CN114050994B - Network telemetry method based on SRv6 - Google Patents
Network telemetry method based on SRv6 Download PDFInfo
- Publication number
- CN114050994B CN114050994B CN202111333786.5A CN202111333786A CN114050994B CN 114050994 B CN114050994 B CN 114050994B CN 202111333786 A CN202111333786 A CN 202111333786A CN 114050994 B CN114050994 B CN 114050994B
- Authority
- CN
- China
- Prior art keywords
- telemetry
- datagram
- node
- sid
- user
- 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
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L45/00—Routing or path finding of packets in data switching networks
- H04L45/12—Shortest path evaluation
- H04L45/123—Evaluation of link metrics
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L43/00—Arrangements for monitoring or testing data switching networks
- H04L43/08—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
- H04L43/0823—Errors, e.g. transmission errors
- H04L43/0829—Packet loss
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L43/00—Arrangements for monitoring or testing data switching networks
- H04L43/08—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
- H04L43/0852—Delays
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L43/00—Arrangements for monitoring or testing data switching networks
- H04L43/08—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
- H04L43/0876—Network utilisation, e.g. volume of load or congestion level
- H04L43/0888—Throughput
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L43/00—Arrangements for monitoring or testing data switching networks
- H04L43/08—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
- H04L43/0876—Network utilisation, e.g. volume of load or congestion level
- H04L43/0894—Packet rate
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L43/00—Arrangements for monitoring or testing data switching networks
- H04L43/10—Active monitoring, e.g. heartbeat, ping or trace-route
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L43/00—Arrangements for monitoring or testing data switching networks
- H04L43/10—Active monitoring, e.g. heartbeat, ping or trace-route
- H04L43/106—Active monitoring, e.g. heartbeat, ping or trace-route using time related information in packets, e.g. by adding timestamps
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L45/00—Routing or path finding of packets in data switching networks
- H04L45/12—Shortest path evaluation
- H04L45/124—Shortest path evaluation using a combination of metrics
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L45/00—Routing or path finding of packets in data switching networks
- H04L45/34—Source routing
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks
Abstract
The invention discloses a network telemetry method based on SRv6, which is characterized in that the traditional in-band network telemetry INT is added into SRv, the real-time information of each node and each link in a network view is collected and stored, the network telemetry is carried out, and the measurement comprises point telemetry, throughput rate packet loss rate in stream telemetry, average time delay in path telemetry, available bandwidth and the like. The invention divides the SID in SRH into three parts: SID header, SID data part, SID trailer, and the format of the SID will vary according to different metrics. In addition, the feasibility of the invention is verified through congestion control, time delay measurement and other cases.
Description
Technical Field
The invention belongs to the field of intersection of network security and network measurement, and particularly relates to a method for enabling SRv6 to perform in-band network tele-control.
Background
The Segment Routing (SR) is a source Routing technology, forms a network architecture facing path connection based on SDN concept, supports the multi-level programmable requirement of future network, and can meet the connection requirement in 5G oversized connection and slicing application scenes. SR-MPLS is an SR solution formed based on the current mainstream MPLS forwarding plane; SRv6 is an SR solution based on IPv6 extensions. SR-MPLS has evolved naturally along with MPLS forwarding mechanisms and has been widely used in transport networks. SRv6 further enhances the network programmability and supports network and service programmability. With the advent of new technologies such as SRV6 and the increasing size of users, networks exhibit "high rate, large-scale, multiple access, unpredictable" characteristics. Traditional network management and approaches have had difficulty addressing the challenges of existing and future networks. Therefore, network managers are urgent to subvert traditional network monitoring and fault removal methods, and propose real-time flexible measurement solutions capable of coping with scenario cases such as network state measurement, network failure detection, fault location and recovery, such as an in-band network telemetry scheme based on SRV 6.
In-band network telemetry is a framework for network data plane collection and reporting of network status without network control plane intervention. In an in-band network telemetry architecture, a switching device forwards data packets carrying telemetry instructions (Telemetry instructions). These telemetry instructions tell the network telemetry enabled network device what network status information should be collected and written as telemetry packets pass through the device.
In summary, the present invention uses the source routing technique of SRv6 and enables in-band network telemetry through modifications to the SID.
Disclosure of Invention
Aiming at the problems, the invention provides a network telemetry method based on SRv, which dynamically acquires real-time information of current resource views of equipment such as an exchanger and the like, wherein the real-time information comprises point telemetry, throughput rate packet loss rate in stream telemetry, average time delay in path telemetry, available bandwidth and the like. And uniformly storing the information in the monitoring server cluster for analysis and processing, and providing corresponding service for the user. The core problem of this invention is how to modify the SID so that it supports in-band network measurements.
