A layered message forwarding method suitable for a non-full-link consortium chain network

Abstract

The application discloses a layered message forwarding method suitable for a non-full-link alliance chain network. First, after the alliance chain network is started, each node starts to pull static meta information of each node in the network. According to the pulled static meta information of the node, each node will cluster all nodes in the alliance chain network at least once locally to obtain a plurality of first-level clusters and second-level clusters. Each node starts cluster learning in the cluster where the node is located to obtain dynamic meta information. Meanwhile, each node will pull dynamic meta information of each node in the alliance chain network at a time, and after pulling enough, each node determines a representative node in other first-level clusters and establishes a link with each representative node, realizes efficient clustering of the node in the non-full-link alliance chain network, and thereafter, communication between each node and nodes of other clusters can be forwarded through the representative node, which helps to improve the broadcast performance of the non-full-link alliance chain network and reduce the burden on the network caused by redundant broadcast messages.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of peer-to-peer networks underlying blockchains, and particularly relates to a layered message forwarding method suitable for non-full-link alliance chains. BACKGROUND

[0002] The main application scenario of alliance chains is specific business cooperation between institutions, such as supply chain finance and electronic forensics. By using alliance chains, these institutions can jointly maintain a reliable database while ensuring data privacy and security. Due to the special application scenario of alliance chains, the consensus mechanism thereof is different from the proof of work (PoW) and other methods of public chains. Alliance chains usually use a Byzantine fault tolerance (BFT) consensus mechanism and rely on leader election, voting algorithms and other methods to complete consensus. The computing power overhead of BFT is much smaller than that of PoW, but the communication overhead of BFT is greatly affected by the size of the nodes and has poor scalability. Therefore, the consensus mechanism of alliance chains has lower requirements for computing resources than public chains, but the entire consensus process is highly dependent on the communication of the underlying distributed peer-to-peer network, and the performance of the underlying distributed peer-to-peer network also determines the upper limit of the performance of the entire consensus process.

[0003] Therefore, it is of great significance to study and optimize the underlying network model of alliance chains. At present, there are few studies on distributed peer-to-peer networks in large-scale alliance chain network scenarios, and most studies remain at the level of layered broadcasting of public chains. Such studies often use clustering algorithms such as K-means for clustering, obtain location coordinates through IP information, cluster nodes with similar physical locations, and select nodes with central locations in the cluster as representative nodes of the cluster to be responsible for intra-cluster broadcasting.

[0004] However, in the non-full-link scenario of alliance chains, such algorithms often perform poorly or even fail. The reasons are as follows:

[0005] 1. The special topology structure caused by the non-full-link network makes the node with a central physical location obtained through geographical location not necessarily the most suitable node for broadcasting in this topology structure.

[0006] 2. Enterprises are usually reluctant to expose the internal topology information of the enterprise and will use external agents to communicate with other enterprises. If other enterprises only use known agent IP information to calculate coordinate positions, the coordinate positions will be distorted, causing the traditional clustering algorithm to fail.

[0007] 3. The widely used K-means algorithm requires a preset K value, but a preset K value is difficult to estimate for an alliance chain cluster whose cluster size changes dynamically. SUMMARY

[0008] The present application aims to provide a layered message forwarding method suitable for a non-full-link alliance chain network, which can effectively solve the problems of special topological structure and position information distortion caused by information security due to non-full-link, thereby causing the failure of the clustering algorithm relied on by traditional layered broadcast. The method proposed by the present application significantly improves the broadcast efficiency of the non-full-link alliance chain network and reduces the network pressure caused by large-scale broadcast.

[0009] In order to achieve the above-mentioned purpose, the present application adopts the following technical solutions:

[0010] One kind is a layered message forwarding method suitable for a non-full-link alliance chain network

[0011] S1: After the alliance chain network is started, each node starts to pull the static meta information of each node of the alliance chain network, and the static meta information includes IP address and public key;

[0012] S2: According to the static meta information of the node pulled, each node will cluster all nodes in the alliance chain network at least once locally, obtain a plurality of first-level clusters and the second-level clusters corresponding to the first-level clusters;

[0013] S3: Each node learns in the cluster where it is located, specifically by establishing a link with other nodes in the cluster and conducting a delay test to obtain the dynamic meta information of the current node;

[0014] S4: Each node will pull the dynamic meta information of each node of the alliance chain network regularly; when each node pulls the dynamic meta information of a preset number of nodes, each node determines the representative node in the other first-level cluster and establishes a link with each representative node;

[0015] S5: Each node sends a broadcast message to the representative nodes of each first-level cluster, and each representative node forwards the broadcast message to other nodes in the cluster where the representative node is located.

