CN112995032B - Segment routing traffic engineering method and device based on limited widest path - Google Patents

Segment routing traffic engineering method and device based on limited widest path Download PDF

Info

Publication number
CN112995032B
CN112995032B CN202110549603.7A CN202110549603A CN112995032B CN 112995032 B CN112995032 B CN 112995032B CN 202110549603 A CN202110549603 A CN 202110549603A CN 112995032 B CN112995032 B CN 112995032B
Authority
CN
China
Prior art keywords
path
segment
network
traffic engineering
node
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
Application number
CN202110549603.7A
Other languages
Chinese (zh)
Other versions
CN112995032A (en
Inventor
郭得科
罗来龙
崔思晨
任棒棒
汪漪
郑龙
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
National University of Defense Technology
Original Assignee
National University of Defense Technology
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by National University of Defense Technology filed Critical National University of Defense Technology
Priority to CN202110549603.7A priority Critical patent/CN112995032B/en
Publication of CN112995032A publication Critical patent/CN112995032A/en
Application granted granted Critical
Publication of CN112995032B publication Critical patent/CN112995032B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

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
    • H04L45/34—Source routing
    • 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
    • H04L45/125—Shortest path evaluation based on throughput or bandwidth

Abstract

The invention provides a segment routing traffic engineering method and a device based on a limited widest path, wherein the traffic engineering method uses a series of segments to encode the widest path after searching the widest path of a stream, and selects a relay node while considering the shortest path in a network in the path. The traffic engineering method can bypass the most congested link to the maximum extent, can guarantee performance, needs less time, and can be suitable for three or more segment routing scenes.

Description

Segment routing traffic engineering method and device based on limited widest path
Technical Field
The invention belongs to the technical field of traffic engineering, and particularly relates to a segment routing traffic engineering method and device based on a limited widest path.
Background
With the rapid development of the internet, the network traffic has problems such as explosive growth, and is limited by the traditional routing algorithm and scheduling method, and the network traffic is not easily distributed uniformly on the link, thereby causing network congestion and reducing the network service quality. Traffic engineering, which targets the analysis, prediction, and regulation of network traffic and improves network performance, therefore, is of paramount importance. Traffic Engineering (TE) is a technology for optimizing network Traffic distribution, and can optimize and schedule network Traffic, thereby realizing network Traffic load balancing, reducing congestion, and improving the utilization rate of network resources.
Recently, segment routing techniques based on IPv6 (SRv 6) have been proposed to avoid network flows being routed through congested links. SRv6 the core idea is to divide a routing path into segments representing the end-to-end logical path, the segments being represented as a series of instructions encapsulated in a header so that each flow can be routed as needed. Fig. 1 depicts a simple example of an SRv6 network, where a flow is transmitted from router R1 to R5. According to the default shortest-circuit method, the flow is transmitted according to the path of R1 → R4 → R5. In conventional routing protocols, in order to change the default routing path, we have to update the link weights and wait for the routing protocol to converge, but this approach is certainly complex and time consuming. In contrast, the SRv6 technique makes it easier and faster to change the default routing path. In fig. 1, R1 will add a Segment Routing Header (SRH) to the packet header. The SRH consists of two parts, a segment left indicator and a segment list. The target address of the current segment is the address of the segment pointed to by the left indicator in the segment list. When a packet is received by R2, the value of the left part of the segment is reduced, and the corresponding segment is activated, i.e. the current destination address is set to 2001:: 5. The packet will then be routed along the shortest path between R2 to R5.
Considering SRv6, the traffic engineering algorithm based on SRv6 network has become a hot problem in traffic engineering research due to its advantages of simplicity, easy deployment, and strong scalability. However, existing methods of Segment Routing Traffic Engineering (SRTE) lack generality because they only consider cases involving two or three segments. As in the prior art document one ("Optimized Network Traffic Engineering using Segment Routing," in Proc. of IEEE INFOCOM, 2015, pp. 657-. In the two-stage model ("Segment routed routing with bound structure in software-defined Networks," in Proc. of IEEE Conference on Local Computer Networks, 2018, pp. 477-. In the prior art document three ("Traffic engineering segmentation routing and knowledge of a carrier IP network," IEEE/ACM Transactions on network, 2018, pp. 1851. the first 1864. Schuller et al extends the SRTE model to only three segments. Due to the tight constraints of two or three segments, the above-described prior approaches may lose the opportunity to improve performance when multiple intermediate segments are available.
Disclosure of Invention
In view of this, the present invention provides a segment routing traffic engineering method and device for a constrained widest path, so as to solve the problem that the existing traffic engineering algorithm can only adapt to strict constraints of two or three segments.
A segment routing traffic engineering method based on a limited widest path comprises the following steps:
step 1: searching and acquiring SRv6 a widest path of all paths between a source node and a destination node corresponding to any flow in the network, where the widest path is a path with a minimum utilization rate among all paths, a utilization rate of each path of all paths is a utilization rate of a largest link among all links of each path, and the largest link is a link of each path with a maximum utilization rate among the links,
step 2: first encoding said widest path with segments in said SRv6 network to obtain an encoded path represented by a combination of segments formed by said segments,
and step 3: and judging whether the number of the segments used in the segment combination for representing the coding path is greater than that of the segment of the SRv6 network, if so, recording the hop counts between each segment and an upstream segment in the coding path, comparing the hop counts corresponding to the segments, deleting the segment with the minimum hop count from the segment combination, and returning to the step 3 until the number of the segments in the segment combination does not exceed that of the segment of the SRv6 network.
Preferably, before the step 1, the method further comprises:
modeling SRv6 networks as undirected graphs
Figure 726681DEST_PATH_IMAGE001
Wherein, in the step (A),
Figure 909401DEST_PATH_IMAGE002
represented as a collection of network routers,
Figure 95662DEST_PATH_IMAGE003
representing physical links between routers, from network node
Figure 22030DEST_PATH_IMAGE004
To a network node
Figure 242927DEST_PATH_IMAGE005
Of a link
Figure 596548DEST_PATH_IMAGE006
All have a given capacity
Figure 535685DEST_PATH_IMAGE007
And weight
Figure 265744DEST_PATH_IMAGE008
∈WTo characterize the link
Figure 341147DEST_PATH_IMAGE009
The cost of transmitting one single-bit stream,
and networking said SRv6
Figure 600090DEST_PATH_IMAGE010
In a stream ofiA stream
Figure 557682DEST_PATH_IMAGE011
By using
Figure 701218DEST_PATH_IMAGE012
It is shown that,
Figure 755762DEST_PATH_IMAGE013
is a stream
Figure 920027DEST_PATH_IMAGE011
The source node of (a) is,
Figure 568177DEST_PATH_IMAGE014
is a stream
Figure 640038DEST_PATH_IMAGE011
The destination node of (2) is,
Figure 707612DEST_PATH_IMAGE015
is a stream
Figure 573937DEST_PATH_IMAGE011
The size of (a) is (b),
wherein the flow is in an n-segment SRv6 network
Figure 709383DEST_PATH_IMAGE011
Can follow a path
Figure 319356DEST_PATH_IMAGE016
Routing of, wherein
Figure 223858DEST_PATH_IMAGE017
At most have
Figure 729926DEST_PATH_IMAGE018
The number of the nodes is one,
Figure 211723DEST_PATH_IMAGE019
each sub-path in (a) isnSegment SRv6 is the shortest path between an upstream node and a downstream node in the network.
Preferably, the step 1 comprises:
order to
Figure 500753DEST_PATH_IMAGE020
Representing flows
Figure 649975DEST_PATH_IMAGE011
Of the source node
Figure 999047DEST_PATH_IMAGE013
And a target node
Figure 171403DEST_PATH_IMAGE014
The set of all the paths in between,
each path in the set
Figure 123178DEST_PATH_IMAGE021
Taking the maximum link utilization of all links in the path as the path
Figure 267852DEST_PATH_IMAGE021
Path utilization of
Figure 912460DEST_PATH_IMAGE022
By using
Figure 244215DEST_PATH_IMAGE023
Indicating a link
Figure 202944DEST_PATH_IMAGE024
Link utilization of
By comparing the
Figure 795599DEST_PATH_IMAGE022
Determining the widest path as the path in the set
Figure 486475DEST_PATH_IMAGE025
Wherein 1 is less than or equal toj≤n，
Figure 695739DEST_PATH_IMAGE026
。
Preferably, the widest path is found in the set at step 1 by an iterative method comprising:
setting an accessed list
Figure 130263DEST_PATH_IMAGE027
And an unaccessed list
Figure 843004DEST_PATH_IMAGE028
In combination with each other
Figure 767097DEST_PATH_IMAGE029
Is shown in
Figure 73445DEST_PATH_IMAGE030
The path in the sub-iterationq j Utilization of, said pathq j To be a slave source node
Figure 436293DEST_PATH_IMAGE028
To the network router setVTo (1)jA node
Figure 144486DEST_PATH_IMAGE031
Is detected by the optical sensor (c) and is,
initialization
Figure 239481DEST_PATH_IMAGE032
、
Figure 157759DEST_PATH_IMAGE033
、
Figure 934085DEST_PATH_IMAGE034
，
Update the table at each iteration
Figure 355839DEST_PATH_IMAGE035
Wherein
Figure 824997DEST_PATH_IMAGE036
，
Figure 964992DEST_PATH_IMAGE037
，
Then will be
Figure 872905DEST_PATH_IMAGE038
Node with minimum corresponding value
Figure 290111DEST_PATH_IMAGE028
Move to
Figure 54804DEST_PATH_IMAGE027
，
When in use
Figure 557461DEST_PATH_IMAGE027
Is empty, from
Figure 3486DEST_PATH_IMAGE035
Find out from
Figure 399832DEST_PATH_IMAGE013
To be at
Figure 945214DEST_PATH_IMAGE039
The widest path corresponding to any node in the set.
Preferably, in the step 2:
combining the segments
Figure 325380DEST_PATH_IMAGE040
Represents the abovef i The widest path coding path of (1), wherein
Figure 975761DEST_PATH_IMAGE041
，
Figure 695455DEST_PATH_IMAGE042
，
And order
Figure 536372DEST_PATH_IMAGE043
Is on the way
Figure 13621DEST_PATH_IMAGE044
In
Figure 863765DEST_PATH_IMAGE045
And
Figure 110070DEST_PATH_IMAGE046
sub-path between, let
Figure 325151DEST_PATH_IMAGE047
To represent
Figure 148750DEST_PATH_IMAGE048
To
Figure 677952DEST_PATH_IMAGE049
Shortest path therebetween, comparison
Figure 903397DEST_PATH_IMAGE043
And
Figure 961483DEST_PATH_IMAGE047
the size of (1) when
Figure 741220DEST_PATH_IMAGE050
Then use
Figure 667588DEST_PATH_IMAGE051
And
Figure 888484DEST_PATH_IMAGE046
address representation of
Figure 507685DEST_PATH_IMAGE043
，
Through each of said sub-paths represented by segments
Figure 181243DEST_PATH_IMAGE043
Representing the coding pathp i w 。
Preferably, thenGreater than or equal to 2, and thenPreferably greater than 3.
Preferably, in the step 1, the iterative method is carried out
Figure 114563DEST_PATH_IMAGE052
Sub-iterations, each iteration requiring execution
Figure 49021DEST_PATH_IMAGE053
A time complexity of the iterative method of
Figure 980068DEST_PATH_IMAGE054
In the step 2, the length of the widest path is not more than
Figure 265556DEST_PATH_IMAGE052
So the iterative method needs to compare the widest path and the shortest path
Figure 736989DEST_PATH_IMAGE055
Then, in the stepIn step 3, the time complexity is not exceeded
Figure 666899DEST_PATH_IMAGE054
All of the hop values of (a) are sorted.
Preferably, thennThe segment SRv6 network is modeled as:
Figure 627901DEST_PATH_IMAGE056
Figure 276052DEST_PATH_IMAGE057
Figure 347913DEST_PATH_IMAGE058
Figure 397908DEST_PATH_IMAGE059
Figure 264233DEST_PATH_IMAGE060
Figure 727576DEST_PATH_IMAGE061
wherein constraint equation (3) in the SRv6 network modeling ensures that all traffic is routed in the SRv6 network and that each traffic passes through only a single onenSegment routing path, in constraint equation (4) in said SRv6 network modeling, if flow
Figure 212915DEST_PATH_IMAGE011
Is a through path
Figure 242051DEST_PATH_IMAGE062
Routed, then
Figure 420222DEST_PATH_IMAGE063
To represent
Figure 167598DEST_PATH_IMAGE064
In (A) belong to
Figure 784524DEST_PATH_IMAGE011
And the constraint (4) indicates that the link utilization of each link is less than the maximum link utilization,
Figure 543533DEST_PATH_IMAGE065
representing flows
Figure 17240DEST_PATH_IMAGE011
At the source node
Figure 127278DEST_PATH_IMAGE013
And the target node
Figure 79054DEST_PATH_IMAGE014
The number of possible paths between the two,
Figure 286044DEST_PATH_IMAGE066
represents from
Figure 806018DEST_PATH_IMAGE067
The number of permutations of h different objects is found,
Figure 262407DEST_PATH_IMAGE068
representing a source node
Figure 887381DEST_PATH_IMAGE013
And
Figure 948878DEST_PATH_IMAGE069
all the shortest paths between the first and second nodes,
Figure 764387DEST_PATH_IMAGE070
to represent
Figure 583438DEST_PATH_IMAGE064
In that
Figure 142596DEST_PATH_IMAGE069
Appear at
Figure 730703DEST_PATH_IMAGE068
The total frequency of occurrence of (a) is,
Figure 920376DEST_PATH_IMAGE071
representing flows
Figure 351357DEST_PATH_IMAGE011
Whether or not to pass
Figure 589571DEST_PATH_IMAGE072
And (4) path routing.
A terminal comprising a memory and a processor, the memory storing a computer program which, when executed by the processor, performs the segment routing traffic engineering method as claimed in any one of the preceding claims.
A computer readable storage medium storing a computer program which, when executed by a processor, performs any of the segment routing traffic engineering methods described above.
From the above, it can be seen that the method and apparatus for routing traffic engineering based on the segment of the constrained widest path according to the present invention, after searching out the widest path of a stream, encodes the widest path using a series of segments, and selects a relay node while considering the shortest path in the network within the path. The process engineering method can bypass the most congested link to the maximum extent, can guarantee performance, needs less time, and can be suitable for three or more segment routing scenes.
Drawings
FIG. 1 is a SRv6 network routing diagram;
FIG. 2 is a schematic diagram of a network constructed in accordance with the present invention;
FIG. 3 is a graph showing a comparison of the performance of N-SR, RWP and SP in 400 flow comparison experiments in topology rand 20;
FIG. 4 is a graphical representation of a comparison of the performance of N-SR, RWP and SP in 400 flow comparison experiments in four topologies;
FIG. 5 is a graphical representation of a comparison of the performance of N-SR, RWP and SP in different numbers of comparative experiments in topology rf 1221.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without any creative effort, shall fall within the protection scope of the present invention. It should be noted that "…" in this description of the preferred embodiment is only for technical attributes or features of the present invention.
In the present invention, we model the network as an undirected graph
Figure 156819DEST_PATH_IMAGE073
Wherein
Figure 189497DEST_PATH_IMAGE002
Represented as a collection of network routers,
Figure 311037DEST_PATH_IMAGE074
representing the physical links between routers. Slave node
Figure 946418DEST_PATH_IMAGE004
To the node
Figure 509117DEST_PATH_IMAGE005
Of a link
Figure 837330DEST_PATH_IMAGE075
All have a given capacity
Figure 118270DEST_PATH_IMAGE076
And weight
Figure 822921DEST_PATH_IMAGE077
Which is shown in
Figure 568023DEST_PATH_IMAGE078
The cost of transmitting one single-bit stream. Suppose there is
Figure 942504DEST_PATH_IMAGE079
A stream
Figure 835373DEST_PATH_IMAGE080
. Free stream
Figure 687923DEST_PATH_IMAGE011
Abstracted as triples having three attributes, i.e.
Figure 349848DEST_PATH_IMAGE081
Wherein
Figure 223126DEST_PATH_IMAGE082
In order to be the source node of the network,
Figure 213079DEST_PATH_IMAGE083
in order for the destination node to be able to,
Figure 993953DEST_PATH_IMAGE015
is the size of the stream. The n-segment routing problem is defined as follows:
n segment routing
Figure 651331DEST_PATH_IMAGE084
: on a network
Figure 492248DEST_PATH_IMAGE085
Each stream of
Figure 969497DEST_PATH_IMAGE086
Can follow a path
Figure 757324DEST_PATH_IMAGE087
Routing of, wherein
Figure 128262DEST_PATH_IMAGE017
At most have
Figure 15447DEST_PATH_IMAGE088
A node, the first
Figure 104626DEST_PATH_IMAGE011
Corresponding path
Figure 368248DEST_PATH_IMAGE089
。
Figure 593693DEST_PATH_IMAGE019
Each sub-path in (a) is
Figure 979675DEST_PATH_IMAGE085
Shortest path between intermediate upstream node and downstream node.
It is to be noted that it is preferable that,
Figure 431516DEST_PATH_IMAGE090
should not contain
Figure 623463DEST_PATH_IMAGE091
And
Figure 578781DEST_PATH_IMAGE014
else, path
Figure 401243DEST_PATH_IMAGE019
There will be one cycle. Let
Figure 199435DEST_PATH_IMAGE092
To represent
Figure 533421DEST_PATH_IMAGE093
And
Figure 999037DEST_PATH_IMAGE014
the number of possible paths in between, then we have:
Figure 664505DEST_PATH_IMAGE094
Figure 887676DEST_PATH_IMAGE095
represents from
Figure 421426DEST_PATH_IMAGE096
The number of permutations of h different objects is found. Thus, for one stream
Figure 616915DEST_PATH_IMAGE011
Is provided with
Figure 312338DEST_PATH_IMAGE097
A possible path, i.e.
Figure 288384DEST_PATH_IMAGE098
，
Figure 235612DEST_PATH_IMAGE062
Representing a path
Figure 144662DEST_PATH_IMAGE099
To (1) ajOne possible path (candidate path), 1 ≦j≤
Figure 417512DEST_PATH_IMAGE092
。
It should be noted that if there are multiple shortest paths between a pair of nodes, the flows are typically evenly distributed between these paths according to the ECMP protocol. Due to the fact that
Figure 412012DEST_PATH_IMAGE099
Candidate sub-path in (1)
Figure 490827DEST_PATH_IMAGE062
Only intermediate nodes that have to be walked are restricted, so that there will be multiple satisfactions
Figure 395329DEST_PATH_IMAGE062
The shortest path of (2). When selecting
Figure 698134DEST_PATH_IMAGE062
After that, we can calculate
Figure 320877DEST_PATH_IMAGE078
The ratio of the carried traffic is as follows:
Figure 937803DEST_PATH_IMAGE100
wherein
Figure 821445DEST_PATH_IMAGE101
To represent
Figure 904939DEST_PATH_IMAGE013
And
Figure 139611DEST_PATH_IMAGE069
all the shortest paths between the first and second nodes,
Figure 232332DEST_PATH_IMAGE102
to represent
Figure 173743DEST_PATH_IMAGE078
In that
Figure 818351DEST_PATH_IMAGE069
Appear at
Figure 415686DEST_PATH_IMAGE101
The total frequency of occurrence of (a).
Setting a binary variable
Figure 171152DEST_PATH_IMAGE071
Representing flows
Figure 904753DEST_PATH_IMAGE011
Whether or not to pass
Figure 923524DEST_PATH_IMAGE062
Path routing and Maximum Link Utilization (MLU) routing in the network
Figure 601631DEST_PATH_IMAGE103
And (4) showing. Then we solve the n-segment routing problem (which can be modeled as:
Figure 301733DEST_PATH_IMAGE104
Figure 14474DEST_PATH_IMAGE105
Figure 204147DEST_PATH_IMAGE106
Figure 510495DEST_PATH_IMAGE107
constraint equation (3) ensures that all traffic is routed in the network and that each traffic only passes through a single n-segment routing path. In constraint equation (4), if flow
Figure 873343DEST_PATH_IMAGE011
Is a through path
Figure 315957DEST_PATH_IMAGE062
Routed, then
Figure 473269DEST_PATH_IMAGE108
To represent
Figure 329229DEST_PATH_IMAGE109
In (A) belong to
Figure 105555DEST_PATH_IMAGE011
The flow rate of (c). Restraint maleEquation (4) indicates that the link utilization of each link should be less than the maximum link utilization.
In the present invention, the n-segment routing problem is an NP-hard problem. We prove by way of example of a network shown in the figure that the n-segment routing problem is indeed an NP-hard problem, and the proving process is as follows:
as shown in fig. 2, assume that
Figure 792889DEST_PATH_IMAGE110
To
Figure 262047DEST_PATH_IMAGE111
Exist in a size of
Figure 402041DEST_PATH_IMAGE112
Is/are as follows
Figure 44375DEST_PATH_IMAGE113
A stream from
Figure 291660DEST_PATH_IMAGE114
To
Figure 56353DEST_PATH_IMAGE115
Exist in a size of
Figure 559010DEST_PATH_IMAGE116
Is/are as follows
Figure 5035DEST_PATH_IMAGE117
And (4) each stream. Attributes of each link
Figure 666960DEST_PATH_IMAGE118
Represents its weight and capacity, wherein
Figure 477921DEST_PATH_IMAGE119
. From FIG. 2, it can be seen that
Figure 326929DEST_PATH_IMAGE120
And
Figure 248748DEST_PATH_IMAGE121
would be the most congested link. To minimize the network MLU, we need to split the flow into two parts to equalize the traffic of the two parts. A part is passed through
Figure 968443DEST_PATH_IMAGE120
E.g. path
Figure 809360DEST_PATH_IMAGE122
And
Figure 286609DEST_PATH_IMAGE123
the other part is passed through
Figure 871174DEST_PATH_IMAGE121
E.g. path
Figure 445374DEST_PATH_IMAGE124
And
Figure 332559DEST_PATH_IMAGE125
. If we can solve the 2-SR problem, we will put the number set
Figure 421738DEST_PATH_IMAGE126
Two subsets can be divided, the sum of each subset being equal. However, this Subset and problem is a classical NP-hard problem, as the relevant principles are explained in detail in the document four "Subset sum NP-complete", according to common general knowledge, which is clearly known to the person skilled in the art. The fourth document can be obtained by a network link http:// www.cs.cornell.edu/courses/cs4820/2018 fa/alternatives/subset sum. The 2-SR problem is NP-hard, as is the n-SR problem.
Since the n-SR problem is np-hard, especially as the topology scale increases, commercial solvers cannot solve quickly. To this end, the present invention proposes a segment routing traffic engineering method based on a limited widest path (RWP) to minimize the maximum link utilization.
The segment routing traffic engineering method based on the limited widest path (RWP) mainly comprises the following steps:
step 1: searching for the Widest Path (WP)
And searching and acquiring SRv6 a widest path in all paths between a source node and a target node corresponding to any flow in the network, where the widest path is a path with a minimum utilization rate in all paths, and the utilization rate of each path in all paths is a maximum utilization rate in all links in each path.
Step 2: the widest path is encoded with a series of segments.
First encoding said widest path with segments in said SRv6 network to obtain an encoded path represented by a combination of segments formed by said segments
And step 3: to meet the maximum segment number constraint, RWP further refines the encoded SR path in step 3.
And judging whether the number of the segments used in the segment combination for representing the coding path is greater than that of the segment of the SRv6 network, if so, recording the hop counts between each segment and an upstream segment in the coding path, comparing the hop counts corresponding to the segments, deleting the segment with the minimum hop count from the segment combination, and returning to the step 3 until the number of the segments in the segment combination does not exceed that of the segment of the SRv6 network.
Before describing the RWP algorithm in detail, the definition of the widest path is given first. Is provided with
Figure 685360DEST_PATH_IMAGE127
Indicating a link
Figure 176384DEST_PATH_IMAGE109
Link utilization of.
Widest path: for the arbitrary stream
Figure 562366DEST_PATH_IMAGE011
Let us order
Figure 748628DEST_PATH_IMAGE128
Representing the corresponding source node
Figure 940575DEST_PATH_IMAGE013
And the corresponding target node
Figure 161472DEST_PATH_IMAGE014
All paths in between, the widest path being the source node
Figure 515093DEST_PATH_IMAGE013
And the target node
Figure 516547DEST_PATH_IMAGE014
The widest path in between. For each corresponding said path
Figure 121972DEST_PATH_IMAGE021
We define the path utilization as the maximum link utilization of all links in the path, i.e., the maximum link utilization
Figure 587588DEST_PATH_IMAGE129
. The widest path is the set
Figure 987476DEST_PATH_IMAGE130
In (1)
Figure 476227DEST_PATH_IMAGE025
Wherein
Figure 9976DEST_PATH_IMAGE131
According to the above definition, the RWP algorithm follows the following three steps.
Step 1: find the widest path.
Figure 939886DEST_PATH_IMAGE011
The widest path of (a) can be found iteratively. We set the accessed list
Figure 900889DEST_PATH_IMAGE027
And an unaccessed list
Figure 549039DEST_PATH_IMAGE028
。
Figure 620900DEST_PATH_IMAGE132
Is shown in
Figure 733213DEST_PATH_IMAGE030
In a sub-iteration from
Figure 740483DEST_PATH_IMAGE013
To
Figure 734984DEST_PATH_IMAGE031
The path utilization of. First of all, initializing
Figure 751481DEST_PATH_IMAGE032
、
Figure 515038DEST_PATH_IMAGE133
、
Figure 21106DEST_PATH_IMAGE134
. In each iteration, we update
Figure 378269DEST_PATH_IMAGE135
Wherein
Figure 57512DEST_PATH_IMAGE136
，
Figure 810661DEST_PATH_IMAGE137
. Then will be
Figure 284368DEST_PATH_IMAGE138
Node with the smallest value
Figure 456723DEST_PATH_IMAGE028
Move to
Figure 549444DEST_PATH_IMAGE027
. When in use
Figure 553172DEST_PATH_IMAGE027
When it is empty, we go through backtracking
Figure 73146DEST_PATH_IMAGE139
Find out from
Figure 529535DEST_PATH_IMAGE082
To be at
Figure 160368DEST_PATH_IMAGE039
The widest path of any node in the tree.
Step 2: and (4) path coding. In finding out
Figure 221865DEST_PATH_IMAGE011
After the widest path we need to represent it using a combination of segments. Let us leave it without loss of generality
Figure 37374DEST_PATH_IMAGE140
To represent
Figure 856426DEST_PATH_IMAGE011
Of the widest path of (1), wherein
Figure 415583DEST_PATH_IMAGE141
、
Figure 331586DEST_PATH_IMAGE142
。
Figure 193363DEST_PATH_IMAGE143
Is on the way
Figure 624344DEST_PATH_IMAGE144
In
Figure 596980DEST_PATH_IMAGE048
And
Figure 633069DEST_PATH_IMAGE046
the sub-path in between. Let
Figure 790381DEST_PATH_IMAGE145
To represent
Figure 584024DEST_PATH_IMAGE048
To
Figure 484984DEST_PATH_IMAGE046
The shortest path between them. If it is not
Figure 375580DEST_PATH_IMAGE146
Then we can use
Figure 579159DEST_PATH_IMAGE048
And
Figure 719153DEST_PATH_IMAGE046
address representation of
Figure 299170DEST_PATH_IMAGE143
。
And step 3: the encoded path is refined. The above steps may result in a path violating the constraint of the maximum number of segments
Figure 841010DEST_PATH_IMAGE147
. Here, an example is given to more clearly illustrate this problem. Suppose we find the widest path at step 1
Figure 543387DEST_PATH_IMAGE149
Then it is represented by 4 segments in step 2, resulting in
Figure 46044DEST_PATH_IMAGE150
. Assuming that the maximum number of stages is fixed to 3, the generated path
Figure 288806DEST_PATH_IMAGE151
Is not feasible. To this end, we assign an attribute to each non-destination segment, which is recorded in
Figure 826098DEST_PATH_IMAGE152
The number of hops between it and the upstream segment. In the case of this example of the present invention,
Figure 496114DEST_PATH_IMAGE153
，
Figure 813963DEST_PATH_IMAGE154
，
Figure 735782DEST_PATH_IMAGE155
. Deletion with lowest
Figure 986635DEST_PATH_IMAGE156
The segments are valued until the constraint is satisfied because the segments have less of a control effect on bypassing the flow over the congested link. Since the maximum number of segments allowed in this example is 3, we only need to delete nodes from the path
Figure 968497DEST_PATH_IMAGE157
To obtain
Figure 570380DEST_PATH_IMAGE158
。
In order to further describe and clarify the algorithmic processes of the process engineering method of the present invention, a specific programming example is provided below, but the process engineering method of the present invention is not limited to be implemented by the following implementation methods:
Require：G=(V,E,W,C) A plurality of streamsf1,f2,…fmThe maximum number of stagesN.
Ensure: segment routing pathp 1 r ,p 2 r ,…,p n r }
1:
Figure 358207DEST_PATH_IMAGE159
2:for all f i ∈{f1,f2,…fm} do
3: p i w =widestpath(G, W′,s i ,d i )
4: p i r =Encodepath(G, p i w , f i )
5: p i r =Refinepath(G, p i w , f i ,n)
6: flow usingf i Along the pathp i r Occupied bandwidth updateW′
7：return{ p 1 r ,p 2 r ,…,p n r }
p i w =widestpath(G, W′,s i ,d i )
1 initializationS=s i ,S′=V-{s i },t=0,B ij (0)=+∞
2:while S′≠Ø do
3:t= t +1
4: for all
Figure 604512DEST_PATH_IMAGE160
5: B ij (t)=min{ B ij (t-1),max{ B ik (t-1),w′ kj }}
6, handleB ij (t) Having a minimum value
Figure 881910DEST_PATH_IMAGE031
FromS′Move toS
7: by backtrackingB sidi To obtainp i w
p i r =Encodepath(G, p i w , f i )
1:p i w =Encodepath(a 0=s i ,a 1 ,a 2 ,,a n-1,a n =d i )
2:j=0,k=1, p i r =[ s i ]
3:while k≠n do
4: if p i w [j,k]= p s [j,k] then
5: k= k +1
6:else
7: j=k-1, p i r =[ p i r →a j ]
p i r =Refinepath(G, p i w , f i ,n)
By observation ofp i w Andp i r find out the pi value of each segment
2: deleting the segment with the minimum value of pi value untilp i r Not exceedingnSegment of
In step 1, we have
Figure 846455DEST_PATH_IMAGE161
Sub-iterations, each iteration requiring execution
Figure 172394DEST_PATH_IMAGE162
A sub-comparison, thus having a time complexity of
Figure 397839DEST_PATH_IMAGE163
. In step 2, the length of the widest path is not greater than
Figure 715644DEST_PATH_IMAGE161
So the algorithm only needs to compare the widest path with the shortest path
Figure 292119DEST_PATH_IMAGE164
Next, the process is carried out. In step 3, the algorithm does not exceed the time complexity
Figure 93853DEST_PATH_IMAGE163
All of
Figure 642646DEST_PATH_IMAGE156
The values are sorted. The time complexity of one iteration can be obtained as
Figure 996267DEST_PATH_IMAGE163
. Due to the fact that
Figure 669825DEST_PATH_IMAGE165
Per flow, total time complexity of the available RWP algorithm is
Figure 665463DEST_PATH_IMAGE166
。
To better contrast the adapted 3-segment or more segment routing of the traffic engineering method provided by the present invention, we performed experiments in four network topologies, including three composite networks (rand20, rand50 and rand80) and one real network (rf 1221). The topologies rand50 and rf1221 are provided by DEFO (an introduction to this particular related knowledge can be obtained from "clear and express forward optimizer" in the link http:// sites. ucluvain. be/info /). The topologies rand20 and rand80 are composite networks with an average degree of 3. We run real streams from DEFO in rand50 and rf1221, while synthetic streams generated by gravity models (m. Roughan, "simple the synthesis of internet traffic," inproc. of ACM sigcomp, October 2005, pp. 93-96) were used in rand20 and rand80 networks. Table I summarizes the details of these topologies. The main indicators of our experiments include MLU and time consumption of the overall network. We compared our algorithm with the Shortest Path (SP) based solution and the results from solving the n-SR model using CPLEX. To this end, in our experiments we changed the maximum number of segments, the network size and the number of flows in the network. We randomly selected streams from the stream data set for 10 runs and found the mean and standard deviation results.
Table I topology and flow information
Figure 740866DEST_PATH_IMAGE167
In topology rand20, we quantify the impact of the maximum number of segments by changing its value on the MLU. As shown in fig. 3, n-SR and RWP achieve better MLU than shortest path. Specifically, the MLU of the 2-SR is half that of the SP, and the 2-RWP can provide nearly 40% performance improvement over the SP. Notably, as shown in FIG. 3, the performance of 3-SR and 4-SR is very close to that of 2-SR. This phenomenon indicates that the 2-SR model can provide an optimal or sub-optimal solution in a small network. Furthermore, the performance of RWP improves with increasing maximum number of stages, and the difference between the result of the RWP algorithm and the theoretical result (n-SR) decreases with increasing number of stages.
In the case of 400 flows, we increase the size of the topology and compare the MLUs for SP, 2-SR and RWP, with the results shown in FIG. 4. In particular, 2-SR achieves the lowest MLU over other methods. In contrast, an SP always has the highest MLU. Our RWP algorithm implements one intermediate MLU but close to 2-SR. FIG. 5 shows the results of MLU when the number of rf1221 streams was varied. Similar to FIG. 4, 2-SR and RWP outperform the SP method. As the flow increases, the MLU of all methods increases.
From the results of the above three experiments, it can be observed that the increase in the segments helps to reduce the MLU in the RWP. The reason for this is that as the number of segments increases, the algorithm will explore more paths and thus hopefully disperse traffic in the network. In addition, given the network size and flow cardinality, the number of paths that the algorithm can explore does not grow as fast as the number of segments. Thus, increasing the number of stages has the marginal effect of reducing the MLU.
Furthermore, by further comparing the time consumption of the SR and RWP in tables II and III, it can be seen that the average execution time of RWP continues to increase as traffic and network size increase. In general, the RWP algorithm provided by the invention is obviously superior to the 2-SR method. As shown in table II, the 2-SR algorithm searches for an optimal path longer and longer (from 2.671s to 488.314s) as the network size increases. In contrast, the time consumption of the RWP algorithm increases only from 1 second to less than 8 seconds as the network size grows. The reason is that the SR model is and the temporal complexity of RWP is O (m × | V | 2). RWP can provide 5-10% more solution in MLU than 2-SR in 10 seconds in most experiments.
Table II run time of 400 streams in four topologies
Figure 468651DEST_PATH_IMAGE168
TABLE III runtime in topology RF1221
Figure 754139DEST_PATH_IMAGE169
In the invention, the existing 2-SR or 3-SR problem is popularized to the situation of n sections, wherein n is more than or equal to 2, after an integer programming model is constructed, the n-SR problem is proved to be np difficult, therefore, a flow engineering method based on the limited widest path section is designed to solve the problem, and experimental results show that the method can provide a solution equivalent to the MLU of the most advanced calculation method while reducing the calculation time.
In addition, the present invention also provides a terminal, which includes a memory and a processor, wherein the memory stores a computer program, and the computer program is executed by the processor to execute any one of the segment routing traffic engineering methods provided by the present invention.
In addition, the present invention also provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to execute any segment routing traffic engineering method provided by the present invention.
While embodiments in accordance with the invention have been described above, these embodiments are not intended to be exhaustive or to limit the invention to the precise embodiments described. Many modifications and variations are possible in light of the above teaching. The embodiments were chosen and described in order to best explain the principles of the invention and the practical application, to thereby enable others skilled in the art to best utilize the invention and various embodiments with various modifications as are suited to the particular use contemplated. The invention is limited only by the claims and their full scope and equivalents.

Claims (10)

1. A segment routing traffic engineering method based on a limited widest path is characterized by comprising the following steps:
step 1: searching and acquiring SRv6 a widest path of all paths between a source node and a target node corresponding to any flow in the network, where the widest path is a path with a minimum utilization rate among all paths, a utilization rate of each path of all paths is a utilization rate of a maximum link among all links of each path, and the maximum link is a link with a maximum utilization rate among all links of each path,
step 2: first encoding said widest path with segments in said SRv6 network to obtain an encoded path represented by a combination of segments formed by said segments,
and step 3: and judging whether the number of the segments used in the segment combination for representing the coding path is greater than that of the segment of the SRv6 network, if so, recording the hop counts between each segment and an upstream segment in the coding path, comparing the hop counts corresponding to the segments, deleting the segment with the minimum hop count from the segment combination, and returning to the step 3 until the number of the segments in the segment combination does not exceed that of the segment of the SRv6 network.
2. The segment routing traffic engineering method according to claim 1, further comprising, before the step 1:
modeling SRv6 networks as undirected graphs
Figure 933592DEST_PATH_IMAGE001
Wherein, in the step (A),
Figure 894595DEST_PATH_IMAGE002
represented as a collection of network routers,
Figure DEST_PATH_IMAGE003
representing physical links between routers, from network node
Figure 231160DEST_PATH_IMAGE004
To a network node
Figure DEST_PATH_IMAGE005
Of a link
Figure 37442DEST_PATH_IMAGE006
All have a given capacity
Figure DEST_PATH_IMAGE007
And weight
Figure 743230DEST_PATH_IMAGE008
To characterize the link
Figure DEST_PATH_IMAGE009
The cost of transmitting one single-bit stream,
and networking said SRv6
Figure 406293DEST_PATH_IMAGE010
In a stream ofiA stream
Figure DEST_PATH_IMAGE011
By using
Figure 696066DEST_PATH_IMAGE012
It is shown that,
Figure DEST_PATH_IMAGE013
is a stream
Figure 837198DEST_PATH_IMAGE011
The source node of (a) is,
Figure 69596DEST_PATH_IMAGE014
is a stream
Figure 106822DEST_PATH_IMAGE011
The destination node of (2) is,
Figure DEST_PATH_IMAGE015
is a stream
Figure 650936DEST_PATH_IMAGE011
The size of (a) is (b),
wherein the flow is in an n-segment SRv6 network
Figure 64600DEST_PATH_IMAGE011
Can follow a path
Figure 948242DEST_PATH_IMAGE016
Routing of, wherein
Figure DEST_PATH_IMAGE017
At most have
Figure 657834DEST_PATH_IMAGE018
The number of the nodes is one,
Figure DEST_PATH_IMAGE019
each sub-path in (a) isnSegment SRv6 is the shortest path between an upstream node and a downstream node in the network.
3. The segment routing traffic engineering method according to claim 2, comprising, at step 1:
order to
Figure 158086DEST_PATH_IMAGE020
Representing flows
Figure 109861DEST_PATH_IMAGE011
Of the source node
Figure 379168DEST_PATH_IMAGE013
And a target node
Figure 758197DEST_PATH_IMAGE014
A set of all paths in between, each path in the set
Figure DEST_PATH_IMAGE021
Taking the maximum link utilization of all links in the path as the path
Figure 949007DEST_PATH_IMAGE022
Path utilization of
Figure DEST_PATH_IMAGE023
By using
Figure 751745DEST_PATH_IMAGE024
Indicating a link
Figure DEST_PATH_IMAGE025
By comparing said link utilization
Figure 406717DEST_PATH_IMAGE026
Determining the widest path as the path in the set
Figure DEST_PATH_IMAGE027
Wherein 1 is less than or equal toj≤n，
Figure 753385DEST_PATH_IMAGE028
。
4. The segment-routed traffic engineering method according to claim 3, characterized in that the widest path is found in the set in step 1 by an iterative method comprising:
setting an accessed list
Figure DEST_PATH_IMAGE029
And an unaccessed list
Figure 431491DEST_PATH_IMAGE030
In combination with each other
Figure DEST_PATH_IMAGE031
Is shown in
Figure 256228DEST_PATH_IMAGE032
The path in the sub-iterationq j Utilization of, said pathq j To be a slave source node
Figure 736013DEST_PATH_IMAGE013
To the network router setVTo (1)jA node
Figure DEST_PATH_IMAGE033
Is detected by the optical sensor (c) and is,
initialization
Figure 253582DEST_PATH_IMAGE034
、
Figure DEST_PATH_IMAGE035
、
Figure 418984DEST_PATH_IMAGE036
，
At each iteration, updating the
Figure DEST_PATH_IMAGE037
Wherein
Figure 47411DEST_PATH_IMAGE038
，
Figure DEST_PATH_IMAGE039
，
Then will be
Figure 411397DEST_PATH_IMAGE040
Node with minimum corresponding value
Figure 67244DEST_PATH_IMAGE030
Move to
Figure DEST_PATH_IMAGE041
，
When in use
Figure 516680DEST_PATH_IMAGE041
Is empty, from
Figure 152060DEST_PATH_IMAGE042
Find out from
Figure 839394DEST_PATH_IMAGE043
To be at
Figure DEST_PATH_IMAGE044
The widest path corresponding to any node in the set.
5. The segment routing traffic engineering method according to claim 4, characterized in that in step 2:
combining the segments
Figure 902027DEST_PATH_IMAGE045
Represents the abovef i The widest path coding path of (1), wherein
Figure DEST_PATH_IMAGE046
、
Figure 573180DEST_PATH_IMAGE047
，
And order
Figure DEST_PATH_IMAGE048
Is on the way
Figure 576034DEST_PATH_IMAGE049
In
Figure DEST_PATH_IMAGE050
And
Figure 852294DEST_PATH_IMAGE051
sub-path between, let
Figure DEST_PATH_IMAGE052
To represent
Figure 882567DEST_PATH_IMAGE053
To
Figure 509857DEST_PATH_IMAGE051
Shortest path therebetween, comparison
Figure DEST_PATH_IMAGE054
And
Figure 283778DEST_PATH_IMAGE052
the size of (1) when
Figure 450098DEST_PATH_IMAGE055
Then use
Figure 120114DEST_PATH_IMAGE050
And
Figure 437963DEST_PATH_IMAGE051
address representation of
Figure 218837DEST_PATH_IMAGE054
，
Through each of said sub-paths represented by segments
Figure 735269DEST_PATH_IMAGE054
Representing the coding pathp i w 。
6. The segment routing traffic engineering method of claim 2, characterized in that the segment routing traffic engineering method is a segment routing traffic engineering methodnGreater than or equal to 2.
7. The segment routing traffic engineering method according to claim 4, characterized in that in step 1, the iterative method is performed
Figure DEST_PATH_IMAGE056
Sub-iterations, each iteration requiring execution
Figure 372924DEST_PATH_IMAGE057
A time complexity of the iterative method of
Figure DEST_PATH_IMAGE058
In the step 2, the length of the widest path is not more than
Figure 505965DEST_PATH_IMAGE056
So the iterative method needs to compare the widest path and the shortest path
Figure 293792DEST_PATH_IMAGE057
In said step 3, the time complexity does not exceed
Figure 166196DEST_PATH_IMAGE058
All of the hop counts of (a) are ordered.
8. The segment-routed traffic engineering method of claim 4, characterized in that, thennThe segment SRv6 network is modeled as:
Figure 178014DEST_PATH_IMAGE059
Figure DEST_PATH_IMAGE060
Figure 798351DEST_PATH_IMAGE061
Figure DEST_PATH_IMAGE062
Figure 717766DEST_PATH_IMAGE063
wherein the constraint equation (3) in the SRv6 network modeling ensures all trafficAre all routed in the SRv6 network and each traffic passes through only a single onenSegment routing path, in constraint equation (4) in said SRv6 network modeling, if flow
Figure 146473DEST_PATH_IMAGE011
Is a through path
Figure DEST_PATH_IMAGE064
Routed, then
Figure 860351DEST_PATH_IMAGE065
To represent
Figure 935361DEST_PATH_IMAGE009
In (A) belong to
Figure 861729DEST_PATH_IMAGE011
And the constraint (4) indicates that the link utilization of each link is less than the maximum link utilization,
Figure DEST_PATH_IMAGE066
representing flows
Figure 738418DEST_PATH_IMAGE011
At the source node
Figure 357618DEST_PATH_IMAGE013
And the target node
Figure 155810DEST_PATH_IMAGE014
The number of possible paths between the two,
Figure 620289DEST_PATH_IMAGE067
represents from
Figure DEST_PATH_IMAGE068
Find h different objects inThe number of the arrays of (2),
Figure 554747DEST_PATH_IMAGE069
representing a source node
Figure 111893DEST_PATH_IMAGE013
And
Figure DEST_PATH_IMAGE070
all the shortest paths between the first and second nodes,
Figure 662960DEST_PATH_IMAGE071
to represent
Figure 196709DEST_PATH_IMAGE009
In that
Figure 454515DEST_PATH_IMAGE070
Appear at
Figure DEST_PATH_IMAGE072
The total frequency of occurrence of (a) is,
Figure 681097DEST_PATH_IMAGE073
representing flows
Figure 719461DEST_PATH_IMAGE011
Whether or not to pass
Figure DEST_PATH_IMAGE074
And (4) path routing.
9. A terminal, characterized in that it comprises a memory and a processor, the memory storing a computer program which, when executed by the processor, performs the segment routing traffic engineering method according to any of claims 1-8.
10. A computer-readable storage medium, in which a computer program is stored which, when being executed by a processor, is adapted to carry out the method of segment routing traffic engineering according to any one of claims 1 to 8.
CN202110549603.7A 2021-05-20 2021-05-20 Segment routing traffic engineering method and device based on limited widest path Active CN112995032B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110549603.7A CN112995032B (en) 2021-05-20 2021-05-20 Segment routing traffic engineering method and device based on limited widest path

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110549603.7A CN112995032B (en) 2021-05-20 2021-05-20 Segment routing traffic engineering method and device based on limited widest path

Publications (2)

Publication Number Publication Date
CN112995032A CN112995032A (en) 2021-06-18
CN112995032B true CN112995032B (en) 2021-08-24

Family

ID=76337057

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110549603.7A Active CN112995032B (en) 2021-05-20 2021-05-20 Segment routing traffic engineering method and device based on limited widest path

Country Status (1)

Country Link
CN (1) CN112995032B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114205289B (en) * 2021-10-22 2023-03-24 清华大学 Traffic engineering calculation method and device considering faults in segmented routing network

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110870261B (en) * 2017-07-07 2022-01-07 华为技术有限公司 PECP segmented routing path segmented label binding extension
US10541923B2 (en) * 2018-02-05 2020-01-21 Ciena Corporation Segment routing traffic engineering based on link utilization
US20220006723A1 (en) * 2018-11-19 2022-01-06 Telefonaktiebolaget Lm Ericsson (Publ) Segment Routing Network
CN112242950B (en) * 2019-07-18 2022-05-17 华为技术有限公司 Method for determining path and related equipment
WO2021021169A1 (en) * 2019-07-31 2021-02-04 Huawei Technologies Co., Ltd Transporting mtnc-id over srv6-enabled dataplane for 5g transport
CN112202670B (en) * 2020-09-04 2022-08-30 烽火通信科技股份有限公司 SRv 6-segment route forwarding method and device

Also Published As

Publication number Publication date
CN112995032A (en) 2021-06-18

Similar Documents

Publication Publication Date Title
CN109150627B (en) Virtual network mapping construction method based on dynamic resource demand and topology perception
CN112738820B (en) Dynamic deployment method and device of service function chain and computer equipment
CN108768736B (en) Optimization method of hybrid service function chain embedding cost
Liu et al. Drl-or: Deep reinforcement learning-based online routing for multi-type service requirements
CN109412963B (en) Service function chain deployment method based on stream splitting
da Silva et al. Evolutionary computation for automatic web service composition: an indirect representation approach
Chattopadhyay et al. A scalable and approximate mechanism for web service composition
CN112995032B (en) Segment routing traffic engineering method and device based on limited widest path
JP5747393B1 (en) Flow aggregation apparatus, method, and program
Bastos et al. A forwarding strategy based on reinforcement learning for Content-Centric Networking
Kuhn et al. Routing schemes and distance oracles in the hybrid model
Shirmarz et al. A novel flow routing algorithm based on non-dominated ranking and crowd distance sorting to improve the performance in SDN
CN114745322B (en) Video flow routing method based on genetic algorithm in SDN environment
CN116455824A (en) Network traffic load balancing method based on reinforcement learning
CN105917621B (en) Method and system for data routing
Perepelkin et al. Research of Multipath Routing and Load Balancing Processes in Software Defined Networks Based on Bird Migration Algorithm
WO2019183962A1 (en) Method for classifying network packet on basis of equal length and equal density segmentation
JP5595342B2 (en) Multiple path search method and apparatus
Qin et al. MCRA: multicost rerouting algorithm in SDN
Qin et al. Minimum cost multi-path parallel transmission with delay constraint by extending openflow
García et al. Multicast routing and virtual network function placement in NFV-SDN networks: A genetic algorithms approach
CN106713165B (en) Method for optimizing load balance in network coding environment
CN108809834B (en) Network topology dispersed short path set calculation method based on path expansion and elimination mechanism
Filippidou et al. Effective and efficient graph augmentation in large graphs
CN115226044B (en) Multicast routing method and system in NFV 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