In order to achieve the purpose of the invention, the technical scheme of the invention is as follows: an in-band network telemetry method based on SRv6 utilizes the relevant fields of SRv6 to obtain effective information such as load, hop count and other network links, nodes and the like, and calculates user interest network measures on a support SRv6 platform according to the information. The method comprises the following steps:
(1) Determining the dimension of the measurement by a user, informing a user interface, and sending the query dimension to an SRV6source router by the user interface, wherein the router sets protocol bits in the SID according to the query dimension (the protocol bits are different, and the formats of the SIDs are different);
(2) The datagram is transmitted in the network, and when the datagram reaches a router supporting SRv, the datagram is acquired and processed according to the requirements of a protocol; when reaching a router which does not support SRv6, the router directly forwards the datagram because the SRH cannot be processed;
(3) When the datagram arrives at the SRV6 sink router, the monitoring server cluster extracts all telemetry data in the datagram and restores the datagram to be delivered to the receiving end so as to realize transparency to the user;
(4) The SRV6 sink router can deliver the data measured in the previous steps to the user interface after certain treatment, and when the user interface receives a new datagram, the user can determine a route for the datagram or the flow according to the user requirement.
As an improvement of the invention, the specific method of the step (1) is as follows:
(1.1) determining the measure of the network telemetry by a user, wherein the measure comprises point telemetry, throughput rate packet loss rate in stream telemetry, average delay and available bandwidth of path telemetry and the like;
(1.2) the user issues the measure task to the user interface;
(1.3) the user interface adds an SRH part to each arrived datagram according to the task of the user, and sets a corresponding protocol bit, so that when the subsequent node receives the datagram, measurement information can be directly written according to the protocol bit;
(1.4) the user interface forwarding the processed datagram to the next hop according to the SR.
In this way, the method is not limited to single measure measurement, the measure is defined by a user, and higher controllability is achieved.
As an improvement of the invention, the design method for processing and writing telemetry information by the node in the step (2) is as follows:
(2.1) the network telemetry method of SRv uses key-value pairs to store data, the primary key having two forms, one node ID and one link ID; when the primary key is a node ID, the node attribute column family contains a time stamp reaching the node, a node utilization rate and the like; when the main key is a link identification sequence number, the link attribute column group comprises link utilization rate, link bandwidth and the like;
(2.2) based on the recognition that the approximation of telemetry data is generally sufficient for most applications, to reduce certain overhead, when a node writes a key pair into a datagram, the key pair is Huffman compressed using the approximation rather than the exact value, at the cost of certain error, by:
let the j-th node ID be p j The corresponding measured value is a character string s j The original binary string s j The 5 bits are divided into a group, and 26 lower case English letters are added with A-F from 0 to 31 respectively, and for the sake of simpler description of specific implementation of Huffman coding, 5 characters are taken as an example for description. Assume that five characters a, B, C, D, E are present and that the frequencies of occurrence, i.e., weights, are 5,4,3,2,1, respectively. The first step is to take two minimum weights as left and right subtrees to construct a new tree, namely 1,2 are taken to construct a new tree, the nodes are 1+2=3,
the broken line is newly generated node, then the newly generated node with the weight value of 3 is put into the rest set in the second step, so the set is changed into {5,4,3,3}, then the minimum two weight values are taken to form a new tree according to the second step, and the like, finally the complete Huffman tree is generated,
the codes corresponding to the characters can be obtained by replacing the corresponding characters with the weights as follows:
a- >11, B- >10, C- >00, D- >011, E- >010. The coding mode ensures that the codes of the characters with high occurrence frequency are shorter, and the codes of the characters with low occurrence frequency are longer, thereby realizing the compression effect of the binary string.
As an improvement of the invention, the specific method of the step (3) is as follows:
(3.1) the monitoring server cluster firstly extracts a protocol field in the SID so as to judge the measuring task at the time;
(3.2) if the protocol field is 011, it represents that the measurement task is congestion control, the data stored in SID is key value pair, link ID, link maximum utilization, server cluster extracts the maximum utilization according to format and stores it according to format, then provides congestion solution for network according to congestion control algorithm (such as HPCC, RED, etc.);
(3.3) if the protocol field is 010, the measurement task is the delay detection of the path telemetry, the data stored in the SID is a key value pair, the node ID, the time stamp of reaching the node, the server cluster extracts the time stamp of reaching each node according to the format and calculates the delay between the point to point and the transmitting end to the receiving end, and because the user does not care about the delay of any time point, only the average delay, the maximum delay and the minimum delay are stored in order to reduce the excessive redundancy cost during the storage.
By this step, different measurement methods for different measures are defined, and corresponding measures are taken while the method is implemented in order to save the same objective of overhead and improvement. Such as storing only the information of the average delay.
