CN114222007B

CN114222007B - Hybrid cloud communication method and system

Info

Publication number: CN114222007B
Application number: CN202210159569.7A
Authority: CN
Inventors: 王�锋
Original assignee: Beijing Lingyun Chuangxiang Technology Co ltd
Current assignee: Beijing Lingyun Chuangxiang Technology Co ltd
Priority date: 2022-02-22
Filing date: 2022-02-22
Publication date: 2022-07-01
Anticipated expiration: 2042-02-22
Also published as: CN114222007A

Abstract

The invention relates to the technical field of hybrid cloud network communication, in particular to a hybrid cloud communication method and a system, the method comprises the following steps of S1, establishing a double-layer NPN architecture, wherein the double-layer NPN architecture comprises a central network controller, a local network controller, a content router and an edge router, and the central network controller is positioned at the upper layer of a network and is used for storing information of all NDN network subdomains in the lower layer of the network covered by the upper layer of the network and performing global forwarding; the local network controller is positioned at a network lower layer and used for managing a content router and an edge router in the NDN network subdomain, the content router performs content storage and communication interface management in the NDN network subdomain, and the edge router is used for realizing interconnection between the NDN network subdomain and other NDN network subdomains; and S2, forwarding the interest packet and the data packet based on the content storage table, the pending interest table and the forwarding table information base in the central network controller.

Description

Hybrid cloud communication method and system

Technical Field

The invention relates to the technical field of hybrid cloud network communication, in particular to a hybrid cloud communication method and system.

Background

Cloud computing is divided into two cases, public cloud and private cloud. Public clouds provide flexible computing services, but stability and data security have been challenged; private clouds provide secure data storage for computing, but lack flexibility. The enterprise user prefers to store the data in the private cloud for safety, but simultaneously hopes to obtain the computing resources of the public cloud to reduce the cost, and in the situation, the hybrid cloud can fuse the public cloud and the private cloud to obtain the best effect, so that the efficiency and the cost are optimized, and the purposes of saving money and being safe are achieved.

In the prior art, an NDN network architecture is usually adopted as a hybrid cloud technology architecture to realize network communication, and the NDN has two roles, namely a consumer (content) and a producer (producer); two types of packets are provided, namely an interest packet and a data packet; the contents are identified by name. If a consumer wants to request a certain content, an interest packet (interest packet) with a corresponding name is produced and forwarded through a network to a node storing the content, and after receiving the interest packet, the node returns a data packet (data packet) carrying the corresponding name along the reverse path of the interest packet. Thus, the NDN network employs a receiver-driven PULL mechanism (PULL). The structures of the Interest packet and the Data packet are shown in fig. 1, the structure of the Interest packet includes a name, a filter, a random number and a rule, and the structure of the Data packet includes a name, original Data, content and a signature.

Each node in the NDN maintains three modules, which are a Content Store (CS), a Pending Interest Table (PIT), and a Forwarding Information Base (FIB), respectively. After receiving the interest packet, the router firstly checks whether the CS caches corresponding content, if so, the router directly replies a DATA packet, if not, the router searches PIT, if so, the router only records an interface receiving the interest packet without forwarding, and if not, the router forwards the interest packet according to a forwarding strategy of FIB, adds corresponding PIT entries, or discards the interest packet according to related information. After receiving the Data packet, the router forwards the Data packet according to the information of the corresponding entry in the PIT, and because the content request is met at the moment, the router deletes the corresponding entry in the PIT after the Data packet is forwarded, and determines the content stored in the CS according to the cache strategy and the related substitution mechanism.

Therefore, in the traditional NDN architecture, forwarding of an interest packet depends on a forwarding information table (FIB), the content of the FIB is maintained in a flooding mode, updating is performed on the basis of the whole network in each updating of the FIB, the updated data volume is large, the cost is high, the algorithm is complex, the updating time is more consumed, and the problem that the needed content cannot be found in the FIB due to updating delay of the FIB exists, and the router directly discards the interest packet to cause communication failure exists; in addition, in the NDN architecture, after the node receives the interest packet, the node returns a data packet carrying a corresponding name along the reverse path of the interest packet, namely, the node returns the original path. However, when the original path is blocked or the path is interrupted, the data packet is directly discarded, which may cause a communication failure.

Disclosure of Invention

The invention aims to solve the problems of large updated data volume, high cost, complex algorithm and more updating time consumption caused by updating in the whole network by adopting a flooding mode of FIB, and provides a hybrid cloud communication method and a system.