[0016] In the S1, the pulling method of each node is that each node sends a PULL instruction to its adjacent node, and the adjacent node returns the latest static meta information it has to the current node as a response.

[0017] The S2 is specifically:

[0018] S2.1: When each node pulls the static meta information of a preset number of nodes, each node will use the coordinate information obtained from the IP address of the static meta information as an input parameter and perform the first round of clustering, and divide all nodes in the alliance chain network into a plurality of first-level clusters according to the distance of the coordinate information;

[0019] S2.2: If each first-level cluster meets the secondary clustering condition, the first-level cluster is subjected to secondary clustering to obtain a plurality of second-level clusters corresponding thereto; if not, S3 is executed.

[0020] The S2.1 is specifically:

[0021] Each node queries the corresponding coordinate information of each node by using the IP address in the pulled static meta-information of the node, and obtains a plurality of primary clusters after inputting the coordinate information of all nodes in the alliance chain network into ISODATA iterative self-organizing clustering.

[0022] In the S2.2, if each primary cluster satisfies the secondary clustering condition, the primary cluster is subjected to secondary clustering to obtain a plurality of secondary clusters corresponding thereto, including:

[0023] If different alliances are contained in each primary cluster, the current primary cluster is first segmented according to the types of alliances to obtain clusters corresponding to different alliances; if the cluster corresponding to the alliance to which the node belongs does not have an internal clustering method, the clusters corresponding to different alliances are recorded as different secondary clusters; if the cluster corresponding to the alliance to which the node belongs has an internal clustering method, the cluster corresponding to the alliance to which the node belongs is subjected to a second round of clustering by using the clustering method to obtain a plurality of secondary clusters.

[0024] The S3 is specifically:

[0025] Each node traverses other nodes in the cluster to which the node belongs, and for a host capable of establishing a direct link or completing communication through a forwarding path, each node obtains the public key of the node corresponding to the host and adds the public key to a Bloom filter to generate a breadth parameter in dynamic meta-information;

[0026] For a host capable of establishing a direct link or completing communication through a forwarding path, each node sends a heartbeat packet at a regular time and records a round-trip average delay to generate a performance parameter in dynamic meta-information, thereby obtaining dynamic meta-information of each node.

[0027] In the S4, the pulling manner of each node is that each node sends a PULL instruction to an adjacent node, and the adjacent node returns the latest dynamic meta-information possessed by the adjacent node to the current node as a response.

[0028] In the S4, each node traverses each primary cluster in the clustering result stored by the node, and selects a plurality of nodes in each primary cluster as representative nodes, so that the public keys of all nodes in the primary cluster can appear in the breadth parameter of at least one representative node.

[0029] In the S4, if there are multiple groups of candidate representative nodes in a primary cluster, the group of candidate representative nodes with the best average delay is selected as the representative node according to the performance parameter.

[0030] II. A computer device

[0031] The computer device comprises a memory and a processor, the memory stores a computer program, and the processor implements the steps of the hierarchical message forwarding method of the non-full-link alliance chain when executing the computer program.

[0032] Three, a computer readable storage medium

[0033] The computer readable storage medium stores a computer program, and the computer program implements the steps of the hierarchical message forwarding method of the non-full-link alliance chain when executed by the processor.

[0034] Four, a computer program product

[0035] The computer program product comprises a computer program / instruction, which implements the steps of the hierarchical message forwarding method of the non-full-link alliance chain when executed by the processor.

[0036] By adopting the technical scheme, the present application has the following beneficial effects:

[0037] 1) By two rounds of clustering, the problem of IP information and actual location distortion caused by the use of proxy communication by numerous enterprises in the alliance chain scenario due to information security reasons is solved. At the same time, the alliance internal self-defined clustering algorithm is supported, and the complex topological structure is adapted.