As an improvement of the invention, the specific method of the step (4) is as follows:
(4.1) transmitting the datagram according to a specified path, collecting a time stamp each time a hop is passed, link utilization and ejecting an address from the SID to continue transmission;
(4.2) after the datagram reaches the monitoring server, the monitoring server stores the measurement information of the path, and calculates time delay, bandwidth and the like according to the method;
(4.3) the monitoring server stores the measurement information of the path, the monitoring server calculates the path state table according to the measurement information after a plurality of times of measurement, and the user can set the selection metric for routing.
By this step, information of a plurality of nodes in the topology is collected, and compared with the conventional method, specific information of a specific node can be collected by the method due to controllability of Srv6 routing, and is used for subsequent analysis and function implementation.
Compared with the prior art, the invention has the following advantages: according to the scheme, a traditional in-band network telemetry INT is added into SRv through a network telemetry method based on SRv, real-time information of each node and each link in a network view is collected and stored, network telemetry is conducted, and the measures comprise point telemetry, throughput rate packet loss rate in stream telemetry, average time delay in path telemetry, available bandwidth and the like. When the method is realized, in order to save bit cost, the invention refers to a Huffman coding mode, and the data is compressed and then transmitted on the premise of not influencing experimental results, thereby achieving better effect of saving bit cost. In addition, due to the specificity of Srv6, specific information of a specific path can be measured using this scheme, thereby collecting necessary information for achieving routing.
Drawings
FIG. 1 is a SRV6 network telemetry architecture;
FIG. 2 is a datagram format for transmission in a network;
FIG. 3 is a default SID format;
FIG. 4 is an average delay SID of measurement path telemetry;
FIG. 5 is a SID for measuring link utilization;
FIG. 6 is an available bandwidth SID for measurement path telemetry;
FIG. 7 is a SID at the time of routing;
FIG. 8 is a SID when performing streaming telemetry;
FIG. 9 is a diagram of a new tree constructed using weights;
fig. 10 is a schematic diagram of a complete huffman tree.
The specific embodiment is as follows:
the present invention will be described in detail below with reference to the accompanying drawings in order to enhance understanding and appreciation of the invention.
Examples: referring to fig. 1-10, an in-band network telemetry method based on SRv uses SRv own relevant fields to obtain effective information such as load, hop count, etc. of network links, nodes, etc., and calculates a user interest network measure on a support SRv6 platform according to the information. The method comprises the following steps:
(1) Determining the dimension of the measurement by a user, informing a user interface, and sending the query dimension to an SRV6source router by the user interface, wherein the router sets protocol bits in the SID according to the query dimension (the protocol bits are different, and the formats of the SIDs are different);
(2) The datagram is transmitted in the network, and when the datagram reaches a router supporting SRv, the datagram is acquired and processed according to the requirements of a protocol; when reaching a router which does not support SRv6, the router directly forwards the datagram because the SRH cannot be processed;
(3) When the datagram arrives at the SRV6 sink router, the monitoring server cluster extracts all telemetry data in the datagram and restores the datagram to be delivered to the receiving end so as to realize transparency to the user;
(4) The SRV6 sink router can deliver the data measured in the previous steps to the user interface after certain treatment, and when the user interface receives a new datagram, the user can determine a route for the datagram or the flow according to the user requirement.
The specific method of the step (1) is as follows:
(1.1) determining the measure of the network telemetry by a user, wherein the measure comprises point telemetry, throughput rate packet loss rate in stream telemetry, average delay and available bandwidth of path telemetry and the like;
(1.2) the user issues the measure task to the user interface;
(1.3) the user interface adds an SRH part to each arrived datagram according to the task of the user, and sets a corresponding protocol bit, so that when the subsequent node receives the datagram, measurement information can be directly written according to the protocol bit;
(1.4) the user interface forwarding the processed datagram to the next hop according to the SR.
The design method for processing and writing telemetry information by the nodes in the step (2) is as follows:
(2.1) the network telemetry method of SRv uses key-value pairs to store data, the primary key having two forms, one node ID and one link ID; when the primary key is a node ID, the node attribute column family contains a time stamp reaching the node, a node utilization rate and the like; when the main key is a link identification sequence number, the link attribute column group comprises link utilization rate, link bandwidth and the like;
(2.2) based on the recognition that the approximation of telemetry data is generally sufficient for most applications, to reduce certain overhead, when a node writes a key pair into a datagram, the key pair is Huffman compressed using the approximation rather than the exact value, at the cost of certain error, by:
let the j-th node ID be p j The corresponding measured value is a character string s j The original binary string s j The 5 bits are divided into a group, and 26 lower case English letters are added with A-F from 0 to 31 respectively, and for the sake of simpler description of specific implementation of Huffman coding, 5 characters are taken as an example for description. Assume that five characters a, B, C, D, E are present and that the frequencies of occurrence, i.e., weights, are 5,4,3,2,1, respectively. The first step is to take two minimum weights as left and right subtrees to construct a new tree, namely 1,2 is taken to form a new tree, and the node is 1+2=3, as shown in fig. 9:
the broken line is the newly generated node, then the newly generated node with the weight value of 3 is put into the rest set in the second step, so the set is changed into {5,4,3,3}, then according to the second step, the minimum two weight values are taken to form a new tree, and the like, and finally the complete Huffman tree is generated, as shown in the following figure 10:
the codes corresponding to the characters can be obtained by replacing the corresponding characters with the weights in fig. 10 as follows: a- >11, B- >10, C- >00, D- >011, E- >010. The coding mode ensures that the codes of the characters with high occurrence frequency are shorter, and the codes of the characters with low occurrence frequency are longer, thereby realizing the compression effect of the binary string.