In order to achieve the above purpose, the invention provides the following technical scheme:

a hybrid cloud communication method, comprising the steps of:

s1, establishing a double-layer NPN architecture, wherein the double-layer NPN architecture comprises a central network controller, a local network controller, a content router and an edge router, wherein,

the central network controller is positioned at the upper layer of the network and is used for storing the information of all NDN network subdomains in the lower layer of the network covered by the upper layer of the network and performing global forwarding;

the local network controller is positioned at a network lower layer and used for managing a content router and an edge router in the NDN network subdomain, the content router performs content storage and communication interface management in the NDN network subdomain, and the edge router is used for realizing interconnection between the NDN network subdomain and other NDN network subdomains;

s2, forwarding the interest packet and the data packet based on the content storage table, the pending interest table, and the forwarding table information base in the central network controller, specifically including the following steps:

s21, according to the information in the interest packet, searching a corresponding data packet in the content storage table, if the corresponding data packet is found, forwarding the data packet and the interest packet through the local network controller according to the request port corresponding to the interest packet recorded in the pending interest table and the path information stored in the forwarding table information base by the data packet; otherwise, go to step S22;

s22, checking the path information in the interest package,

if the interest packet carries the path information, the forwarding table information base is updated according to the path information in the interest packet, and then the data packet is found and forwarded according to the path information; if the interest packet does not carry the path information, acquiring a path from the request node to the target node, adding the generated path information into the interest packet, and then finding and forwarding the data packet according to the generated path information.

As a preferred embodiment of the present invention, the path from the requesting node to the target node is obtained by looking up a local content forwarding table and a local neighbor table in the local network controller, where the local content forwarding table records a content name and a domain number storing the content stored in a sub-domain of the NDN network, and the local neighbor table stores a number of a border router of an adjacent domain.

As a preferred embodiment of the present invention, when the forwarding table information base is updated according to the path information, if there is no entry corresponding to the interest packet in the forwarding table information base, the forwarding table information base requests the local controller to update according to the flooding range and according to the preset frequency, and the local controller actively pushes the update according to the update requesting information.

As a preferable aspect of the present invention, the preset frequency is 30 seconds per time.

As a preferred scheme of the present invention, when a communication failure occurs during forwarding or backtracking, the present invention further comprises the following steps:

the local network controller marks the communication fault path, selects new path information from the path without communication fault, stores the new path information into the pending interest table, and forwards the data packet according to the new path information.

As a preferred solution of the present invention, the local network controller marks the communication failure path through a local communication interface table LIT, where the local communication interface table LIT maintains a communication interface owned by a content router in the NDN network sub-domain and a state of the communication interface, where the state of the communication interface includes: normal idle, active and fault.

As a preferred scheme of the present invention, the local network controller marks a communication failure path, selects new path information from a path in which a communication failure does not occur, stores the new path information in an to-be-determined interest table, and forwards a data packet according to the new path information.

Based on the same conception, the hybrid cloud communication system comprises a central controller, wherein the central controller stores a content storage table, an undetermined interest table and a forwarding table information base, and is used for executing the following steps: the method comprises the steps that a corresponding data packet is searched in a content storage table according to information in an interest packet, if the corresponding data packet is searched, the data packet conducts forwarding of the data packet and the interest packet according to a request port corresponding to the interest packet recorded in an undetermined interest table, path information needed for forwarding is stored in a forwarding table information base, and the method further comprises a local controller.

As a preferred embodiment of the present invention, the local controller stores a local content forwarding table LFT and a local neighbor table LNT, the path calculated from the requesting node to the destination node is found in the local controller by looking up the local content forwarding table LFT and the local neighbor table LNT, the local content forwarding table LFT records a content name stored in the local NDN domain and a domain number storing the content, and the local neighbor table LNT stores a number of a border router ER of an adjacent domain.

As a preferred scheme of the present invention, the local controller further stores a local communication interface table LIT, when a communication fault occurs during forwarding or backtracking, the local controller marks a communication fault path in the local communication interface table LIT, selects new path information from paths in which the communication fault does not occur in the local communication interface table LIT, stores the new path information in an undetermined interest table, and forwards a data packet according to the new path information.

Compared with the prior art, the invention has the beneficial effects that:

1. the invention breaks through the prior art that a network structure is based on one layer of NPN, changes the flow of realizing interest packet and data packet forwarding through three tables of CS, PIT and FIB, adopts a two-layer NPN network structure of a central controller and a local controller, and has the advantages that a control layer and a forwarding layer are arranged in a layered mode, an NDN network domain table and a global FIB table of the central controller are specially used for storing paths, a local controller is specially used for forwarding, the local controller forwards the data packet to an adjacent sub-network only according to an instruction without considering the paths, and the complexity of updating the FIB table is reduced through the cooperation of the central controller and the local controller.

2. In the process of forwarding the interest packet and the data packet by adopting the method and the system, when communication fails due to updating of FIB, if no path information is stored in the information in the interest packet, the path from the request node to the target node is calculated, and the generated path information is added into the interest packet; if the information in the interest packet stores the path information, the forwarding table information base is updated according to the path information, the data packet corresponding to the interest packet is found according to the updated forwarding table information base, and the data packet and the interest packet are traced back. The method provides a solution when the content sought by the interest package cannot be found in the FIB, and reduces the possibility of communication failure.

3. The method calculates the path from the request node to the target node, ensures that the backtracking of the interest packet and the data packet always has path information, fully utilizes various communication modes among multiple clouds, utilizes various transmission paths as far as possible to send data, fully utilizes transmission bandwidth, and improves the success rate and efficiency of communication.

4. In the method of the invention, a local communication interface list LIT is also used, the list stores the communication interface owned by the CSR in the domain and the state of the interface, when the communication link fails, the state marking is carried out on the corresponding line, which is beneficial to quickly finding out the available path through the marking information, and the data packet can not be lost.

Drawings

Fig. 1 is a secondary network architecture based on a central network controller and a local controller in embodiment 1 of the present invention.

Fig. 2 is a customized structure of the local content forwarding table LFT, the local communication interface table LIT, and the local neighbor table LNT in embodiment 1 of the present invention.

Fig. 3 is a flowchart of an interest packet forwarding mechanism in embodiment 1 of the present invention.

Fig. 4 is a data packet processing flow in embodiment 1 of the present invention.

Detailed Description

The present invention will be described in further detail with reference to test examples and specific embodiments. It should be understood that the scope of the above-described subject matter is not limited to the following examples, and any techniques implemented based on the disclosure of the present invention are within the scope of the present invention.

Example 1

On the basis of the original NPN architecture, a network architecture is changed into a two-layer architecture based on a central network controller and a local network controller, the network structure is shown in figure 1, the central network controller belongs to a control plane and is positioned in a central cloud, and the central network controller is used for storing complete information of all NDN network domains in a network and making a global forwarding decision, so that the network management capability and robustness can be improved. And the local network controller belongs to a control plane and is positioned in a non-central cloud. For managing content routers (CSRs) and Edge Routers (ERs) of a local network. Each NDN network domain has a home network controller. And a two-layer architecture is adopted, and the newly added central controller is used for storing more alternative paths in order to have a higher-layer database. Under the architecture of the Internet of things, multiple clouds and a mixed cloud, more than one path can be accessed between two hosts. Meanwhile, the management flexibility can be realized by utilizing the advantages of a Software Defined Network (SDN). For example, in the current single plane NDN, after a node receives an interest packet, a data packet carrying a corresponding name is returned along a reverse path of the interest packet, that is, "return on original path". However, when the original path is blocked or the path is interrupted, the data packet is directly discarded, which may cause a communication failure. However, with a central controller with a higher view angle, when the original route is blocked or the route is interrupted, the local controller can be queried about other alternative content routers with the content label. If the local network controller does not have an alternative path, the local network controller will attempt to query the central network controller for updates.

In addition, when a line link is interrupted or blocked, the local network controller where the link is located will know the message through the interface table, and after updating its own table, synchronize the message to the central network controller, and the central network controller will also update the message.

Among them, the content router CSR and the edge router ER belong to the forwarding plane. The content router can complete node content storage and interest packet forwarding, and manage the storage content and the communication interface in the NDN domain. The border router is used for realizing the interconnection of the local domain and other domains, and the border router also stores FIB information of other NDN domains issued by the central controller.

Further, the central controller CCGW is divided into two tables:

NDN network domain table (NRT): the method is used for storing the noun of each network domain and the name of the corresponding local controller LCGW, and maintaining the online and offline information of each NDN domain.

And the global FIB table is used for caching the path of the currently effective CS content. FIBs are used in the forwarding process of packets of interest, like routing tables in IP networks.

The local controller LCGW is functionally divided into two modules, each of which has a plurality of tables:

the content management module comprises a local content forwarding table LFT, a local communication interface table LIT and a local neighbor table LNT. The LFT mainly records a name of a content stored in the local NDN domain, a domain number in which the content is stored, and a tag record table including the content managed by the local network controller. The names of all content router ids are recorded. Example (a): content labeled Alpha may be obtained via content router 2/7/5.

And the local neighbor table LNT exists on the content router in the same region. The other information used to indicate which area a content label came from, along with which edge router this content came from, and some necessary to maintain the neighbor routers, are recorded in the neighbor table. Example (c): content labeled Gama comes from the R1 area and is learned from the border router ER 2. The local communication interface table LIT stores the numbers of the border routers ER of two adjacent domains, and maintains the classmate interfaces owned by the CSRs in the domain and the states of the interfaces, wherein state 1 indicates normal idle, state 0 indicates working, and state 2 indicates failure. The method is existed on a router with an external interface and used for storing link information such as bandwidth and line type. One host hostA, for example, in the R1 region, needs to look for Delta content located in hostB in the R2 region. The forwarding path is: hostA — query Delta location, get content router id. The interest packet is sent to the content router, and the content router queries a forwarding table to acquire the detailed information of the content: the home area and the border router that reaches this area. And forwarding the interest packet to a boundary router, wherein the boundary router inquires link information destined for the area, and selects wan type networks again according to the priority, preferably the lan type networks.

Examples of local content forwarding tables LFT, local communication interface tables LIT and local neighbor tables LNT are shown in fig. 2.

And a route management module. When a forwarding failure condition of an interested flow of a certain CSR occurs in the local NDN domain (corresponding to state 2 of the LIT), the module performs calculation according to the network information and the node information in the content maintenance module, so as to provide a solution for the condition of the node, and the calculation steps and the solution are introduced in the third section of forwarding mechanism.

The flow chart of the interest packet forwarding mechanism is shown in fig. 3, and the flow chart of the data packet processing is shown in fig. 4, and the method comprises the following steps:

1. the interest package is generated by the application layer, e.g. the administrator marks an application as "content", the data tables of this application will all be recorded in the CS.

2. In the NDN forwarding mechanism based on the SDN, when a router CSR receives an interest packet, the process is performed step by step according to the following steps:

1) and searching for PIT. If the content item exists, adding a request port, returning when the data table arrives, otherwise, constructing a new item of the word name by the PIT, adding the port information of the arrival end into the PIT, and then turning to the step 2).

2) And searching a CS table, if the content data exists, packaging the data, backtracking through port information recorded by the PIT, deleting corresponding entries in the PIT, discarding interest packets, and finishing backtracking of the data packets. If no content data exists, go to step 3).

3) Checking the path information in the interest packet, if the interest packet carries the path information, updating the FIB according to the path information, then forwarding the data table according to the path information, and going to the step 7), and if the interest packet does not contain the path information, going to the step 4).

4) The router looks up the FIB. Go to step 5) if there is a matching entry, otherwise go to step 6).

5) The FIB requests updates from the local controller LCGW at a certain frequency (e.g., 30 seconds) based on the flooding horizon, and the local controller LCGW accepts active push updates from the central controller CCGW. If the FIB updating is started, the step 6) is carried out, otherwise, the interest packet is forwarded according to the content of the FIB, and the step 7) is carried out.

6) The router generates a path request to the local controller LCGW, the local controller LCGW searches the storage position of the content or the position of the edge router, then calculates the optimal path from the request node to the target node, returns the whole path information to the request node, and the request node adds the path information to the head of the interest packet and changes the path mark field. Finally, return to step 3).

7) If the interest packet is successfully forwarded, ending the forwarding, otherwise returning to the step 1)

In addition, in the conventional NDN network, an arrival port of an interest packet is recorded in a certain router node PIT, and all data packets can be returned "as they are" according to a port number of the PIT "when arriving at the node, but the data packets are discarded when a path is blocked or interrupted. In the system architecture proposed herein, the local controller LCGW performs unified nanotube management for such scenarios, stores all paths, and greatly reduces the packet discard probability. The processing flow of the data packet is as follows:

1) and searching for PIT. If the PIT has no requested content item, the data is not the content required by the node, the data is directly discarded, the forwarding is finished, and if the data exists, the step 2) is carried out.

2) Looking up the CS table, if the content item does not exist in the CS, the node updates the content, then sends the updated CS to the local controller LCGW for updating the LFT, and goes to step 3), if the CS has the item, the data packet is discarded.

3) And the node forwards the data packet according to the corresponding interface data in the PIT, and if the forwarding is successful, the step 5) is carried out, and if the forwarding process has unexpected conditions such as path blocking or path interruption, the step 4) is carried out.

4) The node reports the path interruption information to the local controller LCGW, the local controller LCGW updates the LIT, the local controller LCGW checks other paths stored by the nodes at two ends of the path according to the interface table LIT after the update is completed, then an optimal solution is calculated from the current remaining path, new path information is returned to the request node for updating the PIT, and then the step 5 is carried out, if no additional communication port exists between the two nodes, the local controller returns an instruction to enable the node router to discard the data packet, and the forwarding is finished.

5) And after the node successfully forwards the data packet, deleting the corresponding entry in the PIT, and finishing the forwarding.

Claims

1. A hybrid cloud communication method, comprising:

s22, checking the path information in the interest package,

if the interest packet carries the path information, the forwarding table information base is updated according to the path information in the interest packet, and then the data packet is found and forwarded according to the path information; if the interest packet does not carry the path information, acquiring a path from the request node to the target node, adding the generated path information into the interest packet, and then finding and forwarding the data packet according to the generated path information;

the local controller stores a local content forwarding table LFT and a local neighbor table LNT, the path from the requesting node to the target node is obtained by searching the local content forwarding table and the local neighbor table in the local network controller, the local content forwarding table records the name of the content stored in the sub-domain of the NDN network and the domain number for storing the content, and the local neighbor table stores the number of the boundary router of the adjacent domain;

when the forwarding table information base is updated according to the path information, if the forwarding table information base does not have the entry corresponding to the interest packet, the forwarding table information base requests the local controller to update according to the flooding range and the preset frequency, and the local controller actively pushes and updates according to the request updating information.

2. The hybrid cloud communication method of claim 1, wherein said predetermined frequency is 30 seconds per time.

3. The hybrid cloud communication method of any one of claims 1-2, wherein a communication failure occurs when forwarding or backtracking, further comprising the steps of:

4. The hybrid cloud communication method of claim 3, wherein the local network controller marks the communication failure path through a local communication interface table LIT, where the local communication interface table LIT maintains the communication interfaces owned by the content routers in the NDN network sub-domain and the states of the communication interfaces, where the states of the communication interfaces include: normal idle, active and fault.

5. The hybrid cloud communication method of claim 3, wherein the local network controller marks a communication failure path, selects new path information from a path in which a communication failure does not occur, stores the new path information in a pending interest table, and forwards a packet according to the new path information.

6. A hybrid cloud communication system comprises a central controller, wherein the central controller stores a content storage table, a pending interest table and a forwarding table information base, and is used for executing the following steps: according to the information in the interest packet, searching the corresponding data packet in the content storage table, if the corresponding data packet is found, the data packet forwards the data packet and the interest packet according to the request port corresponding to the interest packet recorded in the pending interest table, and storing the path information required for forwarding in the forwarding table information base,

the method comprises the steps that a double-layer NPN architecture is established, wherein the double-layer NPN architecture comprises a central network controller, a local network controller, a content router and an edge router, the central network controller is located at an upper network layer and used for storing information of all NDN network sub-domains in a lower network layer covered by the upper network layer and carrying out global forwarding, the local network controller is located at a lower network layer and used for managing the content router and the edge router in the NDN network sub-domains, the content router carries out content storage and communication interface management in the NDN network sub-domains, and the edge router is used for realizing interconnection between the NDN network sub-domains and other NDN network sub-domains;

based on the content storage table, the pending interest table and the forwarding table information base in the central network controller, the forwarding of the interest packet and the data packet is realized, and the method specifically comprises the following steps:

s22, checking the path information in the interest package,

the local controller stores a local content forwarding table LFT and a local neighbor table LNT, the path from the requesting node to the target node is found in the local network controller by looking up the local content forwarding table and the local neighbor table, the local content forwarding table records the name of the content stored in the sub-domain of the local NDN network and the domain number storing the content, and the local neighbor table stores the number of the boundary router of the adjacent domain;

7. The hybrid cloud communication system of claim 6, wherein the local controller further stores a local communication interface table LIT, and when a communication fault occurs during forwarding or backtracking, the local controller marks a communication fault path in the local communication interface table LIT, selects new path information from paths in the local communication interface table LIT in which the communication fault does not occur, stores the new path information in an to-be-determined interest table, and forwards a data packet according to the new path information.