[0038] 2) By increasing dynamic meta-information, nodes can select representative nodes according to the dynamic meta-information of nodes in other clusters, solve the problem of representative node selection, and make nodes with strong communication ability successfully elected as representative nodes to be responsible for cluster broadcast.

[0039] 3) ISODATA iterative self-organizing clustering can automatically adjust the number of categories and clustering centers in the clustering process, so that the clustering result is closer to the objective reality. The influence of the initial clustering number selected in advance on the clustering result is avoided.

[0040] 4) Using the Bloom filter, a probabilistic data structure, can effectively compress the amount of dynamic meta-information, and significantly reduce the network burden on the premise that the upper consensus protocol has node fault tolerance capability. BRIEF DESCRIPTION OF DRAWINGS

[0041] Figure 1 The flowchart of the present application.

[0042] Figure 2 The network topology graph of the implementation case. DETAILED DESCRIPTION

[0043] The present application will be further described in detail below in combination with the drawings and specific embodiments. It should be understood that the specific embodiments described herein are only used to illustrate and explain the present application, and are not used to limit the present application.

[0044] As Figure 1 shown, the layered message forwarding method suitable for a non-full-link alliance chain network comprises the following steps:

[0045] S1: The topology diagram of the alliance chain network is as shown in Figure 2 After the alliance chain network is started, each node begins to pull the static meta information of each node in the alliance chain network, and the static meta information includes an IP address and a public key.

[0046] In S1, the pulling manner of each node is that each node sends a PULL instruction to its adjacent node, and the adjacent node returns the latest static meta information it has to the current node as a response.

[0047] S2: According to the pulled static meta information of the nodes, each node will locally cluster all the nodes in the alliance chain network at least once to obtain a plurality of first-level clusters and second-level clusters corresponding to the first-level clusters.

[0048] S2 specifically comprises:

[0049] S2.1: After each node pulls the static meta information of a preset number of nodes, each node uses the coordinate information obtained from the IP address of the static meta information as an input parameter and performs the first round of clustering to divide all the nodes in the alliance chain network into a plurality of first-level clusters according to the distance of the coordinate information. The specific value of the preset number is set by a configuration file. In actual application scenarios, when a user wants to quickly start, the proportion of the preset number is set a little smaller, and when a user wants to cluster more accurately, the proportion of the preset number is set a little larger.

[0050] S2.1 specifically comprises:

[0051] Each node queries the specific coordinate information corresponding to each node by using the IP address in the pulled static meta information of the nodes, and then obtains a plurality of first-level clusters after clustering all the nodes in the alliance chain network by using the coordinate information of all the nodes as an input and performing ISODATA iterative self-organizing clustering.

[0052] In this embodiment, since enterprise A adopts a proxy server and links with external enterprises, all the nodes in this enterprise are in cluster_0 cluster, rather than in Shenzhen and Beijing clusters, in the first round of clustering results. The first round of clustering results are shown in Table 1.

[0053] Table 1 is a first round of clustering results table

[0054] Node ID IP Address Public Key Cluster Result Real Location A0 ip_proxy:port_0 key_a_0 cluster_0 Beijing A1 ip_proxy:port_1 key_a_1 cluster_0 Beijing A2 ip_proxy:port_2 key_a_2 cluster_0 Beijing A3 ip_proxy:port_3 key_a_3 cluster_0 Shenzhen A4 ip_proxy:port_4 key_a_4 cluster_0 Shenzhen A5 ip_proxy:port_5 key_a_5 cluster_0 Shenzhen B0 ip_b_0:port key_b_0 cluster_1 Seattle B1 ip_b_1:port key_b_1 cluster_1 Seattle B2 ip_b_2:port key_b_2 cluster_1 Seattle C0 ip_c_0:port key_c_0 cluster_2 London C1 ip_c_1:port key_c_1 cluster_2 London C2 ip_c_2:port key_c_2 cluster_2 London

[0055] S2.2: If each primary cluster satisfies the secondary clustering condition, the primary cluster is subjected to secondary clustering to obtain a plurality of secondary clusters corresponding to the primary cluster; if not, S3 is executed.