The specific method of the step (3) is as follows:
(3.1) the monitoring server cluster firstly extracts a protocol field in the SID so as to judge the measuring task at the time;
(3.2) if the protocol field is 011, it represents that the measurement task is congestion control, the data stored in SID is key value pair, link ID, link maximum utilization, server cluster extracts the maximum utilization according to format and stores it according to format, then provides congestion solution for network according to congestion control algorithm (such as HPCC, RED, etc.);
(3.3) if the protocol field is 010, the measurement task is the delay detection of the path telemetry, the data stored in the SID is a key value pair, the node ID, the time stamp of reaching the node, the server cluster extracts the time stamp of reaching each node according to the format and calculates the delay between the point to point and the transmitting end to the receiving end, and because the user does not care about the delay of any time point, only the average delay, the maximum delay and the minimum delay are stored in order to reduce the excessive redundancy cost during the storage.
The specific method of the step (4) is as follows:
(4.1) transmitting the datagram according to a specified path, collecting a time stamp each time a hop is passed, link utilization and ejecting an address from the SID to continue transmission;
(4.2) after the datagram reaches the monitoring server, the monitoring server stores the measurement information of the path, and calculates time delay, bandwidth and the like according to the method;
(4.3) the monitoring server stores the measurement information of the path, the monitoring server calculates the path state table according to the measurement information after a plurality of times of measurement, and the user can set the selection metric for routing.
The specific embodiment is as follows:
as shown in fig. 1, a schematic diagram of an SRV6 network telemetry architecture is shown.
The first router of the transmitting end is an SRV6source router and is mainly responsible for receiving a query task from a user interface, adding SRH to an arrived datagram according to the query task and setting a protocol field in the SID; the terminal router of the receiving end is an SRV6 sink router, the monitoring server cluster directly acquires the telemetry information from the SRV6 sink router and restores the datagram, thereby realizing transparency to the user; a part of routers which do not support SRv6 exist in the network, so that the SRH extension head cannot be processed, and the direct forwarding datagram has the common IPv6 forwarding capability; the user interface provides an interface for the user to accept telemetry tasks issued by the user and to inform the SRV6source router.
To better illustrate the steps of the SRv6 based network telemetry method, the format of datagrams transmitted in a visual network is first described. As shown in fig. 2, the datagram is composed of three parts. The first part is an IPv6 message header, the middle part represents a SRv specific SRH and is used for storing a 128bit IPv6 address format SID list; and finally, the part is a message load and transmits user information.
To describe the network telemetry method based on SRv in detail, the allocation of 128-bit SIDs when network telemetry is performed is described next. As shown in the default SID format of fig. 3, the SID is divided into three parts, a header, a trailer, and a data part. Wherein, in order to mark the beginning or ending of a SID, a specific 8-bit binary string 01111110 is used as the beginning and ending mark of the SID, in order to prevent the SID data part from existing 01111110, the data part is scanned, if 5 continuous 1 s appear, a 0 is directly added, and the opposite operation is carried out on the receiver, so as to realize transparent transmission; the protocol part is 3 bits in total, 8 functions are provided for the user, and the task for measuring at this time is congestion control as 011; the SID data portion stores telemetry data for the node in key-value pairs. When stored, the following sequence is followed: a node ID of 4 bits; telemetry data compressed to below 20 bits; an end symbol. The main function of the filling bit is to fill the SID to 128 bits, thereby realizing memory alignment; since there are multiple SIDs in one SRH, to facilitate the server cluster to directly intercept all SIDs, an EF flag bit is used for the flag—when the SID is the last SID of the SRH, its EF is 1, and the EF bits of the remaining SIDs are all 0.
The invention will now be described with reference to specific examples and figures of the invention. The invention comprises the following steps:
(1) Determining the dimension of the measurement by a user, informing a user interface, and sending the query dimension to an SRV6source router by the user interface, wherein the router sets protocol bits in SIDs according to the query dimension, and the formats of the SIDs are different if the protocol bits are different;
(2) The datagram is transmitted in the network, and when the datagram reaches a router supporting SRv, the datagram is acquired and processed according to the requirements of a protocol; when reaching a router which does not support SRv6, the router directly forwards the datagram because the SRH cannot be processed.
