CN111490934A - Multi-path routing system based on stream burstiness - Google Patents
Multi-path routing system based on stream burstiness Download PDFInfo
- Publication number
- CN111490934A CN111490934A CN202010332340.XA CN202010332340A CN111490934A CN 111490934 A CN111490934 A CN 111490934A CN 202010332340 A CN202010332340 A CN 202010332340A CN 111490934 A CN111490934 A CN 111490934A
- Authority
- CN
- China
- Prior art keywords
- forwarding
- label
- data message
- flowet
- path
- 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.)
- Granted
Links
Images
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
- H04L45/24—Multipath
-
- 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
- H04L45/14—Routing performance; Theoretical aspects
-
- 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
- H04L45/50—Routing or path finding of packets in data switching networks using label swapping, e.g. multi-protocol label switch [MPLS]
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention discloses a multi-path routing system based on flow burstiness, which comprises a gateway node and a forwarding node, wherein when a new Flowlet data message reaches a gateway, the gateway node searches for an unused label for the new Flowlet data message, records the use of the label, firstly sends a control signaling message to establish a label instruction table for a subsequent forwarding node, and finally adds the label to the data message to convert the data message into a message carrying the label for forwarding, and the forwarding node firstly receives the control signaling message, reads the content of the control signaling message, adds the content into the label instruction table, and then directly searches for the label to forward the message carrying the label when receiving the message carrying the label.
Description
Technical Field
The invention belongs to the technical field of multipath transmission, and particularly relates to a multipath routing system based on stream burstiness.
Background
Today, Internet routing protocols such as Open Short Path First (OSPF) and Border Gateway Protocol (BGP) are mainly designed to provide end-to-end "best effort" services. These protocols implement the concept of "best effort" by providing a single path to the destination subnet using distributed autonomous protocols. The huge success of the internet causes the number of various internet applications to increase explosively, the various internet applications obviously do not meet the traditional best effort service, and new requirements are provided for indexes such as link bandwidth, end-to-end minimum time delay and high-reliability links; the huge traffic generated by massive internet applications also enables operators to continuously deploy new network equipment and lay higher-speed bottom links, and the traditional routing protocol cannot fully utilize the network structure with larger and larger scale, and is difficult to provide flexible transmission service for upper-layer services. Therefore, new network technologies continue to emerge, where multipath transmission provides a new transmission technology.
The multi-network card interface of the routing device is realized, and the Internet topology naturally has the characteristic of multi-path in a multi-host deployment mode of an autonomous system ISP (Internet service Provider). The multi-path transmission breaks through the original single-path thought, and is fundamentally more effective than the current single-path transmission. Just as packet switching is more efficient than circuit switching because it takes full advantage of the spatial and temporal multiplexing capabilities on a single link, the multiplexing gain effect of the network's temporal and spatial dimensions is reflected in multipath transmissions. Besides fully utilizing the bottom transmission capability of the network, the multi-path transmission can also be used as a technical means for realizing the network functions of bandwidth aggregation, minimizing end-to-end time delay, improving robustness, increasing stability, realizing load balance, safely dispersing routing and the like.
At present, a plurality of multipath source routing technologies are researched in a wireless mobile network, and a plurality of multipath routing protocols, routing algorithms and system frameworks which are suitable for the wireless mobile network are also provided. The wireless network routing protocol is developed from the initial dynamic source routing to the multi-path source routing, and a routing algorithm aiming at the problems of minimum resource consumption, end-to-end time delay, data safety and the like under multi-paths is continuously improved.
And the Internet, have evolved over the years, with many network devices and protocols already running on the core system, and it is obviously a costly project to fully adapt these devices and protocols to multipath routing. An existing MultiPath routing system (for example, an MPTCP (MultiPath TCP) protocol and an equivalent ECMP (Equal Cost MultiPath) protocol) is difficult to be compatible with an existing protocol running in the Internet at present, flexibility is low, and the existing MultiPath routing system is difficult to adapt to a large-scale and variable network environment, and deployment of the system is difficult.
The multi-path routing also causes disorder of messages of a single flow at an aggregation point, and when the time delay between paths is different, the later sent messages may arrive in advance, which causes disorder of the messages, thereby causing misjudgment of a transmission layer protocol (TCP) as network congestion and finally causing the performance reduction of the transmission layer protocol. The transmission performance of the system is also a problem to be solved by the multi-path routing system.
It can be seen that although multipath has many advantages in reducing delay, improving robustness, and increasing system stability, the existing multipath transmission system cannot be fully compatible with the conventional system, cannot meet the increasing network scale, and cannot provide sufficient flexibility in deployment and function.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provide a multi-path routing system based on stream burstiness, which realizes multi-path routing by setting a label in a TT L field by utilizing the characteristics of network traffic burstiness and improves the routing performance while being compatible with the prior protocol.