[0056] If each primary cluster satisfies the secondary clustering condition, the primary cluster is subjected to secondary clustering to obtain a plurality of secondary clusters corresponding to the primary cluster, including:

[0057] If each primary cluster contains different alliances, the current primary cluster is first segmented according to the types of alliances to obtain clusters corresponding to different alliances; if the cluster corresponding to the alliance to which the node belongs does not have an internal clustering method, the clusters corresponding to different alliances are recorded as different secondary clusters; if the cluster corresponding to the alliance to which the node belongs has an internal clustering method, the second round of clustering is performed on the cluster corresponding to the alliance to which the node belongs using the clustering method, to obtain a plurality of secondary clusters, that is, the nodes in the current cluster are further subdivided into a plurality of secondary clusters.

[0058] In this embodiment, the nodes in enterprise A perform the second round of clustering, and use the self-defined clustering according to the machine room location information to complete the internal second round of clustering. In the clustering result, A0, A1, and A2 become a cluster, A3, A4, and A5 become a cluster, and the nodes of other companies will not perform the second round of clustering. Table 2 below is the second round of clustering result of the nodes in enterprise A running locally, and the nodes of other companies still maintain the first round of clustering result.

[0059] Table 2 is a second round of clustering result table

[0060]

[0061]

[0062] S3: Each node performs intra-cluster learning in the cluster to which the node belongs. If each node has only a primary cluster, intra-cluster learning is performed in the primary cluster. If there is a secondary cluster, intra-cluster learning is performed in the primary cluster. Specifically, the dynamic meta-information of the current node is obtained by establishing a link with other nodes in the cluster and performing a delay test, including performance parameters and breadth parameters.

[0063] S3 is specifically:

[0064] Each node traverses other nodes in the cluster to which the node belongs. For a host that can establish a direct link or complete communication through a forwarding path, each node obtains the public key of the node corresponding to the host and adds it to the Bloom filter to generate the breadth parameter in the dynamic meta-information.

[0065] For a host that can establish a direct link or complete communication through a forwarding path, each node sends a heartbeat packet at regular intervals and records the round-trip average delay to generate the performance parameter in the dynamic meta-information, thereby obtaining the dynamic meta-information of each node.

[0066] The meta-information table maintained by the nodes in enterprise A is shown in Table 3 below, and the meta-information table maintained by the nodes in enterprises B and C is shown in Table 4 below. The nodes in enterprises B and C do not know that the nodes in enterprise A have performed the second round of clustering, and still regard the nodes in enterprise A as belonging to the same cluster.

[0067] Table 3 is the information table stored locally by nodes A0-A5

[0068] Node ID IP Address Public Key Performance Parameter Breadth Parameter Cluster Result Real Location A0 ip_proxy:port_0 key_a_0 5ms bloom_a_0 cluster_00 Beijing A1 ip_proxy:port_1 key_a_1 10ms bloom_a_1 cluster_00 Beijing A2 ip_proxy:port_2 key_a_2 2ms bloom_a_2 cluster_00 Beijing A3 ip_proxy:port_3 key_a_3 20ms bloom_a_3 cluster_01 Shenzhen A4 ip_proxy:port_4 key_a_4 10ms bloom_a_4 cluster_01 Shenzhen A5 ip_proxy:port_5 key_a_5 20ms bloom_a_5 cluster_01 Shenzhen B0 ip_b_0:port key_b_0 5ms bloom_b_0 cluster_1 Seattle B1 ip_b_1:port key_b_1 10ms bloom_b_1 cluster_1 Seattle B2 ip_b_2:port key_b_2 10ms bloom_b_2 cluster_1 Seattle C0 ip_c_0:port key_c_0 5ms bloom_c_0 cluster_2 London C1 ip_c_1:port key_c_1 10ms bloom_c_1 cluster_2 London C2 ip_c_2:port key_c_2 10ms bloom_c_2 cluster_2 London

[0069] Table 4 is the information table stored locally by nodes B0-B2 and C0-C2

[0070] Node ID IP Address Public Key Performance Parameter Breadth Parameter Cluster Result Real Location A0 ip_proxy:port_0 key_a_0 5ms bloom_a_0 cluster_0 Beijing A1 ip_proxy:port_1 key_a_1 10ms bloom_a_1 cluster_0 Beijing A2 ip_proxy:port_2 key_a_2 2ms bloom_a_2 cluster_0 Beijing A3 ip_proxy:port_3 key_a_3 20ms bloom_a_3 cluster_0 Shenzhen A4 ip_proxy:port_4 key_a_4 10ms bloom_a_4 cluster_0 Shenzhen A5 ip_proxy:port_5 key_a_5 20ms bloom_a_5 cluster_0 Shenzhen B0 ip_b_0:port key_b_0 5ms bloom_b_0 cluster_1 Seattle B1 ip_b_1:port key_b_1 10ms bloom_b_1 cluster_1 Seattle B2 ip_b_2:port key_b_2 10ms bloom_b_2 cluster_1 Seattle C0 ip_c_0:port key_c_0 5ms bloom_c_0 cluster_2 London C1 ip_c_1:port key_c_1 10ms bloom_c_1 cluster_2 London C2 ip_c_2:port key_c_2 10ms bloom_c_2 cluster_2 London

[0071] S4: Each node will periodically pull the dynamic meta-information of each node in the alliance chain network, and the dynamic meta-information includes performance parameters and breadth parameters, etc., which are used as data indicators for selecting cluster representative nodes. After each node pulls the dynamic meta-information of several nodes, each node determines the representative nodes in other primary clusters and establishes a link with each representative node. Specifically, the nodes with higher performance parameters and breadth parameters in other primary clusters are regarded as the representative nodes of the cluster.

[0072] In S4, each node pulls by sending a PULL instruction to its adjacent nodes, and the adjacent nodes return the latest dynamic meta-information they have to the current node as a response.

[0073] In S4, each node traverses each primary cluster in the clustering result stored by the node, and selects a preset number of nodes in each primary cluster as representative nodes, so that the public keys of all nodes in the primary cluster can appear in the breadth parameter of at least one representative node. The specific value of the preset number is set through a configuration file. In actual application scenarios, when a user wants to quickly start, the proportion of the preset number is set a little smaller, and when a user wants more accurate clustering, the proportion of the preset number is set a little larger. If there are multiple groups of candidate representative nodes in a primary cluster, the group of candidate representative nodes with the best average latency is selected as the representative nodes according to the performance parameters.

[0074] In this embodiment, in the information table stored locally by the nodes in enterprises B and C, A0-A5 are still regarded as a cluster, but according to the dynamic meta-information, it is impossible to use one node to cover all public keys, and it is necessary to select one from A0-A2 and A3-A5 to complete the coverage, so enterprises B and C will select two representative nodes for enterprise A, and the selected nodes are A2 and A4 with high performance parameters. Similarly, nodes B0 and C0 are selected as representative nodes responsible for broadcasting in enterprises B and C.

[0075] S5: Each node sends a broadcast message to the representative node of each primary cluster, and each representative node forwards the broadcast message to other nodes in the cluster where the representative node is located.

[0076] The embodiment of the application provides a layered message forwarding device suitable for a non-full-link alliance chain network.

[0077] The embodiment of the application provides a layered message forwarding device suitable for a non-full-link alliance chain network.

[0078] The functions and roles of the units in the device are achieved in the implementation process of the corresponding steps in the above method, and will not be described here.

[0079] The technical features of the above-described embodiments can be combined in any way, and to make the description concise, all possible combinations of the technical features in the above-described embodiments are not described, but as long as the combinations of the technical features do not exist, they should be considered as the scope of the present application.

[0080] The embodiment of the application further provides a readable storage medium, which stores a program, and the program is executed by a processor to realize the layered message forwarding method suitable for the non-full-link alliance chain network.

[0081] The readable storage medium can be an internal storage unit of any data processing capable device of any of the preceding embodiments, such as a hard disk or a memory. The readable storage medium can also be an external storage device, such as a plug-in hard disk, a smart media card (SMC), an SD card, a flash card, etc. Further, the readable storage medium can include both an internal storage unit of any data processing capable device and an external storage device. The readable storage medium is used to store the computer program and other programs and data required by the any data processing capable device, and can also be used to temporarily store data that has been output or is to be output.

[0082] While the embodiments of the present application have been illustrated and described, it will be clear to those skilled in the art that various changes, modifications, substitutions and alterations can be made thereto without departing from the principles and spirit of the present application, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A hierarchical message forwarding method suitable for a non-full-link federation chain network, characterized in that, The method comprises the following steps: S1: After the consortium chain network is started, each node starts to pull static meta information of each node in the consortium chain network, and the static meta information comprises an IP address and a public key; S2: According to the pulled static meta information of the nodes, each node locally clusters all nodes in the consortium chain network at least once to obtain a plurality of first-level clusters and second-level clusters corresponding to the first-level clusters; S3: Each node performs intra-cluster learning in the cluster where the node is located, specifically by establishing a link with other nodes in the cluster and performing a delay test to obtain dynamic meta information of the current node; S4: Each node periodically pulls dynamic meta information of each node in the consortium chain network; after each node pulls dynamic meta information of a preset number of nodes, each node determines a representative node in each first-level cluster and establishes a link with each representative node; S5: Each node sends a broadcast message to the representative nodes of each first-level cluster, and each representative node forwards the broadcast message to other nodes in the cluster where the representative node is located. 2.The layered message forwarding method for a non-full-link federation chain network according to claim 1, wherein, In the S1, the pulling manner of each node is that each node sends a PULL instruction to an adjacent node, and the adjacent node returns the latest static meta information possessed by the adjacent node to the current node as a response. 3.The layered message forwarding method for a non-full-link federation chain network according to claim 1, wherein, The S2 specifically comprises: S2.1: After each node pulls static meta information of a preset number of nodes, each node locally uses coordinate information obtained by the IP address of the static meta information as an input parameter and performs first-level clustering to divide all nodes in the consortium chain network into a plurality of first-level clusters according to the distance of the coordinate information; S2.2: If each first-level cluster meets a second-level clustering condition, the first-level cluster is subjected to second-level clustering to obtain a plurality of second-level clusters corresponding to the first-level cluster; if not, the S3 is executed.

4. The layered message forwarding method for a non-full-link federation chain network according to claim 3, wherein, The S2.1 specifically comprises: Each node queries the coordinate information corresponding to each node by using the IP address in the pulled static meta information of the nodes, and then inputs the coordinate information of all nodes in the consortium chain network to perform ISODATA iterative self-organizing clustering to obtain a plurality of first-level clusters.

5. The layered message forwarding method for a non-full-link federation chain network according to claim 3, wherein, In the S2.2, if each first-level cluster meets a second-level clustering condition, the first-level cluster is subjected to second-level clustering to obtain a plurality of second-level clusters corresponding to the first-level cluster, which comprises: If the first-level cluster contains different alliances, the first-level cluster is first divided according to the types of the alliances to obtain clusters corresponding to different alliances; if the cluster corresponding to the alliance where the node is located does not have an internal clustering method, the clusters corresponding to different alliances are recorded as different second-level clusters; if the cluster corresponding to the alliance where the node is located has an internal clustering method, the second-level clustering is performed on the cluster corresponding to the alliance where the node is located by using the clustering method to obtain a plurality of second-level clusters.

6. The layered message forwarding method for a non-full-link federation chain network according to claim 1, wherein, The S3 specifically comprises: Each node traverses other nodes in the cluster where the node is located, and for a host that can establish a direct link or complete communication through a forwarding path, each node acquires a public key of the host corresponding node and adds the public key to a Bloom filter to generate a breadth parameter in the dynamic meta information; For the host that can establish a direct link or complete communication through a forwarding path, each node periodically sends a heartbeat packet and records a round-trip average delay to generate a performance parameter in the dynamic meta information, thereby obtaining dynamic meta information of each node.

7. The hierarchical message forwarding method for a non-full-link federation chain network according to claim 1, wherein, In the S4, each node pulls by sending a PULL instruction to its adjacent node, and the adjacent node returns the latest dynamic meta-information it has to the current node as a response. 8.The layered message forwarding method for a non-full-link federation chain network according to claim 1, wherein, In the S4, each node traverses each primary cluster in the clustering result stored by the node, and selects several nodes as representative nodes in each primary cluster, so that the public keys of all nodes in the primary cluster can appear in the breadth parameter of at least one representative node.

9. The layered message forwarding method for a non-full-link federation chain network according to claim 8, wherein, In the S4, if there are multiple groups of candidate representative nodes in a primary cluster, the group of candidate representative nodes with the best average time delay is selected as the representative nodes according to the performance parameter. 10.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-9. The processor, when executing the computer program, implements the steps of the method in any one of claims 1 to 9.

Images