(3) When the datagram arrives at the SRV6 sink router, the monitoring server cluster extracts all telemetry data in the datagram and restores the datagram to be delivered to the receiving end so as to realize transparency to the user.
(4) The SRV6 sink router can deliver the data measured in the previous step to a user interface after certain processing. When a new datagram is received by the user interface, a route may be determined for that datagram or flow as desired by the user.
In order to describe in detail the working steps of an in-band network telemetry method based on Srv6, the different working modes of the method under different measures are specifically described next in 5 use cases.
Specific example 1:
if the telemetry task is the average delay in path telemetry, the protocol field is 010, and the workflow of the in-band network telemetry method based on Srv6 is as follows.
As shown in fig. 4 by the SID of the delay when the network is measured, the data portion in the SID is now clocked by the time the datagram arrives at the node, and the overhead of the time stamp is forced to be reduced to 16 bits. This is because 32 bits are typically required if one is to represent an exact time with a 2-ary system. At this point, a SID can only store 2 node timestamps, which is unacceptably high. In practical use cases, very accurate delay data may not be required. In contrast, excessive jitter and relatively unusual values are noticeable. Therefore, when storing the time stamp, the 32-bit time stamp data is compressed to 16 bits using the means of huffman coding described above.
When the average time delay in path telemetry is measured using the SRv6 based network telemetry method, the specific steps are as follows:
(1.1) the user issues the measure task to the user interface, which sets the source router so that it sets the protocol bits in all SIDs to 010;
(1.2) when the datagram j arrives at the node i, acquiring an arrival time stamp and the node ID, compressing the 32-bit time stamp to 16 bits according to huffman coding, and storing according to the format of fig. four;
(1.3) when datagram j arrives at sink router, the monitoring server cluster obtains all SIDs from datagram j and extracts the data part. Because the data portion contains time stamps and node IDs for reaching all nodes, the time delay between the point to point and between the receiving end to the transmitting end can be calculated at this time;
and (1.4) when a plurality of datagrams reach the sink router, the monitoring server cluster has enough path delay samples, and the monitoring server can calculate according to the path delay samples to acquire the average delay of the path telemetry.
Specific example 2:
if the telemetry task implements congestion control based on network telemetry of SRv, the protocol field is 011, and an in-band network telemetry method based on Srv6 works as follows.
First, when congestion control is performed using SRv6, instead of proposing a new congestion control method, SRv is used to collect data required for other congestion control methods, such as HPCC, RED, thereby reducing bit overhead. In order for SRv to support congestion control, only the highest utilization of the link, not per hop, needs to be maintained in the SID packet header. As shown in SID in congestion control in fig. five, the protocol bit is set to 011 as required, and when congestion control is performed using a network telemetry method based on SRv, the key value pair of SID data portion is (link ID, link highest utilization), the specific steps are as follows:
(2.1) the user issues the measure task to the user interface, which sets the source router so that it sets the protocol bits in all SIDs to 011;
(2.2) as datagram j passes through node i, the following calculation is performed in the packet:
wherein:
t represents base RTT, B represents link bandwidth, T, B are constants, T represents time occupation of a new packet, byte is packet size, qlen represents queue length when a data packet is dequeued.
The calculated link utilization U is written into the datagram in the format of fig. 5.
(2.3) when datagram j arrives at the sink router, the monitoring server cluster obtains all SIDs from it, extracting the data part. As described above, only the highest utilization rate of the link is needed to be saved, so that the data in the server is updated only when the new link utilization rate is greater than the corresponding utilization rate in the current server cluster;
specific example 3:
if the telemetry task is based on SRv measuring the available bandwidth for path telemetry, the protocol field is 101, and an in-band network telemetry method based on Srv6 works as follows.
First, when SRv is used to measure the available bandwidth for path telemetry, the method of active measurement is adopted because the complexity of passive measurement is too high and its accuracy is severely dependent on the performance of the packet capturer, and furthermore, passive measurement requires specific rights and cooperation, limiting the effectiveness of passive measurement to some extent. In this patent, a packet interval model (PGM) is adopted, the basic idea is to record the transmission time when a datagram is transmitted at the transmitting end, and to analyze the arrival time after flowing through the network under test with capacity C at the receiving end, so as to calculate the transmission time interval and the reception time interval. The available bandwidth can be calculated from the relationship between the two. The method belongs to a direct detection method in active measurement, and can reduce the influence on bandwidth caused by the conditions of route change and the like to a certain extent. The method comprises the following specific steps:
(3.1) the user issues the measure task to the user interface, which sets the source router so that it sets the protocol bits in all SIDs to 101;
(3.2) datagramsi, when passing through the source router, use t ij Representing the time of arrival of datagram j at router i, the ID of datagram and the time of transmission t are written in the SID of datagram 1i When the datagram i arrives at the sink router, the ID and arrival time t of the datagram are written in the SID of the datagram 2i ；
(3.3) the sender continues to send the datagram, and when datagram j passes through the source router, writes the ID of datagram and the sending time t in the SID of datagram 1j When the datagram j arrives at the sink router, the ID and arrival time t of the datagram are written into the SID of the datagram 2j ；
(3.4) calculating the transmission time interval delta of the transmitting end according to the n-th datagram when the n-th datagram passes through the sink router in And receiving end delta out Is a function of the reception time interval of (a):
(3.5) to this point, the calculation formula of the available bandwidth a is:
specific example 4:
if the telemetry task performs path selection based on SRv, the protocol field is 001, and an in-band network telemetry method based on Srv6 works as follows.
(4.1) the user interface setting the SID protocol field in the arriving datagram to 00l and assigning a path to this datagram according to the format of the following diagram, placing it in the first SID as shown in fig. 7;
| source address | First hop address | Second hop address | Third hop address | … | Last hop address | Destination address |
(4.2) transmitting the datagram i according to a designated path, collecting time stamps according to the format of fig. 7 each time a hop is passed, and using the link; after each jump, an address is sprung out from the SID so as to continue transmission;
(4.3) in the transmission process, when one node is reached, the information such as the sending time, the link utilization rate, the time stamp and the like is obtained by collecting or operation according to the method, and the information is stored in the SID according to the format of figure 7. In the storage process, if the overhead is excessive, the coding scheme is used for compression to reduce the bit overhead.
(4.4) after the datagram i reaches the monitoring server, the monitoring server stores the measurement information of the path, and calculates time delay, bandwidth and the like according to the method;
(4.5) repeating the operations of (4.1) to (4.4) but each time the user interface specifies a path for an arriving datagram to gather information for multiple links is random. Information for all links is not collected in actual use, for two reasons: 1. the number of possible paths from the transmitting end to the receiving end is excessive, which may include many redundant or redundant paths, and if each path is collected, the excessive overhead is unacceptable. 2. In actual path selection, it is not necessary to find a best path among all paths, but only to find a path which is reachable and relatively less expensive. So far, the monitoring server side can gather different tables according to the information collected for a plurality of times, and the table model is as follows:
list one
Wherein V is ij Representing the calculation of the link metrics from node i to node j, which may be latency, bandwidth, link utilization, different metrics are stored in different tables.
(4.5) the monitoring server side shares the measurement information matrix to the user interface for routing. When a certain datagram is accumulated, the user interface operates according to the following formula based on the measure matrix:
wherein, time ij Representing the delay from node i to node j, V ij Representing the link utilization from node i to node j. a and b are parameters which are self-determined by the user and can be used for reflecting higher utilization rate or less time which is more important for the user, and different parameters are set according to actual use conditions. The user interface generates a routing table therefrom as follows:
and the user interface selects the path with the minimum sigma cost as the optimal route according to the calculation result, and puts the path into the SID for transmission.
Specific example 5:
if the telemetry task performs packet loss rate and throughput rate measurement based on SRv, the protocol field is 100, and at this time, an in-band network telemetry method based on Srv6 works as follows.
The packet loss rate and the throughput rate are similar in measurement method, and the required cost is small, so that the measurement can be performed together.
(5.1) the user interface sets the SID protocol field in the arrived datagram to 100, which means that the measurement task is stream telemetry, and the measurement required to be measured is throughput rate and packet loss rate;
(5.2) the user interface writing information such as the transmission address, the destination address, the datagram ID, etc. into the SID for transmission according to the format of fig. 8;
(5.3) the user interface does the same for the next 1000 messages. And (3) selecting a proper route to a destination address according to the routing method in the step four, replying a user interface after the monitoring server cluster receives the message, and calculating the packet loss rate and throughput by the user interface according to the number of the received replied datagrams and the consumed time.
Experiment 1: simulation environment simulation test
The purpose of the experiment is as follows:
test SRv accuracy of measurements taken while operating in-band network telemetry
The experimental steps are as follows:
1. arranging an SRV6 simulation environment in the mini et;
2. performing point telemetry, stream telemetry, path telemetry according to the patent steps, and measuring throughput rate, packet loss rate, average time delay and available bandwidth in a network;
3. collecting and processing the measured telemetry values
Experiment 2: SRv6 in-band telemetry contrast
The purpose of the experiment is as follows:
test SRv efficiency comparison when running in-band network telemetry and when running INT alone
The experimental steps are as follows:
1. arranging an SRV6 simulation environment in the mini et;
2. performing point telemetry, stream telemetry, path telemetry according to the patent steps, measuring throughput rate, packet loss rate, average time delay and available bandwidth in the network and using overhead;
3. arranging the same network topology in a mini et;
4. point telemetry, stream telemetry, path telemetry, measuring throughput in the network, packet loss, average delay and available bandwidth and overhead;
5. the overhead and accuracy of the two measurement methods are compared.
It should be noted that the above-mentioned embodiments are only preferred embodiments of the present invention, and are not intended to limit the scope of the present invention, and the equivalent substitutions or alternatives made on the basis of the above-mentioned technical solutions are all included in the scope of the present invention.
Claims (3)
1. A SRv 6-based in-band network telemetry method, comprising the steps of:
step 1: determining the dimension of the measurement by a user, informing a user interface, and issuing the query dimension to an SRV6source router by the user interface, wherein the router sets a protocol bit in the SID according to the query dimension;
step 2: the datagram is transmitted in the network, and when the datagram reaches a router supporting SRv, the datagram is acquired and processed according to the requirements of a protocol; when reaching a router which does not support SRv6, the router directly forwards the datagram because the SRH cannot be processed;
step 3: when the datagram arrives at the SRV6 sink router, the monitoring server cluster extracts all telemetry data in the datagram and restores the datagram to be delivered to the receiving end so as to realize transparency to the user;
step 4: the SRV6 sink router processes the data measured in the previous steps and delivers the processed data to the user interface, when the user interface receives a new datagram, a route is determined for the datagram or the flow according to the user requirement;
the specific method of the step 1 is as follows:
1.1, a user decides the measure of the network telemetry, wherein the measure comprises point telemetry, throughput rate packet loss rate in stream telemetry, average delay of path telemetry and available bandwidth;
1.2, the user issues a measure task to a user interface;
1.3 the user interface adds SRH part to each arrived datagram according to the task of the user, and sets corresponding protocol bit, so that when the subsequent node receives the datagram, the measurement information is directly written according to the protocol bit; the default SID format shows that the SID is divided into a head part, a tail part and a data part, and the protocol part is 3 bits altogether, 8 functions are provided for a user, and 011 represents the task of the measurement at the time as congestion control; the SID data portion stores the telemetry data of the node in key-value pairs, which, when stored, are in the following order: a node ID of 4 bits; telemetry data compressed to below 20 bits; an ending symbol;
1.4 the user interface forwards the processed datagram to the next hop according to the SR;
the design method for processing and writing telemetry information by the node in the step 2 is as follows:
2.1: SRv6 the network telemetry method uses key value pair mode to store data, the main key has two modes, one is node ID and the other is link ID; when the primary key is a node ID, the node attribute column family contains a time stamp reaching the node and a node utilization rate; when the main key is a link ID, the link attribute column group contains the link utilization rate and the link bandwidth;
2.2 based on the insight that the approximation of telemetry data is sufficient for most applications, when a node reports a key pair to the data, the key pair is Huffman compressed, using the approximation instead of the exact value, at the cost of some error, by:
let the j-th node ID be p j The corresponding measured value is a character string s j The original binary string s j The 5 bits are divided into a group, and 26 lowercase English letters are added with A-F from 0 to 31 respectively.
2. The in-band network telemetry method of claim 1 based on SRv, wherein the specific method of step 3 is as follows:
3.1, the monitoring server cluster firstly extracts a protocol field in the SID so as to judge the measuring task at the time;
3.2 if the protocol field is 011, the measurement task is congestion control, the data stored in SID is key value pair, link ID, link maximum utilization, server cluster extracts the maximum utilization according to format and stores according to format, then provides congestion solution for network according to congestion control algorithm;
3.3 if the protocol field is 010, the measurement task is the delay detection of the path telemetry, the data stored in the SID is a key value pair, the node ID, the time stamp of reaching the node, the server cluster extracts the time stamp of reaching each node according to the format and calculates the delay between the point to point and the transmitting end to the receiving end, and because the user does not care about the delay of any time point, only the average delay, the maximum delay and the minimum delay are stored in order to reduce the excessive redundant expenditure during the storage.
