset value represents that the messages are the messages received by the second port in the outgoing direction.

3. The method of claim 2,

the obtaining of the number of the packets received by the first port and carrying the first preset value includes: obtaining the quantity of the messages which are received by the first port and carry the CWR bits as a first preset value;

4. The method of claim 2 or 3, wherein determining the ratio of the traffic received in the ingress direction and the traffic received in the egress direction comprises:

5. The method of claim 1, wherein determining to adjust the threshold of the queue corresponding to the port according to the ratio comprises:

and determining the threshold value of the queue corresponding to the port according to the preset threshold value and the ratio.

6. A message forwarding apparatus, comprising: the device comprises a judging module, an obtaining module, a determining module, an adjusting module and a forwarding module;

the judging module is used for judging whether the network equipment is congested or not, and if so, the acquiring module acquires the flow received by a port of the network equipment in an incoming direction and the flow received by a port of the network equipment in an outgoing direction;

the adjusting module is used for determining and adjusting the threshold value of the queue corresponding to the port according to the proportion;

the forwarding module is used for determining how to forward the message according to the threshold value when the flow is received again;

when the ratio between the flow received in the incoming direction and the flow received in the outgoing direction is determined, the adjusting module is specifically configured to:

7. The apparatus according to claim 6, wherein the obtaining module is specifically configured to obtain a number of packets received by a first port, where the number of packets carries a first preset value, where the first preset value indicates that the packet is a packet received by the port in an incoming direction; and acquiring the quantity of the messages which are received by the second port and carry a second preset value, wherein the second preset value represents that the messages are the messages received by the second port in the outgoing direction.

8. The apparatus of claim 7, wherein the obtaining module is specifically configured to obtain a number of packets received by the first port, where the number of packets carries a CWR bit with a first preset value; and obtaining the number of the messages with the ECE bit of a second preset value received by the second port.

9. The apparatus according to claim 7 or 8, wherein the determining module is specifically configured to determine a ratio between a traffic received in an ingress direction and a traffic received in an egress direction according to the obtained number of the packets received by the first port and carrying the CWR bit with the first preset value and the obtained number of the packets received by the second port and having the ECE bit with the second preset value.

10. The apparatus of claim 6, wherein when determining to adjust the threshold of the queue corresponding to the port according to the ratio, the adjusting module is specifically configured to: