This application is a National Stage application under 35 U.S.C. § 371 of International Application No. PCT/JP2020/003696, having an International Filing Date of Jan. 31, 2020, which claims priority to Japanese Application Serial No. 2019-020141, filed on Feb. 6, 2019. The disclosure of the prior application is considered part of the disclosure of this application, and is incorporated in its entirety into this application.
The present invention relates to a control apparatus, a control method, and a program.
SR (Segment Routing) has been receiving attention as a network technique for realizing flexible TE (Traffic Engineering) and SFC (Service Function Chaining).
For example, it is possible to induce traffic in a link with a low usage rate by constantly acquiring the bandwidth usage rate of the link and performing control through SR using a traffic influx port (GW (Gateway), etc.) with respect to a large amount of traffic, such as a DDoS (Distributed Denial of Service) attack, or a burst-like increase in traffic resulting from viewing of a specific video or the like.
SR-MPLS, in which SR is applied to an MPLS (Multi-Protocol Label Switching) network, can achieve an improvement in scalability and a simplification of MPLS. In SR-MPLS, labels called SIDs (Segment IDs) are used, and route information is exchanged by adding SIDs to a route advertisement using a protocol (augmented IGP) obtained by augmenting an IGP (Interior Gateway Protocol). Since labels stacked in a packet fulfill the role of a route information table during packet transfer, as long as the SR policy is understood, an intermediate router does not need to hold all route information.
In SR-MPLS, three types of SIDs called Node-ID, Adj-SID (Adjacency-SID), and Peer-SID are used. A Node-ID is defined for each node and is used when transferring to a node in accordance with the shortest route of an IGP. An Adj-SID is defined for each interface and is used when explicitly designating a link to be passed through. A Peer-SID is used for transfer between ASs (Autonomous Systems).
In TE using SR-MPLS, it is envisioned that many labels are stacked in a packet. For example, in a use case such as congestion avoidance, it is necessary to perform TE in which a link through which traffic is to be passed is directly designated. In order to robustly perform route control with respect to a change in link cost resulting from breakdown, it is important to individually designate the links to be subjected to TE. As is evident based on
As a result, it is also conceivable that the number of labels stacked in the packet will exceed the processing limit of the router, and in this case, flexible traffic control using SR is difficult. Also, the overhead increases accompanying an increase in the number of labels stacked in the packet.
Conventionally, a case has been considered in which the number of stacked labels is temporarily reduced by replacing multiple SIDs (labels) with a single BSID (Binding SID) (performing compression) (NPL 1, 2). For example, a controller compresses multiple labels <16003, 16004> into one BSID <40164> and advertises the result in some routers (GW, etc.). When a router in which the BSID <40164> was advertised receives a packet to which the BSID <40164> was added, the router converts (expands) the BSID <40164> into the original list of labels <16003, 16004> and performs transfer in accordance with the expanded SIDs.
However, if BSIDs are to be allocated in units of ASs, many labels will be needed for TE in an AS. Even if an AS is divided into multiple areas and BSIDs are allocated in units of areas, the BSIDs that are to be pushed in an ingress router increase in number accompanying an increase in the number of areas, and therefore multiple labels will be needed.
The present invention proposes a scheme for reducing the number of labels stacked in a packet in SR.
A control apparatus according to an aspect of the present invention is
Also, a control method according to an aspect of the present invention is
Also, a program according to an aspect of the present invention causes a computer to function as the units of the above-described control apparatus.
The present invention makes it possible to reduce the number of labels stacked in a packet in SR.
Hereinafter, embodiments of the present invention will be described with reference to the drawings.
Logical Configuration of Network
First, a logical configuration of a network for illustrating an embodiment of the present invention will be described with reference to
A control apparatus 100 called a controller can acquire NW information such as topology information from the ASBRs and ABRs, and can cause the ASBRs and ABRs to execute processing for designation (also called policies). In this manner, a router that can act as a peer to the controller 100 and can be controlled by the controller 100 is referred to as an anchor node. In the following embodiment, an example will be described in which the anchor nodes are ASBRs and ABRs. The anchor nodes are not limited to the ASBRs and the ABRs, and any communication apparatus in the network may be designated as an anchor node. Also, it is not necessary for all of the ASBRs and ABRs to be anchor nodes. The anchor nodes are designated in advance by an operation manager of the network. The controller 100 may also manage one AS, or may manage multiple ASs.
For example, it is envisioned that TE is performed in which the links to be passed through are designated individually in order to make changes in the link cost resulting from the occurrence of breakdown robust. For example, in the network shown in
Here, a scheme for replacing multiple SIDs (labels) with a single BSID (NPL 1, 2) will be considered with reference to
First, a case is envisioned in which BSIDs are allocated in units of ASs. For example, in
Next, it is envisioned that the BSIDs are allocated in units of areas. For example, the SIDs used in TE in the sub-areas (1) of AS1 in
Accordingly, a label count reduction scheme according to which the processing limit of the router is not exceeded is needed.
Label Count Reduction Scheme
A scheme in which a restriction is satisfied when route information constituted by L SIDs is provided (the node that satisfies an upper limit U (U<L) of the label push count and in which the BSID can be expanded being an anchor node) will be described with reference to
In step S101, the controller 100 acquires the upper limit U of the label push count of each router.
The controller 100 acquires the topology information of each AS/area from the anchor nodes (ASBRs and ABRs) using a protocol such as BGP-LS (BGP-Link State). Information (Maximum SID Depth) on the number of pushable labels that are advertised in IGP is included in the topology information. The controller 100 determines the minimum value of the Maximum SID Depths of the routers in the network managed by the controller 100 as U. Note that U does not need to be determined uniquely in the network, and can also be determined for each area. If U is to be determined for each area, the controller 100 may determine the minimum value Ui of the area i for each area and use the minimum value Ui in the calculation of step S105. U may also be a value determined in advance based on the specification of the communication apparatus used in the network. Note that since U is the restriction condition of route control (SR), any value may be used, as long as it is in a range that does not exceed the processing limit of the communication apparatus on the route determined in step S102 below.
In step S102, the controller 100 creates route information (SR-TE path) that is constituted by L SIDs based on a required condition of TE.
The required condition is determined according to a user requirement, and examples thereof include a TE requirement such as delay and avoidance of a specific link, a service chaining requirement, and a target requirement. Based on the required condition, the controller 100 specifies SIDs of nodes, links, and peers to be passed through, and creates an SR-TE path.
The SIDs included in the SR-TE path may also be specified in real time based on information held by the controller 100. Anchor nodes such as ASBRs and ABRs are designated in advance. The controller 100 holds a correspondence relationship between the address of the ASBR and the AS number. Also, in step S101, the controller 100 holds SIDs (Node-IDs, Adj-SIDs, Peer-SIDs) acquired from the ASBR using a protocol such as BGP-LS. The controller 100 can create a list of SIDs constituting an SR-TE path using these SIDs.
In step S103, the controller 100 specifies the SIDs of the anchor nodes (H_1, . . . , H_M) that can expand BSIDs from among the SIDs constituting the SR-TE path.
Here, among the SIDs constituting the SR-TE path generated in step S102, it is possible to specify the SIDs of the anchor nodes (H_1, . . . , H_M) that can expand BSIDs from the information acquired and stored by the controller 100 using a protocol such as BGP-LS, and information used for specifying the SIDs in step S102.
In step S104, the controller 100 divides a list of L SIDs into M BSIDs such that the final label of each BSID is the SID of an anchor node (H_1, . . . , H_M).
Since the anchor node has a role of expanding the next BSID according to an instruction from the controller 100, the controller 100 divides the list of SIDs constituting the SR-TE path in units of anchor nodes (H_1, . . . , H_M). Specifically, when an SID included in the SR-TE path matches an SID of an anchor node, the controller 100 performs division immediately after that SID. Also, since it is understood that an ASBR serving as an anchor node is present when an SID included in the SR-TE path matches a Peer-SID, the controller 100 performs division immediately after that SID. The controller 100 sets a BSID for each divided path.
In step S105, the controller 100 determines whether or not the label push count in each anchor node (H_1, . . . , H_{M−1}) falls below U, and calculates the excess number R_i=U−(label push count of H_i).
A portion of the list of SIDs can be compressed into a BSID by dividing the list of SIDs into BSIDs. The labels <1, 2, 3, H_1> obtained by performing division in
The labels that are to be pushed by the anchor nodes H_0, H_1, and H_2 after the labels shown in
In step S106, in the case where R_i> the label push count of H_1 is satisfied, the controller 100 reduces the number of BSIDs by merging adjacent BSIDs. Note that the label push count resulting from merging needs to be U or less. The controller 100 updates R_i after the BSIDs are merged.
If it is envisioned that the labels <4, H_2> that are pushed by H_1 and the labels <5, 6, H_3> that are pushed by H_2 in
In step S107, the controller 100 executes collapsing of the list of SIDs using a BSID in which R_i>0 is satisfied.
An existing technique (NPL 3, 4) can be used for the collapsing of the list of SIDs. Specifically, it is possible to confirm whether or not it is possible to replace some of the paths in the list of SIDs with Node-IDs serving as endpoints, and if replacement is possible, the list of SIDs can be collapsed. For example, it is envisioned that three Adj-SIDs have been used to designate a link for passing from a Headend node to a Tailend node shown in
In step S108, if there is no solution in step S107 above (i.e., if R_i>0 is satisfied), the controller 100 prevents the upper limit of the label push count from being exceeded by terminating the SR-TE path once in the segment of the BSID that exceeds the upper limit U of the label push count.
For example, as shown in
Alternatively, the controller 100 may also redefine the list of BSIDs that were compressed for each anchor node as a new BSID. For example, by redefining BSID2 and BSID3 in
After the labels are compressed according to the above-described procedure, the controller 100 uses a protocol such as PCEP (Path Computation Element Protocol) or Netconf to distribute the SR policy (information for expanding the SR-TE path and the BSIDs included in the SR-TE path) to the anchor nodes on the route.
Functional Configuration of Controller
The NW information acquisition apparatus 110 acquires NW information such as topology information from anchor nodes and the like in the managed network. The SR route creation apparatus 120 creates route information to be distributed as an SR policy based on the NW information acquired by the NW information acquisition apparatus 110 and a required condition from an application. This route information is obtained by expressing labels of communication apparatuses or links through which to pass in the form of a list of SIDs. The route compression apparatus 130 compresses labels of route information created by the SR route creation apparatus 120 into a BSID. The compression of the labels is performed in accordance with the flowchart shown in
The information management unit 112 uses a protocol such as BGP-LS to acquire link state information of an IGP and route information of BGP, IP, or the like from a communication apparatus in the managed network via the communication unit 111. The communication apparatuses from which the link state information and the route information are acquired are mainly anchor nodes, but the information may also be acquired from other communication apparatuses. Here, the Maximum SID Depth advertised in the IGP is acquired, and the upper limit U of the label push count is determined. The upper limit U of the label push count is transmitted to a later-described route compression apparatus 130 and is stored in the compressed information table 135. The information management unit 112 performs computation such as sorting in which the label information is used as a key, and thereafter stores the acquired link state information and route information in the link state information table 115 and the route information table 116. Note that when the link state information is stored in the link state information table 115, the link state information computation unit 113 performs computation of the link usage rate or the like, and thereafter stores the link state information in the link state information table 115 in a predetermined format. The information management unit 112 determines the transmission interval and the data content to be transmitted based on the data transmission information table 114, and transmits the data in the link state information table 115 and the route information table 116 to the SR route creation information 120 via the communication unit 111. At this time, the information management unit 112 may also transmit all of the information every time to the SR route creation apparatus 120 in accordance with the content of the data transmission information table 114, and may also transmit only information indicating a difference from the previous instance of transmission.
The SR route creation unit 122 determines the route information of SR based on the policy information determined by the application (designation of nodes through which to pass, or the like, avoidance of a specific link, etc.), and the NW information received at a certain interval from the NW information acquisition apparatus 110 via the communication unit 121. This corresponds to step S102 of
The anchor nodes are designated in advance, and the anchor node information extraction unit 132 acquires information such as the BGP table from the NW information acquisition apparatus 110 via the communication unit 131, extracts information on the predesignated anchor nodes from the acquired information, and stores the extracted information in the anchor node information table 134.
The compression computation unit 133 acquires the route information of SR from the SR route creation apparatus 120 via the communication unit 131. The compression computation unit 133 divides the labels included in the route information in units of anchor nodes. At this time, the label information of the anchor nodes that can expand the BSIDs is acquired from the anchor node information table 134 and division of the route information is performed. This corresponds to steps S103 and S104 of
The information management unit 142 acquires compressed route information from the route compression apparatus 130 via the compression unit 141 and stores the acquired route information in the compressed route management table 143. The information management unit 142 transmits the SID list of the compressed route management table 143 to the ingress router and distributes the content of the BSID to the relevant anchor node via the communication unit 141. For example, in the case of SR policy 2, the information management unit 142 transmits the SID list of the route to the ingress router, distributes the information <10, 20, 30> for expanding the BSID:100 to the router A, and distributes the information <16005> for expanding the BSID:200 to the router B. If the BSIDs have been merged into BSID′, the information management unit 142 distributes the information for expanding the BSID′ to the corresponding anchor node. For example, in the case of SR policy 2′, the information management unit 142 distributes the information <10, 20, 30, 16005> for expanding the BSID′:100 to the router A. Note that when the route is terminated partway, the information management unit 142 transmits the new route information to the anchor node that terminated the route via the communication unit 141. A communication protocol such as PCEP can be used as the method for distributing the route.
In step S201, the NW information acquisition apparatus 110 acquires the NW information such as the topology information from a communication apparatus in the managed network. In step S202, the NW information acquisition apparatus 110 processes the acquired NW information and transmits the result to the SR route creation apparatus 120 in step S203. In step S204, the SR route creation apparatus 120 creates the route information of SR taking into account a requirement from an application and NW information, and transmits the result to the route compression apparatus 130 in step S205. In step S206, the route compression apparatus 130 creates a BSID by specifying the SIDs of the anchor nodes from among the SIDs included in the candidate routes and dividing the SIDs. Thereafter, the route compression apparatus 130 determines whether or not the processing limit of the router is exceeded during transmission through BSID, and performs merging of the BSIDs and review of the route information according to the determination result. In step S207, the route compression apparatus 130 transmits the compressed route to the SR policy distribution apparatus 140. In step S208, the SR policy distribution apparatus 140 registers the compressed route and transmits the compressed route to the communication apparatus in step S209.
As described above, according to the embodiment of the present invention, a router that is suitable for expansion of the BSID can be selected from among the routers (anchor nodes) that can expand the BSID, and the number of labels that have been expanded can be suppressed to the processing limit of the router or less. In the method of the embodiment of the present invention, after dividing segments in units of anchor nodes into BSIDs, merging can be performed in view of an insufficient state of the label count. This makes it possible to determine which router is to be set as an anchor node. Also, the number of instances of terminating the SR-TE path can be reduced.
Also, since the label count can be reduced, flexible TE can be realized also in an environment in which many labels are required while the load on a router and overhead are suppressed.
If the BSIDs are allocated in units of ASs as shown in
As shown in
As shown in
If the BSIDs are allocated in units of ASs as shown in
As shown in
As shown in
Example of Hardware Configuration
Supplement
Although an apparatus according to the embodiment of the present invention has been described with reference to functional block diagrams for the sake of convenience in the description, the apparatus according to the embodiment of the present invention may also be realized using hardware, software, or a combination thereof. For example, the embodiment of the present invention may also be realized using a program for causing a computer to realize the functions of the apparatus according to the embodiment of the present invention, a program for causing a computer to execute the procedures of the method according to the embodiment of the present invention, and the like. The functional units may also be used in combination according to need. The method according to the embodiment of the present invention may also be realized in an order different from the order shown in the embodiment.
Although a method for reducing the number of labels stacked in a packet in SR was described above, the present invention is not limited to the above-described embodiment and can be modified and adapted in various ways within the scope of the claims.
Number | Date | Country | Kind |
---|---|---|---|
2019-020141 | Feb 2019 | JP | national |
Filing Document | Filing Date | Country | Kind |
---|---|---|---|
PCT/JP2020/003696 | 1/31/2020 | WO |
Publishing Document | Publishing Date | Country | Kind |
---|---|---|---|
WO2020/162361 | 8/13/2020 | WO | A |
Number | Name | Date | Kind |
---|---|---|---|
20120069740 | Lu | Mar 2012 | A1 |
20170064717 | Filsfils et al. | Mar 2017 | A1 |
20180198706 | Ceccarelli | Jul 2018 | A1 |
20200213223 | Peng et al. | Jul 2020 | A1 |
Number | Date | Country |
---|---|---|
109218189 | Jan 2019 | CN |
WO 2017198319 | Nov 2017 | WO |
WO 2018033769 | Jun 2019 | WO |
Entry |
---|
Filsfils et al., “Segment Routing Policy Architecture draft-ietf-spring-segment-routing-policy-02.txt,” SPRING Working Group Internet-Draft, Oct. 22, 2018, 33 pages. |
Guedrez et al., “Label Encoding Algorithm for MPLS Segment Routing,” dated Mar. 2017, retrieved from URL <https://www.researchgate.net/publication/311607490_Label_Encoding_Algorithm_for_MPLS_Segment_Routing>, 7 pages. |
Sivabalan et al., “Carrying binding label/segment-ID in PCE-based Networks, draft-sivabalan-pce-binding-label-sid-05,” PCE Working Group Internet-Draft, Oct. 19, 2018, 12 pages. |
Tantsura et al., “The critical role of Maximum SID Depth (MSD)hardware limitations in Segment Routing ecosystem and how to work around those,” NANOG71, Oct. 2, 2017, retrieved from URL <https://pc.nanog.org/static/published/meetings/NANOG71/1424/20171004_Tantsura_The_Critical_Role_v1.pdf>, 21 pages. |
Togari et al., “A study of label pushing reduction mechanism for SR-MPLS,” IEICE General Conference Lecture Proceedings, Mar. 5, 2019, Communication 2:p. 82, 3 pages (With English Translation). |
Number | Date | Country | |
---|---|---|---|
20220131796 A1 | Apr 2022 | US |