3. The in-band network telemetry method of claim 1 based on SRv, wherein the specific method of step 4 is as follows:
4.1 datagram is transmitted according to the appointed route, each time a jump is passed, the time stamp is collected, the link utilization rate is increased, and an address is popped out from SID so as to continue transmission;
4.2, after the datagram reaches the monitoring server, the monitoring server stores the measurement information of the path, and calculates time delay and bandwidth according to the method;
and 4.3, the monitoring server stores the measurement information of the path, calculates a path state table according to the measurement information after a plurality of times of measurement, and sets a selection metric by a user for routing.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111333786.5A CN114050994B (en) | 2021-11-11 | 2021-11-11 | Network telemetry method based on SRv6 |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111333786.5A CN114050994B (en) | 2021-11-11 | 2021-11-11 | Network telemetry method based on SRv6 |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114050994A CN114050994A (en) | 2022-02-15 |
| CN114050994B true CN114050994B (en) | 2023-07-25 |
Family
ID=80208445
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111333786.5A Active CN114050994B (en) | 2021-11-11 | 2021-11-11 | Network telemetry method based on SRv6 |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114050994B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114978976B (en) * | 2022-04-14 | 2023-04-18 | 北京邮电大学 | Data anomaly detection method and device for SRv6 converged network |
| CN115442275B (en) * | 2022-07-27 | 2024-02-27 | 北京邮电大学 | Hybrid telemetry method and system based on hierarchical trusted streams |
| CN115766552A (en) * | 2022-11-04 | 2023-03-07 | 西安电子科技大学 | Network measurement method and device based on SRv6 and INT |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020001393A1 (en) * | 2018-06-25 | 2020-01-02 | 华为技术有限公司 | Method and network node for sending network performance parameter and calculating network performance |
| 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 |
| CN113067773A (en) * | 2021-03-16 | 2021-07-02 | 中国科学技术大学 | Method for fusing segment routing and in-band remote measurement based on protocol non-perception |
| CN113328915A (en) * | 2021-06-02 | 2021-08-31 | 北京工业大学 | Associated network measuring method based on SRv6 |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8601155B2 (en) * | 2006-08-16 | 2013-12-03 | Oracle America, Inc. | Telemetry stream performance analysis and optimization |
| US10212088B2 (en) * | 2016-11-07 | 2019-02-19 | Cisco Technology, Inc. | Tactical traffic engineering based on segment routing policies |
| US20190260657A1 (en) * | 2018-02-21 | 2019-08-22 | Cisco Technology, Inc. | In-band performance loss measurement in ipv6/srv6 software defined networks |
| CN112350934A (en) * | 2019-08-07 | 2021-02-09 | 中国电信股份有限公司 | Data transmission method, network node and data transmission system |
| CN111130842B (en) * | 2019-11-22 | 2022-04-15 | 东南大学 | Dynamic network map database construction method reflecting network multidimensional resources |
| CN112565017A (en) * | 2020-12-04 | 2021-03-26 | 盛科网络（苏州）有限公司 | Time delay measuring method and device based on SRv6 SID |
| CN112491926A (en) * | 2020-12-11 | 2021-03-12 | 迈普通信技术股份有限公司 | SRv6 path quality measuring method, device, electronic equipment and storage medium |
-
2021
- 2021-11-11 CN CN202111333786.5A patent/CN114050994B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020001393A1 (en) * | 2018-06-25 | 2020-01-02 | 华为技术有限公司 | Method and network node for sending network performance parameter and calculating network performance |
| 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 |
| CN113067773A (en) * | 2021-03-16 | 2021-07-02 | 中国科学技术大学 | Method for fusing segment routing and in-band remote measurement based on protocol non-perception |
| CN113328915A (en) * | 2021-06-02 | 2021-08-31 | 北京工业大学 | Associated network measuring method based on SRv6 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114050994A (en) | 2022-02-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114050994B (en) | Network telemetry method based on SRv6 | |
| US11159432B2 (en) | Data transmission method, and switch and network control system using the method | |
| US10320691B1 (en) | Visibility packets | |
| USRE48645E1 (en) | Exporting real time network traffic latency and buffer occupancy | |
| CN103338150B (en) | Communication network architecture method for building up, device, server and router | |
| KR100997182B1 (en) | Flow information restricting apparatus and method | |
| US11184284B2 (en) | Data packet forwarding method and apparatus | |
| CN105556916B (en) | The information statistical method and device of network flow | |
| US8923152B2 (en) | Random data stream sampling | |
| CN107295036B (en) | Data sending method and data merging equipment | |
| CN108512760A (en) | The method for routing of QoS of survice is ensured based on SDN | |
| US20170300595A1 (en) | Data packet extraction method and apparatus | |
| CN109391560A (en) | Notifying method, agent node and the computer equipment of network congestion | |
| CN101729391A (en) | Method, node and system for acquiring link aggregation group information | |
| CN116114233A (en) | Automatic flow management | |
| JP6055553B2 (en) | Queue scheduling method, apparatus and system | |
| CN114666681A (en) | Stateful in-band network telemetry method and system | |
| CN115225550B (en) | On-demand whole network telemetry device based on clustering path planning algorithm | |
| CN107426101B (en) | Quantum cluster fragment transmission method based on layering | |
| CN112291076A (en) | Packet loss positioning method, device and system and computer storage medium | |
| CN114157595B (en) | Communication system, data processing method and related equipment | |
| CN114884893A (en) | Forwarding and control definable cooperative traffic scheduling method and system | |
| CN116319468B (en) | Network telemetry method, device, switch, network, electronic equipment and medium | |
| CN116527559B (en) | Message forwarding processing and sending method, device, equipment and medium for network slice | |
| CN108632163A (en) | Transmission method, border router and the core router of data packet |
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 |