In order to achieve the above object, the multi-path routing system based on stream burstiness of the present invention includes a gateway node and a forwarding node, wherein:
the gateway node is configured to receive a data packet sent by a user and perform Flowlet division, where a time interval adopted during division is greater than a maximum delay difference of all paths between source and destination addresses of the data packet, and then process the data packet, that is, determine a forwarding policy for the data packet based on Flowlet, perform path calculation, and forward the data packet after distributing a label, and specifically includes:
the gateway node maintains a Flowet state table, a forwarding policy table and a used label table, wherein the Flowet state table is used for storing matching items of various Flowets, corresponding labels, paths and duration of the Flowets; the forwarding policy table is used for storing forwarding policies corresponding to the Flowlets, and the specific forwarding policies can be set in advance according to needs; the used label table is used for storing the labels which are used in the current system and the corresponding failure time;
the gateway node processing the data message comprises the following steps:
s1: judging whether the Flowet in which the data message is positioned exists in a Flowet state table or not, namely extracting a matching item from the data message, searching whether the Flowet exists or not in the Flowet state table, if not, entering a step S2, otherwise, entering a step S6;
s2: searching a forwarding strategy table for a forwarding strategy corresponding to the Flowet where the data message is located, namely extracting a matching item of the data message, matching in the forwarding strategy table, if the corresponding forwarding strategy does not exist, entering step S33, otherwise, entering step S4;
s3, setting the highest 2 bits in TT L field of current data message as '00', namely marking as traditional route data message, and then forwarding;
s4: extracting a forwarding strategy of the Flowet where the data message is located from the forwarding strategy table, and calculating N parallel forwarding paths according to the forwarding strategy, wherein N is set according to actual needs;
s5: selecting one label from unused labels as the label of the Flowet where the current data message is located, setting the failure time of the Flowet according to the requirement, recording the label and the failure time into a used label table, and recording the Flowet matching item, the corresponding label, the path and the duration of the Flowet into a Flowet state table;
s6: judging whether the Flowlet control information is invalid, if so, entering a step S7, otherwise, entering a step S9;
s7: extracting a label and N forwarding paths of the Flowet where the current data message is located from the Flowet state table, entering step S8, and simultaneously pushing the failure time corresponding to the Flowet label back to a Flowet interval by using the label table;
s8, generating corresponding control signaling messages for each forwarding path respectively, wherein the source address and the destination address of the control signaling messages are the same as the current data messages, the contents of the control signaling messages are the labels of the Flowlets where the data messages are located, the corresponding failure time and the forwarding paths, the highest 2 bits of the TT L field in the header are set as '01', and then sending the control information messages;
s9, setting the highest 1 bit of TT L field in the header of the data message as '1', and the 2 nd bit to the 4 th bit as the label of the Flowet where the data message is located, converting the original data message into a message carrying the label to send;
after receiving a data message, the forwarding node firstly extracts the highest 1 bit in the TT L field, if the data message is 1, the data message is a message carrying the label, if not, further extracts the 2 nd bit in the TT L field, if the data message is 0, the forwarding node is a traditional routing data message, otherwise, the forwarding node is a control signaling message, and the processing modes of the three data messages are respectively as follows:
for the traditional routing data message, the forwarding node forwards the data message according to a traditional routing method.
For the control signaling message, the forwarding node extracts the source and destination addresses, the labels, the instructions and the failure time of the Flowlet, records the source and destination addresses, the labels, the instructions and the failure time into a label instruction table, and then forwards the control signaling message to the next node according to the path information in the control signaling message.
For the message carrying the label, the forwarding node firstly judges whether the forwarding node is a destination node, if so, the forwarding node extracts data and sends the data to an upper layer application, if not, the forwarding node extracts a source destination address and a label of the data message, inquires a label instruction table to obtain a corresponding instruction for operation, forwards the instruction to the next node, and meanwhile, pushes the failure time corresponding to the label of the Flowlet back to a Flowlet interval in the label instruction table.
The invention relates to a multi-path routing system based on flow burstiness, which comprises a gateway node and a forwarding node, wherein when a new Flowlet data message reaches a gateway, the gateway node searches for an unused label for the new Flowlet data message, records the use of the label, firstly sends a control signaling message to establish a label instruction list for a subsequent forwarding node, and finally converts the data message into a message carrying the label for forwarding; the forwarding node receives the control signaling message, reads the content in the control signaling message, adds the content into the label instruction list, and directly searches for a label to forward the control signaling message when receiving the label-carrying message subsequently.
The invention has the following beneficial effects:
1. on the premise of the existing IP network structure, the format of the message to the forwarding specification obeys the specification of the existing network protocol, and only the bits of the TT L field part of the IP message are endowed with new meanings, so that the method can be directly deployed in the existing network structure without modification;
2. the invention solves the problem of message disorder in a multi-path routing system: the multi-path routing system uses the Flowet as a basic transmission unit, the maximum time delay difference among different paths is calculated by a gateway node, and then the Flowet interval can be used for counteracting the path time delay difference, so that the problem of message disorder of multi-path transmission is solved;
3. the invention simplifies the structure of the forwarding node, and improves the speed of searching the message forwarding path by using the label;
4. the routing calculation of the invention is more flexible: the gateway node may match different forwarding paths for different traffic, thereby selecting different paths using different algorithms.
Drawings
FIG. 1 is a diagram illustrating the setting of the TT L field in the present invention;
FIG. 2 is a block diagram of an embodiment of a multi-path routing system based on flow burstiness according to the present invention;
FIG. 3 is a flow chart of the processing of a data packet by a gateway node in the present invention;
FIG. 4 is a block diagram of a gateway node in the present invention;
fig. 5 is a schematic diagram of the hybrid deployment of the multi-path routing system and the conventional uncontrolled network.
Detailed Description
The following description of the embodiments of the present invention is provided in order to better understand the present invention for those skilled in the art with reference to the accompanying drawings. It is to be expressly noted that in the following description, a detailed description of known functions and designs will be omitted when it may obscure the subject matter of the present invention.
Examples
To better explain the technical solution of the present invention, first, the technology adopted in the present invention will be briefly described.
● Flowlet technique
Flowet (micro Flow) is a kind of transmission unit between message and Flow (Flow). Batch processing techniques are commonly used throughout the network system of application, middleware, operating system, transport and link layers, and batch processing can result in short-term bursts of datagrams. Therefore, network traffic has a sudden "train-like behavior" at a microscopic time, and is not a uniform continuous message. Dividing the messages with the time interval between the continuous messages exceeding a certain interval threshold value into different groups, wherein the messages of the same group form a group of burst messages, and the group of burst messages is called a Flowlet.
The maximum time delay difference of a plurality of paths is △ t, and the time interval between two Flows is t, the time difference of the Flow sent by a sending end to a receiving end is t- △ t or y + △ t, and the Flow sent first can be ensured to arrive first as long as (t- △ t) >0 or (t + △ t) >0 is ensured, the Flow sent first can not arrive first, as the time is positive, t + △ t is naturally greater than 0, so as long as (t- △ t) >0 is ensured, namely the time interval of the Flow is greater than the time delay difference of the paths, the problem of the correct flow control among the Flows can be avoided as the internal sequence of the Flow can not change in the transmission process, the correct sequence of the Flow is ensured, and the correct sequence of the Flow messages can be ensured to be achieved as long as the time is not to be naturally greater than the time delay difference of the Flow.
● source routing
The source routing refers to the way that the routing node which receives the data message firstly determines the whole routing of the datagram, and no longer every routing node participates in the path decision, but the identification of the routing is transmitted to the subsequent routing node by using the label in the data header. Therefore, the problem that each routing node is required to independently determine the next hop of the datagram by checking the target IP address of the data packet before inquiring the routing table of the routing node in the traditional IP routing is avoided, time and hardware resources are saved, and the performance of an application program is prevented from being reduced possibly.
● TT L carries a label
The label carried in the message header needs to be modified, and the traditional protocol cannot be violated because the label is compatible with the traditional protocol, the TT L field of the message header is selected to carry the label information, the TT L field is reduced by 1 when passing through a forwarding device, and as long as borrowing is not generated, the value of the high bit of the TT L is always kept unchanged, because Europe and mathematics start to operate from the low bit, the data characteristic of the TT L can allow the label information to be placed on the high 4 bits, and the placed label information cannot be modified as long as the low TT L value is not consumed.
Fig. 1 is a schematic diagram of setting of a TT L field in the invention, as shown in fig. 1, three types of messages exist in a multi-path routing system of the invention, and are distinguished by special codes in a TT L field, the first two bits are 00 to represent a traditional routing data message, the first two bits are 01 to represent a control signaling message, and the first bit is 1 to represent a label-carrying message, wherein x represents a bit used for limiting a traditional hop count, and y represents a bit of label information.
Fig. 2 is a block diagram of an embodiment of a multi-path routing system based on stream burstiness according to the present invention. As shown in fig. 2, the multi-path routing system based on stream burstiness of the present invention includes a gateway node and a forwarding node, and the following two nodes are respectively described in detail.
The gateway node is used for receiving the data message sent by the user and performing Flowlet division, wherein the time interval adopted during division is larger than the maximum time delay difference of all paths between the source address and the destination address of the data message, then processing the data message, namely determining a forwarding strategy for the data message based on the Flowlet, performing path calculation, and forwarding the data message after distributing a label. The specific work of the gateway node for processing the data message comprises the following aspects:
the gateway node maintains a Flowet state table, a forwarding policy table and a used label table, wherein the Flowet state table is used for storing matching items of various Flowets, corresponding labels, paths and duration of the Flowets, and the duration of the Flowets is the maximum delay difference of parallel forwarding paths calculated by the gateway node for the Flowets. Table 1 is an example of an entry of the Flowlet status table.
TABLE 1
In general, in order to control the size of the table, the gateway node monitors the Flowlet and deletes information corresponding to the Flowlet in the Flowlet status table when the gateway node is not used for a while.
The forwarding policy table is used for storing forwarding policies corresponding to the respective flowlets, that is, forwarding modes of the Flowlet data messages, and the specific forwarding policies can be set in advance according to needs. Table 2 is an example of an entry of the forwarding policy table.
TABLE 2
The used tags table is used to store tags that are currently in use in the system and their corresponding expiration times. Table 3 is an example of a used tag table.
| Used label | Time to failure |
| 1 | t1 |
| 7 | t2 |
| … | … |
TABLE 3
Fig. 3 is a flow chart of the processing of data packets by a gateway node in the present invention. As shown in fig. 3, the processing steps of the gateway node on the data packet in the present invention include:
s301: and judging whether the Flowet in which the data message is positioned exists in a Flowet state table or not, namely extracting a matching item from the data message, searching whether the Flowet exists in the Flowet state table or not, if not, entering a step S302, and otherwise, entering a step S306.
S302: matching the forwarding strategy:
searching a forwarding strategy table for a forwarding strategy corresponding to the Flowlet where the data message is located, namely extracting a matching item of the data message, matching in the forwarding strategy table, if no corresponding forwarding strategy exists, entering step S303, otherwise, entering step S304.
S303: marking as a traditional routing data message:
and setting the highest 2 bits in the TT L field of the current data message as '00', namely marking the data message as a traditional routing data message, and then forwarding the data message.
S304: calculating a forwarding path:
and extracting the forwarding strategy of the Flowet where the data message is located from the forwarding strategy table, and calculating N parallel forwarding paths according to the forwarding strategy, wherein N is set according to actual needs. That is to say, the Flowlet data message can be transmitted through N parallel forwarding paths in the invention.
S305: dispensing Flowlet label:
selecting one label from the unused labels as the label of the Flowet where the current data message is located, setting the failure time of the Flowet according to the requirement, recording the label and the failure time into a used label table, and recording the Flowet matching item, the corresponding label, the path and the duration of the Flowet into a Flowet state table. The process advances to step S308.
It can be seen that in the present invention, the distribution of new label information is not realized by using periodic interaction, but is triggered by a new Flowlet, and this way can adapt to the burstiness of the Flowlet and improve the response speed of the establishment of control information.
In general, the expiration time is set to the interval of the flowet where the current data message is located. The label failure time adopts relative time, so that the problem of time synchronization among different devices can be avoided, and the transmission time of the control signaling message on the path can be offset by adopting the relative time by the subsequent forwarding node.
S306: and judging whether the Flowlet control information is invalid, if so, entering a step S307, and otherwise, entering a step S309.
S307: extracting a Flowlet label and a forwarding path:
and extracting the label and N forwarding paths of the Flowet where the current data message is located from the Flowet state table, entering the step S308, and simultaneously pushing the failure time corresponding to the Flowet label back to a Flowet interval by using the label table.
For the update of the expiration time, the tag information can not be expired before the current Flowlet flow is not transmitted, and meanwhile, the old tag information can be ensured to be expired before the next Flowlet arrives. The failure time of the label is automatically maintained through the continuity of the Flowlet without introducing extra overhead, and the overhead of control information synchronization is greatly reduced.
S308: generating a control signaling message:
respectively generating corresponding control signaling messages for each forwarding path, wherein the source address and the destination address of each control signaling message are the same as those of the current data message, the contents of the control signaling messages are the label of the Flowet where the data message is located, the corresponding failure time and the forwarding path, setting the highest 2 bits of the TT L field in the header to be '01', and then sending the control information message.
S309: forwarding the data message:
setting the highest 1 bit of TT L field in the header of the data message as '1', and setting the 2 nd bit to the 4 th bit as the label of the Flowet where the data message is located, and converting the original data message into a message carrying the label for sending.
The forwarding node maintains a label instruction table, which includes the flowet identifier, which includes the source and destination addresses of the flowet and the label, and its corresponding instruction and expiration time, and the instruction includes the operations that the current forwarding node needs to complete in the forwarding path, which generally includes the adjustment mode of TT L value, the port and link to be forwarded, table 4 is an example of the label instruction table.
| Flowlet identification | Instructions | Time to failure |
| Source address, destination address, 1 | m1 | t1 |
| Source address, destination address, 7 | m2 | t2 |
| … | … | … |
TABLE 4
Because the gateway node can send out three data messages, the forwarding node has different processing modes aiming at each data message. The specific method comprises the following steps:
after receiving the data message, the forwarding node firstly extracts the highest 1 bit in the TT L field, if the data message is 1, the data message is a message carrying a label, if the data message is not further extracted the 2 nd bit in the TT L field, if the data message is 0, the forwarding node is a traditional routing data message, otherwise, the forwarding node is a control signaling message, and the processing modes of the three data messages are respectively as follows:
for the traditional routing data message, the forwarding node forwards the data message according to a traditional routing method.
For the control signaling message, the forwarding node extracts the source and destination addresses, the labels, the instructions and the failure time of the Flowlet, records the source and destination addresses, the labels, the instructions and the failure time into a label instruction table, and then forwards the control signaling message to the next node according to the path information in the control signaling message.
For the message carrying the label, the forwarding node firstly judges whether the forwarding node is a destination node, if so, the forwarding node extracts data and sends the data to an upper layer application, if not, the forwarding node extracts a source destination address and a label of the data message, inquires a label instruction table to obtain a corresponding instruction for operation, forwards the instruction to the next node, and meanwhile, pushes the failure time corresponding to the label of the Flowlet back to a Flowlet interval in the label instruction table.
When the tag instruction list is inquired, the source and destination addresses and the tags are used as tag identifications, the value range of the tags is very small, the tag values are used as physical storage positions of data, and corresponding information can be obtained by one-step searching; the value range of the IP address is very large, and the value formed by the source address pair and the destination address pair is irregular, so that a large amount of storage space is consumed if the method is the same as that of a label, and the system cannot work. Therefore, in this embodiment, a 16-bit CRC (Cyclic Redundancy Check) algorithm is used to query the source address and the destination address of the data packet, and then the tag value is directly calculated to obtain the storage location of the instruction, thereby improving the query efficiency.
In the invention, the speed of forwarding the data message by the forwarding node through the lookup label is greatly improved compared with the traditional lookup IP routing table, and meanwhile, the simplification of the forwarding node does not need to realize three-layer network functions such as routing decision and the like. Compared with the traditional routing node, the forwarding node has great advantages in terms of processing speed and implementation complexity.
As can be seen from the above description, most of the functions of the routing decision are carried by the gateway node in the present invention, and the operation is relatively complex. Therefore, in order to improve the operation efficiency of the gateway node, the embodiment improves the specific structure of the gateway node. Fig. 4 is a structural diagram of a gateway node in the present invention. As shown in fig. 4, the gateway node in the present invention includes a centralized function processing unit, a path control engine, and a database. Each module is described in detail below.
The centralized function processing unit is used for finishing the processing of the data message, and comprises the steps of determining a forwarding strategy, generating and sending a control signaling message, modifying and sending the data message, extracting a source and destination address of the data message when determining that the forward-to-forward strategy needs to calculate a parallel forwarding path, packaging the source and destination address of the data message and the forwarding strategy into a path request, sending the path request to the path control engine, and then receiving the parallel forwarding path fed back by the path control engine to further generate the control signaling message.
The path control engine is used for responding to the path request, extracting a source and destination address and a forwarding strategy from the path request, and calculating to obtain a corresponding parallel forwarding path to be fed back to the centralized function processing unit.
The database is used for storing a Flowlet state table, a forwarding policy table and a used label table.
The specific processing steps of the centralized function processing unit for the data message comprise:
1) after receiving the data message, extracting a matching item, inquiring a Flowet state table in a database, judging whether the Flowet in which the data message is positioned exists in the Flowet state table, if not, entering a step 2), otherwise, entering a step S6;
2) extracting a matching item of the data message, inquiring a forwarding policy table in a database, judging whether the matching is successful, if the matching is not successful, entering the step 3), and if not, entering the step 4);
3) setting the highest 2 bits in the TT L field of the current data message as '00', namely marking as the traditional routing data message, and then forwarding;
4) extracting a forwarding strategy of the Flowlet where the data message is located from a forwarding strategy table of a database, extracting a source destination address of the data message, packaging the source destination address and the forwarding strategy together into a path request, and sending the path request to a path control engine;
5) after receiving a parallel forwarding path fed back by a path control engine, selecting one of unused labels as a label of a Flowet where a current data message is located, setting failure time of the Flowet according to needs, wherein the failure time adopts relative time, then recording the label and the failure time into a used label table of a database, and recording a Flowet matching item, a corresponding label, a path and duration time of the Flowet into a Flowet state table of the database;
6) judging whether the Flowlet control information is invalid, if so, entering a step 7), and if not, entering a step 9);
7) extracting a label and N forwarding paths of the Flowet where the current data message is from a Flowet state table of the database, entering step S8, and simultaneously pushing the failure time corresponding to the Flowet label back to a Flowet interval by using the label table;
8) respectively generating corresponding control signaling messages for each forwarding path, wherein the source address and the destination address of each control signaling message are the same as those of the current data message, the contents of the control signaling messages are the label of the Flowet where the data message is located, the corresponding failure time and the forwarding path, the highest 2 bits of the TT L field in the header are set as '01', and then sending a control information message;
9) setting the highest 1 bit of TT L field in the header of the data message as '1', and the 2 nd bit to the 4 th bit as the label of the Flowlet where the data message is located, converting the original data message into a message carrying the label and sending the message;
in order to facilitate the path control engine to perform path calculation and update, a topology information table and a path information table may also be set in the database, and are used to store network topology information and existing parallel forwarding path information. Through the setting of the topology information table, the path control engine can obtain the latest topology information again under the condition of receiving the notification of the change of the topology information, so that the processing process of path calculation is accelerated, and the interaction time with the database is reduced. The setting of the path information table enables the path information in the path forwarding table to be directly modified without interaction with the original service request party when the path is re-planned. Because the path adjustment process caused by link change is transparent to the centralized function processing unit, the consistency of data reading and modification of the path adjustment process and the centralized function processing unit is ensured through the atomicity of the database, the complexity of the path control engine is greatly reduced, and the path control engine can put more resources on path computation.
In order to realize the flexibility of the network structure, the centralized function processing unit, the path control engine and the database are interacted through corresponding network interfaces, so that three function modules can be deployed on one machine, and can also be deployed on a plurality of machines for redundancy configuration, thereby improving the safety and reliability of the system.
The multi-path routing system based on stream burstiness can be deployed alone or in a mixed way with a traditional uncontrolled network when the system is actually deployed, as shown in FIG. 5, a part of the network is the traditional uncontrolled network when the system is deployed in a mixed way, the traditional uncontrolled network has IP message transmission capability, and the traditional uncontrolled network is transparent only to be regarded as a point-to-point transmission route for the multi-path routing system of the invention, and conversely, the multi-path routing system of the invention is transparent for the traditional network, and messages generated by the multi-path routing system are treated as common traditional IP messages, and the multi-path routing system and the traditional network are mutually transparent and work independently, so that the key point for realizing the mode is that the TT L field is set, and by virtue of the mathematical characteristic of TT L, the invention uses TT L as an intermediate bridge, not only completes the function of hop limit in the traditional network, but also can transmit required messages in a mode that the TT L can not be normally deployed by the traditional network, and even if the traditional network is completely damaged by the multi-path routing system, the TT 54 can not be used for correcting the messages.
Although illustrative embodiments of the present invention have been described above to facilitate the understanding of the present invention by those skilled in the art, it should be understood that the present invention is not limited to the scope of the embodiments, and various changes may be made apparent to those skilled in the art as long as they are within the spirit and scope of the present invention as defined and defined by the appended claims, and all matters of the invention which utilize the inventive concepts are protected.
Claims (4)
1. A multi-path routing system based on flow burstiness, comprising a gateway node and a forwarding node, wherein:
the gateway node is configured to receive a data packet sent by a user and perform Flowlet division, where a time interval adopted during division is greater than a maximum delay difference of all paths between source and destination addresses of the data packet, and then process the data packet, that is, determine a forwarding policy for the data packet based on Flowlet, perform path calculation, and forward the data packet after distributing a label, and specifically includes:
the gateway node maintains a Flowet state table, a forwarding policy table and a used label table, wherein the Flowet state table is used for storing matching items of various Flowets, corresponding labels, paths and duration of the Flowets; the forwarding policy table is used for storing forwarding policies corresponding to the Flowlets, and the specific forwarding policies can be set in advance according to needs; the used label table is used for storing the labels which are used in the current system and the corresponding failure time;
the gateway node processing the data message comprises the following steps:
s1: judging whether the Flowet in which the data message is positioned exists in a Flowet state table or not, namely extracting a matching item from the data message, searching whether the Flowet exists or not in the Flowet state table, if not, entering a step S2, otherwise, entering a step S6;
s2: searching a forwarding strategy table for a forwarding strategy corresponding to the Flowet where the data message is located, namely extracting a matching item of the data message, matching in the forwarding strategy table, if the corresponding forwarding strategy does not exist, entering step S33, otherwise, entering step S4;
s3, setting the highest 2 bits in TT L field of current data message as '00', namely marking as traditional route data message, and then forwarding;
s4: extracting a forwarding strategy of the Flowet where the data message is located from the forwarding strategy table, and calculating N parallel forwarding paths according to the forwarding strategy, wherein N is set according to actual needs;
s5: selecting one label from unused labels as the label of the Flowet where the current data message is located, setting the failure time of the Flowet according to the requirement, recording the label and the failure time into a used label table, and recording the Flowet matching item, the corresponding label, the path and the duration of the Flowet into a Flowet state table;
s6: judging whether the Flowlet control information is invalid, if so, entering a step S7, otherwise, entering a step S9;
s7: extracting a label and N forwarding paths of the Flowet where the current data message is located from the Flowet state table, entering step S8, and simultaneously pushing the failure time corresponding to the Flowet label back to a Flowet interval by using the label table;
s8, generating corresponding control signaling messages for each forwarding path respectively, wherein the source address and the destination address of the control signaling messages are the same as the current data messages, the contents of the control signaling messages are the labels of the Flowlets where the data messages are located, the corresponding failure time and the forwarding paths, the highest 2 bits of the TT L field in the header are set as '01', and then sending the control information messages;
s9, setting the highest 1 bit of TT L field in the header of the data message as '1', and the 2 nd bit to the 4 th bit as the label of the Flowet where the data message is located, converting the original data message into a message carrying the label to send;
after receiving a data message, the forwarding node firstly extracts the highest 1 bit in the TT L field, if the data message is 1, the data message is a message carrying the label, if not, further extracts the 2 nd bit in the TT L field, if the data message is 0, the forwarding node is a traditional routing data message, otherwise, the forwarding node is a control signaling message, and the processing modes of the three data messages are respectively as follows:
for the traditional routing data message, the forwarding node forwards the data message according to a traditional routing method.
For the control signaling message, the forwarding node extracts the source and destination addresses, the labels, the instructions and the failure time of the Flowlet, records the source and destination addresses, the labels, the instructions and the failure time into a label instruction table, and then forwards the control signaling message to the next node according to the path information in the control signaling message.
For the message carrying the label, the forwarding node firstly judges whether the forwarding node is a destination node, if so, the forwarding node extracts data and sends the data to an upper layer application, if not, the forwarding node extracts a source destination address and a label of the data message, inquires a label instruction table to obtain a corresponding instruction for operation, forwards the instruction to the next node, and meanwhile, pushes the failure time corresponding to the label of the Flowlet back to a Flowlet interval in the label instruction table.
2. The system of claim 1, wherein the gateway node comprises a centralized function processing unit, a path control engine, and a database, wherein:
the centralized function processing unit is used for finishing the processing of the data message, and comprises a forwarding strategy determination unit, a control signaling message generation and transmission unit, a data message modification and transmission unit, a path control engine and a forwarding strategy control unit, wherein when the forwarding strategy improvement unit determines that a parallel forwarding path needs to be calculated, the source and destination addresses of the data message are extracted, and the source and destination addresses and the forwarding strategy are packaged together into a path request and transmitted to the path control engine, and then the parallel forwarding path fed back by the path control engine is received to further generate the control signaling message;
the path control engine is used for responding to the path request, extracting a source and destination address and a forwarding strategy from the path request, and calculating to obtain a corresponding parallel forwarding path to be fed back to the centralized function processing unit;
the database is used for storing a Flowlet state table, a forwarding policy table and a used label table.
3. The flow burstiness-based multipath routing system of claim 1, wherein the label failure time is an interval of a Flowlet where a current data packet is located.
4. The multi-path routing system based on the flow burstiness as claimed in claim 1, wherein when the forwarding node queries the tag instruction table, the forwarding node queries a source address and a destination address of the data packet by using a 16-bit CRC algorithm, and then directly calculates a tag value to obtain a storage location of the instruction.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010332340.XA CN111490934B (en) | 2020-04-24 | 2020-04-24 | Multi-path routing system based on stream burstiness |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010332340.XA CN111490934B (en) | 2020-04-24 | 2020-04-24 | Multi-path routing system based on stream burstiness |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111490934A true CN111490934A (en) | 2020-08-04 |
| CN111490934B CN111490934B (en) | 2021-09-14 |
Family
ID=71813647
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010332340.XA Expired - Fee Related CN111490934B (en) | 2020-04-24 | 2020-04-24 | Multi-path routing system based on stream burstiness |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111490934B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2022247307A1 (en) * | 2021-05-28 | 2022-12-01 | 华为技术有限公司 | Message transmission method and apparatus |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103354520A (en) * | 2013-07-02 | 2013-10-16 | 华为技术有限公司 | Label processing method and device employing the same |
| WO2016070535A1 (en) * | 2014-11-04 | 2016-05-12 | 中兴通讯股份有限公司 | Method and apparatus for rearrangement in multi-protocol label switching ring network |
| CN107579920A (en) * | 2017-09-25 | 2018-01-12 | 盛科网络（苏州）有限公司 | Transmission method, device, storage medium and the processor of data flow |
| CN108648550A (en) * | 2018-05-15 | 2018-10-12 | 陆昊民 | A kind of Accounting Course system |
| CN109561026A (en) * | 2018-12-10 | 2019-04-02 | 福州大学 | Burst method for stream processing based on software defined network carrier state data plane |
| US20190319882A1 (en) * | 2016-12-27 | 2019-10-17 | Huawei Technologies Co., Ltd. | Transmission Path Determining Method and Apparatus |
-
2020
- 2020-04-24 CN CN202010332340.XA patent/CN111490934B/en not_active Expired - Fee Related
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103354520A (en) * | 2013-07-02 | 2013-10-16 | 华为技术有限公司 | Label processing method and device employing the same |
| WO2016070535A1 (en) * | 2014-11-04 | 2016-05-12 | 中兴通讯股份有限公司 | Method and apparatus for rearrangement in multi-protocol label switching ring network |
| US20190319882A1 (en) * | 2016-12-27 | 2019-10-17 | Huawei Technologies Co., Ltd. | Transmission Path Determining Method and Apparatus |
| CN107579920A (en) * | 2017-09-25 | 2018-01-12 | 盛科网络（苏州）有限公司 | Transmission method, device, storage medium and the processor of data flow |
| CN108648550A (en) * | 2018-05-15 | 2018-10-12 | 陆昊民 | A kind of Accounting Course system |
| CN109561026A (en) * | 2018-12-10 | 2019-04-02 | 福州大学 | Burst method for stream processing based on software defined network carrier state data plane |
Non-Patent Citations (2)
| Title |
|---|
| ZHIQIANG GUO等: ""EasyLB: Adaptive Load Balancing Based on Flowlet Switching for Wireless Sensor Networks"", 《SENSORS 2018》 * |
| 熊建辉: ""MPLS 转发机制的研究与仿真实现"", 《萍乡学院学报》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2022247307A1 (en) * | 2021-05-28 | 2022-12-01 | 华为技术有限公司 | Message transmission method and apparatus |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111490934B (en) | 2021-09-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US7590756B2 (en) | Method and system for transferring data in a communications network using redundant communication paths | |
| CN101496348B (en) | Technique for multiple path forwarding of label-switched data traffic | |
| CN102263697B (en) | Method and device for sharing aggregated link traffic | |
| US7778204B2 (en) | Automatic maintenance of a distributed source tree (DST) network | |
| EP2421205A1 (en) | Flooding-based routing protocol having average-rate and burst-rate control | |
| CN101277477B (en) | Method, apparatus and system for equalizing flux | |
| WO2018010529A1 (en) | Method and apparatus for an information-centric mac layer | |
| CN101651630A (en) | Switching method and switching device based on link fault | |
| CN101741709A (en) | Method and system for establishing label switched path and network node | |
| CN104717259B (en) | The multi-path data Transmission system and method for distributed transfer server network assistance | |
| US20230370298A1 (en) | Message header processing method and apparatus, storage medium and electronic device | |
| CN107454000B (en) | Network data transmission device and method | |
| CN109525304B (en) | Space intelligent network architecture integrating perception, calculation and storage | |
| WO2009036678A1 (en) | Method, device and network system for message forwarding | |
| CN101141376B (en) | Method to establish TE path, path computation equipment and system | |
| EP4191966A1 (en) | Method and device for processing data message, storage medium, and electronic device | |
| CN116192726A (en) | Forwarding path establishment method, forwarding path establishment device and computer readable storage medium | |
| Li et al. | A reliable and efficient forwarding strategy in vehicular named data networking | |
| CN113965246A (en) | Application layer route forwarding optimization method based on-satellite UPF | |
| CN112688869A (en) | Data reliable transmission method based on dynamic routing algorithm in weak network environment | |
| EP1699169A1 (en) | Wireless base station, wireless mobile device, and wireless access network for reducing signalling traffic | |
| CN111490934B (en) | Multi-path routing system based on stream burstiness | |
| Jin et al. | MANET for Disaster Relief based on NDN | |
| CN103905307A (en) | MPLS multi-topology path label notification method and device | |
| CN104247346A (en) | Pseudowire extended group messaging in a packet switched network |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20210914 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |