CN115118663B - Method for obtaining network congestion information by combining in-band network telemetry - Google Patents
Method for obtaining network congestion information by combining in-band network telemetry Download PDFInfo
- Publication number
- CN115118663B CN115118663B CN202210738290.4A CN202210738290A CN115118663B CN 115118663 B CN115118663 B CN 115118663B CN 202210738290 A CN202210738290 A CN 202210738290A CN 115118663 B CN115118663 B CN 115118663B
- Authority
- CN
- China
- Prior art keywords
- int
- data
- field
- header
- congestion
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 21
- 238000004891 communication Methods 0.000 claims abstract description 10
- 238000013461 design Methods 0.000 claims abstract description 4
- 230000005540 biological transmission Effects 0.000 claims description 20
- 230000007246 mechanism Effects 0.000 claims description 14
- HRULVFRXEOZUMJ-UHFFFAOYSA-K potassium;disodium;2-(4-chloro-2-methylphenoxy)propanoate;methyl-dioxido-oxo-$l^{5}-arsane Chemical compound [Na+].[Na+].[K+].C[As]([O-])([O-])=O.[O-]C(=O)C(C)OC1=CC=C(Cl)C=C1C HRULVFRXEOZUMJ-UHFFFAOYSA-K 0.000 claims description 12
- 238000003780 insertion Methods 0.000 claims description 9
- 230000037431 insertion Effects 0.000 claims description 9
- 238000012545 processing Methods 0.000 claims description 7
- 239000000284 extract Substances 0.000 claims description 5
- 230000000737 periodic effect Effects 0.000 claims 1
- 230000000875 corresponding effect Effects 0.000 description 12
- 238000005516 engineering process Methods 0.000 description 5
- 238000012544 monitoring process Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 230000006399 behavior Effects 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 235000008694 Humulus lupulus Nutrition 0.000 description 1
- 230000003542 behavioural effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012806 monitoring device Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000011664 signaling Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/10—Flow control; Congestion control
- H04L47/12—Avoiding congestion; Recovering from congestion
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/10—Flow control; Congestion control
- H04L47/19—Flow control; Congestion control at layers above the network layer
- H04L47/193—Flow control; Congestion control at layers above the network layer at the transport layer, e.g. TCP related
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
In the method for acquiring network congestion information by combining in-band network telemetry, in a TCP/IP protocol-oriented network, a transmitting end transmits data to a receiving end for communication, if congestion is transmitted in the network when passing through a certain node, INT layer data is added to the node where congestion occurs on the basis of congestion sensing by using a traditional ECN method, and the data acquisition work of the node is completed by the INT layer data; and if congestion is found again, only the data of the congestion node is required to be added after the INT layer data until all the congestion exchanger data are collected and forwarded to the receiving end, at the moment, the receiving end strips the header to extract the data and generate corresponding message data to be sent to the receiving end, the communication is finished, and the TCP connection is disconnected. The invention combines the control plane to design the network topology, monitors the network congestion in real time, can more comprehensively analyze the running state of the current TCP network, and accurately positions the congestion and judges the congestion condition.
Description
Technical Field
The invention belongs to the technical field of network TCP communication, and particularly relates to a method for acquiring network congestion information by combining in-band network telemetry.
Background
Passive network telemetry INT technology based on the P4 language also follows on the programmable data plane. INT technology is a framework for collecting and reporting network status by a network data plane proposed by Barefoot, arista, dell, intel and VMware together in 2015, and is a typical in-band measurement technology for solving the problem that the forwarding path and the forwarding delay are invisible. The technology does not need network control plane intervention, collects end-to-end real-time state information in a data path directly, and finally analyzes monitoring data through network management software deployed on a collector, extracts useful information and realizes monitoring of network performance. The basic process of INT telemetry is as follows: first, an INT traffic sender (application, terminal host protocol stack, management program, intelligent network card, sender top-of-rack switch) at the network edge embeds a specific data packet matching field called telemetry instruction into a normal data packet and sends the normal data packet to the network to be probed. Immediately as the data packet traverses the network along the designated path, the telemetry field tells the INT-enabled programmable device which network telemetry information needs to be collected and encapsulates the relevant information into normal data packets, carrying to the INT traffic sink, which is also at the network edge. And finally, extracting network telemetry data encapsulated in normal data packets at an INT flow receiving end, and reporting the network telemetry data to the INT monitoring equipment for processing.
In network transmission, a TCP transmission protocol becomes a practical transmission backbone frame of the internet at present due to its strong flexibility and reliable transmission capability, most of the networks are transmitted through TCP, and network problems such as misconfiguration, hardware faults, software errors and the like are necessarily and frequently occurred in a complex network environment such as reality. Among them, the network congestion problem is most common, and explicit congestion notification Explicit Congestion Notification (ECN), which is an extension of the TCP/IP protocol and supports end-to-end networks, is defined in RFC 3168, 2001. If end-to-end in the network successfully negotiates ECN, the ECN capable router may set a flag in the IP header when congestion occurs, signaling that congestion is imminent, rather than dropping packets directly. ECN reduces the number of lost packets of TCP, and by avoiding retransmissions, reduces delay (especially jitter), improving the performance of the application. However, the ECN congestion mechanism can only judge congestion occurrence, can not locate congestion nodes, and collects accurate congestion information to reflect congestion degree.
Disclosure of Invention
The invention aims to provide a method for acquiring network congestion information by combining in-band network telemetry, which is realized by adding INT in a data packet, and according to the execution mechanism of the specific network, the congestion place needs to count the INT metadata of equipment per se according to the indication of the INT header so as to complete congestion data acquisition. The system not only can judge whether congestion occurs, but also can position the congestion node and collect accurate congestion information to reflect the congestion degree.
The technical scheme adopted by the invention is as follows: a method of obtaining network congestion information in conjunction with in-band network telemetry,
step one, in a network facing to a TCP/IP protocol, TCP connection is established between hosts, and a transmitting end transmits data to a receiving end for communication;
step two, in the data transmission process, a plurality of switches are experienced, when data arrives at the switches, data packets are required to be analyzed for each node switch, TCP detects congestion through ECN, and if congestion occurs, a designed mechanism is applied to analyze and judge the packets, add the packets and modify the packet modules; otherwise, the data packet is processed normally. In the node switches passing through the route forwarding rule, the data collected by each congestion-generating switch only need to be added after INT Metadata is collected before;
and thirdly, after collecting the data of all the congestion nodes, when the data packet is forwarded to a receiving end of the system, the receiving end strips the header to extract Metadata to generate corresponding ACK data and sends the corresponding ACK data to a sending end, and after the transmission is finished, the established TCP connection is released.
The present invention is also characterized in that,
the second mechanism comprises the following modules:
s1, analyzing and judging a packet module, and carrying out IP address analysis, port analysis, protocol analysis and congestion judgment on a network interface layer, a network layer, a transmission layer and an application layer of a data packet of the node switch;
s2, adding a packet module, designing a protocol for adding INT, comprising: INT Shim header, INT header and INT Metadata specified by the InstructionnMASK field;
s3, modifying the packet module, wherein the protocol is used for modifying the header field of the upper layer of the data packet, calculating the data length of the INT Metadata of the node and modifying the checksum field.
The specific process of analyzing the data packet is as follows:
acquiring an Ethernet header of the data packet, and judging whether the Ethernet header is 0x800 identical to an IPv4 protocol number according to an EtherType field;
checking a network layer header of the data packet if the number is the same as the IPv4 protocol number, wherein ToS=1 or ToS=2 in IPv4, and the device supports ECN; if the queue size is greater than the threshold, congestion occurs, tos=3; judging whether the Protocol field is the same as the Protocol number of the TCP data packet or not and is equal to 0x06;
if the reserved field in the transport layer is 7, the protocol, the mechanism and the system designed by the invention are started if the reserved field is 7, and if the reserved field is 7, the reserved field is the same as the TCP protocol number and congestion is sent. First, an INT header space is allocated for the data packet, and second, an S2 add packet module is performed.
The invention increases the specific location of the packet module between the TCP layer and the load.
The add packet module mainly includes an INT shim header and an INT header:
the INTshim header includes: the Type field occupies 4 bits, and indicates that the data packet is of the INT data packet Type by setting corresponding content; NPT field, which occupies 2 bits, representing the next header field; the Length field takes 8 bits, representing the Length of the INT header and Metadata; the DSCP field, which takes 6 bits, carries the original DSCP value.
The INT header includes: ver field, which occupies 4 bits, version of the INT Metadata header, here version 2; a D field, which occupies 1 bit, representing discard; an E field which occupies 1 bit, and judges whether or not the remaining switches are used up (1 is used up, metadata insertion operation is ended, and 0 is continued operation); an M field which occupies 1 bit, determining whether a link Maximum Transmission Unit (MTU) has been reached (1 is to end Metadata insertion operation, 0 is to be operable continuously); a hopmacadatalen field, which takes 5 bits, representing the length of INT metadata that each node needs to add; a Count field, which occupies 8 bits and indicates that the number of INT layers is increased; an InstructionMask field, which occupies 16 bits, each bit of which represents an INT metadata type, which indicates the combination of individual INT metadata types that each node needs to count, e.g., switch identification number, ingress port timestamp, egress port timestamp, packet buffer queue length, etc.
Adding a packet module, firstly, inserting an INTShim header, and initializing a header information field; then insert INT header, also carry on the initialization setting to its containing field at the same time; finally, the total field of the IP layer is updated.
Because INT pad data overhead is low, the path length is typically no more than 5 hops inside the data center.
Modifying the upper header field of the data packet further comprises:
ECN field set 3 and DSCP field set 0x17 in network layer IPv 4; the flag entering the protocol reserves a field of 7 for the TCP header; the INT data packet Type of the INT Sheim header Type field is 1, namely an MD working mode; transmitting the DSCP value of the original IPv4 into the INT shimDSCP value; continuing the subsequent operation when confirming that the INT header E mark and the M mark are both 0, and ending Metadata insertion otherwise; then modify the Ver field, default INT Metadata header version 2; updating the IPv4 header length and the INTshim header length; and finally, calculating the checksum and updating the checksum field.
The length of INT Metadata data generated by congestion of each node is 4 XN bytes, and N is an integer greater than or equal to 1.
Step three, the receiving end strips the header to extract Metadata to generate corresponding ACK data and sends the corresponding ACK data to the sending end, specifically:
when the data packet is forwarded to the last switch of the system, whether the target IP address is the direct connection device of the network device is judged. The receiving end extracts all congestion information Metadata, generates corresponding ACK and transmits the ACK back to the sending end through TCP communication; and the rest data is the restored source data packet, and is submitted to the upper layer processing on the receiving end.
Specifically, the receiving end generates a TCP ACK, which is an IP header and a message that the TCP header carries congestion information Metadata, and the sending end determines that the TCP ACK method may be: when the URG/PSH/RST/SYN/FIN is set to 0 and the data length is the length of congestion information Metadata, the corresponding TCP ACK is determined.
The invention has the beneficial effects that the invention introduces INT aiming at the ECN limitation of the current TCP/IP protocol, realizes the congestion control optimization of the network by designing the protocol and a mechanism for expanding the data packet header by the INT, and can collect finer grained data so as to meet the accurate judgment of the network performance. The invention applies INT technology on ECN mechanism, and counts the INT Metadata of the device itself according to the indication of INT header at congestion node to complete data collection and forward back to the transmitting end, and has the advantages that:
(1) The remote measurement information of accurate data plane congestion on the path can be rapidly acquired, and the accuracy and timeliness are improved;
(2) INT telemetry data processing is optimized. The last hop sends the data packet to the receiving end instead of returning to the INT monitoring device, the receiving end extracts congestion information, encapsulates the information into an ACK, and returns the ACK to the sending end.
Drawings
FIG. 1 is an application scenario of the present invention;
FIG. 2 is a schematic diagram of an in-band network telemetry back-to-destination mode system;
FIG. 3 is a flow chart of a mechanism design;
FIG. 4 is a flow chart for parsing a data packet;
FIG. 5 is an add-on package design flow;
FIG. 6 is a data packet format;
FIG. 7 is a schematic diagram of an INT Shim header format provided by the present invention;
FIG. 8 is a schematic diagram of an INT header format provided by the present invention;
FIG. 9 is a flow for modifying upper header fields of a packet;
fig. 10 is a TCP ACK message structure.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, embodiments of the present invention will be described in further detail below with reference to the accompanying drawings.
In implementing the network system, python, mininet and P4 tools are used. These techniques may facilitate network prototyping. Python is attached with a Scapy library which simplifies the programming operation of network data packets, and in a program, the module can realize the sending, monitoring and analyzing of the network data packets. This module is lower than Nmap. Various scanning attack behaviors in the network can be intuitively known. In Scapy, each protocol is a class, and an INT protocol class is instantiated to create an INT packet. In this embodiment, a class of newly added INT packets is created, and INT layers include INT shim header, INT header, and Metadata.
Whereas Mininet provides a method of simulating a real network on a local computer by writing a simple Python script. The topology used in the present invention is shown in fig. 1, where there are 4 hosts and 4 switches in the application scenario, h1 and h11 are connected to s1, h2 and h22 are connected to s4, and reducing the bandwidth between s2 and s3 to 512kbps in the topology json file forms a bottleneck. Congestion is created when a low rate of traffic is sent from h1 to h2 and a high rate of iperf traffic is sent from h11 to h 22. Data plane processing logic for P4 switch uses P4 16 The syntax and semantic format of the standard definition is written. The data plane logic runs on top of the v1model architecture of the open source software switch model BMv (Behavioral Model version 2). BMv2 is a software switch with protocol independent forwarding architecture that can produce program-customized data forwarding behavior by running P4 programs.
Next, based on the above experimental environment, data transmission of the in-band network telemetry back-to-destination mode system is realized on the basis of TCP communication, as shown in fig. 2, the h1 host sends TCP data, through a plurality of switches, if the switches send congestion, an INT data layer is added, and finally INT Metadata information is extracted on the receiving end host h2 and fed back to the host h1 as ACK.
The method for acquiring network congestion information by combining in-band network telemetry of the invention, as shown in fig. 3, comprises the following steps:
step one, in a network facing to a TCP/IP protocol, TCP connection is established between hosts, and a transmitting end transmits data to a receiving end for communication;
step two, in the data transmission process, a plurality of switches are experienced, when data arrives at the switches, data packets are required to be analyzed for each node switch, TCP detects congestion through ECN, and if congestion occurs, a designed mechanism is applied to analyze and judge the packets, add the packets and modify the packet modules; otherwise, the data packet is processed normally. In the node switches passing through the route forwarding rule, the data collected by each congestion-generating switch only need to be added after INT Metadata is collected before;
the mechanism comprises the following modules:
s1, analyzing and judging a packet module, and carrying out IP address analysis, port analysis, protocol analysis and congestion judgment on a network interface layer, a network layer, a transmission layer and an application layer of a data packet of the node switch, wherein the analysis and judgment module is shown in FIG. 4;
analyzing the transmitted data packet, obtaining an Ethernet header of the data packet, and judging whether the Ethernet header is 0x800 identical to the IPv4 protocol number according to the EtherType field. Checking a network layer header of the data packet if the number is the same as the IPv4 protocol number, wherein ToS=1 or ToS=2 in IPv4, and the device supports ECN; if the queue size is greater than the threshold, congestion occurs, tos=3; judging whether the Protocol field is the same as the Protocol number of the TCP data packet or not and is equal to 0x06; if the reserved field in the transmission layer is 7, the protocol, the mechanism and the system designed by the invention are started if the reserved field in the transmission layer is 7. First, an INT header space is allocated for the data packet, and second, an S2 add packet module is performed.
S2, adding a packet module, designing a protocol for adding INT, comprising: INT Shim header, INT header and INT Metadata specified by the InstructionnMASK field;
as shown in fig. 5, first, insert an INTShim header and perform initialization setting of header information fields; then insert INT header, also carry on the initialization setting to its containing field at the same time; finally, the total field of the IP layer is updated.
The position of the add INT is between the TCP layer and the payload, and the packet format is shown in fig. 6.
The message format of the INT Shim packet is shown in FIG. 7, wherein:
(1) The 4-bit Type field represents the INT packet Type, definition 1 represents the INT-MD Type;
(2) The NPT field of 2-bit indicates the next header field;
(3) The 8-bit Length field indicates the Length of the INT header and Metadata;
(4) The 6-bit DSCP field carries the original DSCP value.
The message format of the INT packet is shown in fig. 8, in which:
(1) The Ver field of the 4-bit indicates the version of the INT Metadata header, here by default version 2;
(2) The D field of 1-bit indicates discarding, and the message is discarded after INT-MD metadata is extracted;
(3) The 1-bit E field is used to determine if the remaining switches are exhausted. 1 is used up, the metadata insertion operation is ended, and 0 is continued operation;
(4) The 1-bit M field is used to determine if the link Maximum Transmission Unit (MTU) has been reached. 1 is ending Metadata insertion operation, 0 is continuing operation;
(5) The HopMetaadataLen field of the 5-bit indicates the length of INT metadata that each node needs to add;
(6) The Count field of 8-bits indicates the number of INT layers to be added. Because INT padding overhead is low, count is typically no more than 5 inside the data center;
(7) An InstructionMask field of 16-bits, each bit representing an INT metadata type, indicates the combination of individual INT metadata types that each hop node needs to count, e.g., switch identification number, ingress port timestamp, egress port timestamp, packet buffer queue length, etc. The actual information collected depends on the support of the INT forwarding device for the collected INT metadata.
When the insertion of the INT header is completed, the acquisition of INT metadata is performed according to the header InstructionMask field.
Assume that the INT metadata collected are:
1) Switch identification code: switchID, record each switch ID, 4 bytes;
2) Node jump delay: hopLatency takes microsecond level as a unit and occupies 4 bytes;
3) Ingress port timestamp: ingressTimestamp takes microsecond level as a unit and occupies 4 bytes;
4) Output port timestamp: the egrastimestamp occupies 4 bytes in units of microseconds.
The length of INT Metadata data generated by congestion of each hop node is 4 XN bytes, and N is an integer greater than or equal to 1.
S3, modifying the packet module, wherein the protocol is used for modifying the header field of the upper layer of the data packet, calculating the data length of the INT Metadata of the node and modifying the checksum field.
The logic control stage, according to the key value key extracted by the header, look up table, match to finish the corresponding action, specifically include, as shown in fig. 9:
ECN field set 3 and DSCP field set 0x17 in network layer IPv 4; the flag entering the protocol reserves a field of 7 for the TCP header; the INT data packet Type of the INT Sheim header Type field is 1, namely an MD working mode; transmitting the DSCP value of the original IPv4 into the INT shimDSCP value; continuing the subsequent operation when confirming that the INT header E mark and the M mark are both 0, and ending Metadata insertion otherwise; then modify the Ver field, default INT Metadata header version 2; updating the IPv4 header length and the INTshim header length; and finally, calculating the checksum and updating the checksum field.
And thirdly, after collecting the data of all the congestion nodes, when the data packet is forwarded to a receiving end of the system, the receiving end strips the header to extract Metadata to generate corresponding ACK data and sends the corresponding ACK data to a sending end, and after the transmission is finished, the established TCP connection is released.
When the data packet is forwarded to the last switch of the system, whether the target IP address is the direct connection device of the network device is judged. The last exchanger sends data to a receiving end, a receiving end TCP layer sets a D field of an INT header to be in a discarding state, extracts all congestion information Metadata, generates a corresponding TCPACK at the receiving end, and sends the TCPACK back to a sending end through TCP communication; and the rest data is the restored source data packet and is transmitted to an upper application program for processing.
Specifically, the receiving end generates a TCP ACK, the message structure is shown in fig. 10, and the method for determining the TCP ACK by the transmitting end may be: when the URG/PSH/RST/SYN/FIN is set to 0 and the data length is the length of congestion information Metadata, the corresponding TCP ACK is determined.
While embodiments of the present invention have been illustrated and described above, it should be understood that the above embodiments are illustrative, and not to be construed as limiting the present invention, and all technical solutions falling under the spirit of the present invention are included in the scope of the present invention. It should be noted that variations, modifications, substitutions and variations can be made by persons skilled in the art to the above described embodiments within the scope of the invention.
Claims (7)
1. The method for acquiring the network congestion information by combining in-band network telemetry is characterized by comprising the following specific operation steps of:
step one, in a network facing to TCP/IP protocol, a transmission control protocol TCP connection is established between hosts, and a transmitting end transmits data to a receiving end for communication;
step two, in the data transmission process, a plurality of switches are experienced, when data arrives at the switches, data packets are required to be analyzed for each node switch, TCP detects congestion through explicit congestion notification ECN, and if congestion occurs, a designed mechanism is applied to analyze and judge the packets, add the packets and modify the packet modules; otherwise, normally processing the data packet, wherein in the node switches passing through the routing forwarding rule, the data collected by each switch with congestion only needs to be added to the data collected before the in-band network telemetry Metadata INT Metadata is collected;
the mechanism in the second step is specifically as follows:
the analysis judging packet module is used for carrying out IP address analysis, port analysis, protocol analysis and congestion judgment on the data packets of the node switch by a network interface layer, a network layer, a transmission layer and an application layer;
an add packet module, designed to add INT, comprising: INT Shim header, INT header and INT Metadata specified by the InstructionnMASK field;
the protocol is used for modifying the header field of the upper layer of the data packet, calculating the data length of the INT Metadata of the node and modifying the checksum field;
and thirdly, after collecting the data of all the congestion nodes, when the data packet is forwarded to a receiving end of the system, the receiving end strips the header to extract Metadata to generate corresponding acknowledgement character ACK data and sends the corresponding acknowledgement character ACK data to a sending end, and after the transmission is finished, the established TCP connection is released.
2. The method for obtaining network congestion information in combination with in-band network telemetry of claim 1, wherein the parsing of the data packets comprises:
acquiring an Ethernet header of the data packet, and judging whether the Ethernet header is 0x800 identical to an IPv4 protocol number according to an EtherType field;
checking a network layer header of the data packet if the number is the same as the IPv4 protocol number, wherein ToS=1 or ToS=2 in IPv4, and the device supports ECN; if the queue size is greater than the threshold, congestion occurs, tos=3; judging whether the Protocol field is the same as the Protocol number of the TCP data packet or not and is equal to 0x06;
checking if the reserved field in the transport layer is 7 if it is the same as the TCP protocol number and congestion is sent, wherein if it is 7, the mechanism of design is started; first, an INT header space is allocated for the data packet, and second, an S2 add packet module is performed.
3. The method for obtaining network congestion information in connection with in-band network telemetry of claim 1, wherein the location of the add-on packet module is between the TCP layer and the load.
4. The method of obtaining network congestion information in connection with in-band network telemetry of claim 1, wherein the INTShim header comprises:
the Type field occupies 4 bits, and indicates that the data packet is of an INT data packet Type by setting corresponding content;
NPT field, 2 bits, representing the next header field;
a Length field, 8 bits, representing the Length of the INT header and Metadata;
a DSCP field, occupying 6 bits, carrying the original DSCP value;
the INT header includes:
ver field, 4 bits, INT Metadata header version, version defaulting to 2;
a D field, which occupies 1 bit and indicates discarding;
e field, which takes 1 bit, judges whether the rest exchanger is used up, 1 is used up, ends Metadata insertion operation, and 0 is continuously operable;
m field, occupy 1 bit, judge whether it has reached the maximum transmission unit MTU of the periodic line, 1 is to finish Metadata to insert the operation, 0 is to continue the operation;
a hopmacadatalen field, which takes 5 bits, representing the length of INT metadata that each node needs to add;
a Count field, which occupies 8 bits and indicates that the number of INT layers is increased; an InstructionMask field, which occupies 16 bits, each bit of which represents an INT metadata type, indicates the combination of individual INT metadata types that each node needs to count, including switch identification number, ingress port timestamp, egress port timestamp, packet buffer queue length.
5. The method for obtaining network congestion information in connection with in-band network telemetry of claim 1, wherein modifying the upper header field of the data packet comprises: ECN field set 3 and DSCP field set 0x17 in network layer IPv 4; the flag entering the protocol reserves a field of 7 for the TCP header; the INT data packet Type of the INT Sheim header Type field is 1, namely an MD working mode; transmitting the DSCP value of the original IPv4 into the INT shimDSCP value; continuing the subsequent operation when confirming that the INT header E mark and the M mark are both 0, and ending Metadata insertion otherwise; then modify the Ver field, default INT Metadata header version 2; updating the IPv4 header length and the INTshim header length; and finally, calculating the checksum and updating the checksum field.
6. The method of claim 1, wherein the length of the INT Metadata data generated by congestion at each node is 4 x N bytes, and N is an integer greater than or equal to 1.
7. The method for obtaining network congestion information in combination with in-band network telemetry according to claim 1, wherein the step three is that the receiving end strips the header to extract Metadata to generate corresponding ACK data and sends the corresponding ACK data to the sending end, specifically:
when the data packet is forwarded to the last exchanger of the system, judging whether the target IP address is the direct-connected equipment of the network equipment or not; the receiving end extracts all congestion information Metadata, generates corresponding ACK and transmits the ACK back to the sending end through TCP communication; and the rest data is the restored source data packet, and is submitted to the upper layer processing on the receiving end.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210738290.4A CN115118663B (en) | 2022-06-27 | 2022-06-27 | Method for obtaining network congestion information by combining in-band network telemetry |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210738290.4A CN115118663B (en) | 2022-06-27 | 2022-06-27 | Method for obtaining network congestion information by combining in-band network telemetry |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115118663A CN115118663A (en) | 2022-09-27 |
| CN115118663B true CN115118663B (en) | 2023-11-07 |
Family
ID=83331526
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210738290.4A Active CN115118663B (en) | 2022-06-27 | 2022-06-27 | Method for obtaining network congestion information by combining in-band network telemetry |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115118663B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115883465B (en) * | 2022-12-01 | 2024-04-19 | 迈普通信技术股份有限公司 | Flow control method, device, server, system and storage medium |
| CN117176839B (en) * | 2023-10-26 | 2024-02-09 | 苏州元脑智能科技有限公司 | Remote measurement message transmission method, device, communication equipment and storage medium |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2005117131A (en) * | 2003-10-03 | 2005-04-28 | Nippon Telegr & Teleph Corp <Ntt> | Tcp traffic control method and control apparatus |
| EP2637371A1 (en) * | 2012-03-09 | 2013-09-11 | British Telecommunications Public Limited Company | Signalling congestion |
| CN108965151A (en) * | 2018-08-27 | 2018-12-07 | 华中科技大学 | A kind of Explicit Congestion control method based on queuing delay |
| CN112422498A (en) * | 2020-09-04 | 2021-02-26 | 网络通信与安全紫金山实验室 | In-band network remote measuring method, system and computer readable storage medium |
| CN113507420A (en) * | 2021-08-05 | 2021-10-15 | 清华大学 | Congestion control method and device |
| US11258719B1 (en) * | 2020-08-24 | 2022-02-22 | Keysight Technologies, Inc. | Methods, systems and computer readable media for network congestion control tuning |
| CN114124826A (en) * | 2021-11-28 | 2022-03-01 | 天津大学 | Low-delay data center network transmission system and method capable of sensing congestion position |
-
2022
- 2022-06-27 CN CN202210738290.4A patent/CN115118663B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2005117131A (en) * | 2003-10-03 | 2005-04-28 | Nippon Telegr & Teleph Corp <Ntt> | Tcp traffic control method and control apparatus |
| EP2637371A1 (en) * | 2012-03-09 | 2013-09-11 | British Telecommunications Public Limited Company | Signalling congestion |
| CN108965151A (en) * | 2018-08-27 | 2018-12-07 | 华中科技大学 | A kind of Explicit Congestion control method based on queuing delay |
| US11258719B1 (en) * | 2020-08-24 | 2022-02-22 | Keysight Technologies, Inc. | Methods, systems and computer readable media for network congestion control tuning |
| CN112422498A (en) * | 2020-09-04 | 2021-02-26 | 网络通信与安全紫金山实验室 | In-band network remote measuring method, system and computer readable storage medium |
| CN113507420A (en) * | 2021-08-05 | 2021-10-15 | 清华大学 | Congestion control method and device |
| CN114124826A (en) * | 2021-11-28 | 2022-03-01 | 天津大学 | Low-delay data center network transmission system and method capable of sensing congestion position |
Non-Patent Citations (2)
| Title |
|---|
| Best nexthop Load Balancing Algorithm with Inband network telemetry;Jiyoon Lim、等;《16th International conference on Network and service management（CNSM） 》;全文 * |
| 无线Ad hoc网络中MAC层拥塞测度的研究;廖秋丽;叶进;宋国青;李紫龙;;桂林电子科技大学学报(02);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115118663A (en) | 2022-09-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US7355971B2 (en) | Determining packet size in networking | |
| CN115118663B (en) | Method for obtaining network congestion information by combining in-band network telemetry | |
| Tan et al. | In-band network telemetry: A survey | |
| US20210092061A1 (en) | Data packet detection method, device, and system | |
| US11258719B1 (en) | Methods, systems and computer readable media for network congestion control tuning | |
| CN106027412B (en) | A kind of TCP jamming control method based on congestion queue length | |
| US7599283B1 (en) | Network traffic synchronization and data compression in redundant network topologies | |
| US7773530B2 (en) | Network traffic synchronization mechanism | |
| US8601155B2 (en) | Telemetry stream performance analysis and optimization | |
| EP3139560A1 (en) | Packet processing method, device and computer storage medium | |
| CN101080914A (en) | Method, system and article for improved network performance by avoiding data corruption | |
| CN101189840A (en) | Data unit relay device and method for controlling the same | |
| US8862768B2 (en) | Method and system for packetizing data for servicing traffic end-to-end | |
| CN112422421B (en) | Multi-path data packet transmission method of heterogeneous network | |
| CN114631290B (en) | Transmission of data packets | |
| US8705524B1 (en) | Systems and methods for embedding metadata in data packets | |
| US6850530B1 (en) | Methods and apparatus for providing and obtaining resource usage information | |
| CN114050994A (en) | SRv 6-based network telemetry method | |
| CN102907046B (en) | For the treatment of the method for message | |
| KR20220029142A (en) | Sdn controller server and method for analysing sdn based network traffic usage thereof | |
| CN116346634A (en) | State sensing information processing method and device of network management and control system and electronic equipment | |
| US11502986B2 (en) | Reducing transmission delay of transmitting data in Wi-Fi | |
| CN117014967A (en) | Mobile communication system, method and user plane node | |
| CN115277450A (en) | OPNET-based virtual-real combined heterogeneous communication network fusion system and application | |
| CN111917624B (en) | Method and system for transmitting control information in VXLAN transmission |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |