阅读说明：本技术 一种基于坐标映射的多模标识网络寻址方法及系统 (Multi-mode identification network addressing method and system based on coordinate mapping ) 是由 李挥 胡嘉伟 邬江兴 伊鹏 朱伏生 李文军 安辉耀 李胜飞 陈世胜 唐宏 张云勇 于 2019-02-21 设计创作，主要内容包括：本发明适用于网络技术改进领域,提供了一种基于坐标映射的多模标识网络寻址方法,包括：S1、通过对网络中的每个节点赋予三维球坐标,将具有无标度性的多模标识网络映射到一个三维双曲空间中；S2、报文的发送者将将其目的地的节点坐标附于每段报文中；S3、路由节点在转发报文时,计算目的地与每个相邻节点之间的双曲距离,并选取最小者作为转发对象。该算法依赖的全局信息较少,且计算简单,易于在本地完成,由此提高了对大规模网络的适应性。(The invention is suitable for the field of network technology improvement, and provides a multi-mode identification network addressing method based on coordinate mapping, which comprises the following steps: s1, mapping the multi-mode identification network with non-scale property into a three-dimensional hyperbolic space by endowing each node in the network with a three-dimensional spherical coordinate; s2, the sender of the message attaches the node coordinates of the destination to each segment of the message; s3, when the routing node forwards the message, the hyperbolic distance between the destination and each adjacent node is calculated, and the minimum one is selected as a forwarding object. The algorithm has the advantages of less dependence on global information, simple calculation and easy local completion, thereby improving the adaptability to large-scale networks.)

1. A multi-mode identification network addressing method based on coordinate mapping is characterized by comprising the following steps:

s1, mapping a multi-mode identification network with non-standard property into a three-dimensional hyperbolic space;

s2, giving a three-dimensional spherical coordinate to each node in the multi-mode identification network in a three-dimensional hyperbolic space;

and S3, when any node in the three-dimensional hyperbolic space is addressed, the node coordinate of the destination of the node is attached to each segment of the message.

S4, selecting hyperbolic distance according to the calculated distance between the nodes

d₁₂＝cosh^-1(coshr₁coshr₂-sinhr₁sinhr₂cosΔθ₁₂)

The minimum adjacent node forwards the message;

where r, θ are derived from the coordinates of the nodes, Δ θ₁₂The central angle of a connecting line between two points and the origin:

2. the coordinate mapping-based multimode identification network addressing method according to claim 1, wherein the mapping in step S1 comprises an angular coordinate mapping and a radial coordinate mapping, the angular coordinate mapping comprising the steps of:

s111, taking longitude and latitude information of the geographical position of the central node angular coordinate with high connectivity;

and S112, judging whether the degree of the non-central node coordinate with low connectivity is larger than or equal to a set value, if so, measuring and calculating the average time delay from the non-central node to each central node based on an IP protocol, and selecting the set value central node with the minimum time delay to calculate the angular coordinate of the non-central node, otherwise, mostly only having one path to the central node, and directly copying the angular coordinate of the non-central node with the highest neighbor degree.

3. The coordinate mapping-based multimode identification network addressing method according to claim 2, characterized in that said radial coordinate mapping comprises the following steps:

and S121, dividing the global network into a plurality of subgraphs and carrying out independent path coordinate operation on each subgraph.

4. The method as claimed in claim 3, wherein the most central node in each sub-graph in step S121 has a radial coordinate r0, and only minor modifications are made to non-central nodes having smaller original radial coordinates.

5. The coordinate mapping based multimode identification network addressing method of claim 4, wherein the maximum likelihood estimation of the path coordinates comprises the following steps:

s1211, according to the statistical rule of the network, the degrees of the nodes are assumed to meet power distribution rho (kappa) -kappa^-γThe lowest degree is k₀Average degree ofThe relationship between the degree and the radial coordinate of the node is as follows:

wherein R is the radius of the sphere;

s1212, according to the statistical rule of the network, assuming that the connection probability of the two nodes is:

x is the hyperbolic distance between two points, where R can be given by the following integral:

wherein the parameter T is temperature, and the aggregation degree of the nodes is controlled; zeta is the curvature of the hyperbolic space, x' isHyperbolic distance between (r,0, 0);

s1213, taking m nodes i with highest central degree in the network₁,i₂...i_mThe other nodes measure and calculate the self and each i_*Time delay of if i_kIf the time delay is the minimum, the node belongs to the subgraph G_k. For degree of possession κ_iThe maximum likelihood estimate of the radial coordinates of the node of (2) is:

if node i ∈ G_j，Is i_jThe radial coordinate of the maximum likelihood estimation value of (1) is:

6. the multi-mode identification network addressing system based on coordinate mapping is characterized by comprising

The network mapping module is used for mapping a multi-mode identification network with non-standard property into a three-dimensional hyperbolic space and endowing each node in the network with a three-dimensional spherical coordinate;

a forwarding module for selecting hyperbolic distance according to the calculated distance between nodes

d₁₂＝cosh^-1(cosh r₁cosh r₂-sinh r₁sinh r₂cos Δθ₁₂)

The minimum adjacent node is used as a forwarding object;

where r, θ are derived from the coordinates of the nodes, Δ θ₁₂The central angle of a connecting line between two points and the origin:

7. the multi-mode identification network addressing system comprising the hyperbolic routing as claimed in claim 6, wherein multiple routing identifications exist in the network, such as content identification, identity identification, geospatial location identification, IP address identification, and the like, and real-time requirements of the network for multiple requirements are met through multiple identification dynamic adaptation and conversion technology, wherein hyperbolic space coordinates are used for addressing routing of multiple novel network identifications including content in a large-scale network.

8. The coordinate mapping-based multimodal identification network addressing system of claim 7 wherein the mapping in the network mapping module comprises an angular coordinate mapping and a radial coordinate mapping, the angular coordinate mapping comprising

The central node coordinate calculation unit is used for acquiring longitude and latitude information of the geographic position of the central node angular coordinate with high connectivity;

and the non-central node coordinate calculating unit is used for judging whether the degree of the non-central node coordinate with lower connectivity is greater than or equal to a set value or not, if so, based on an IP protocol, calculating the average time delay from each central node, and selecting the set value central nodes with the minimum time delay to calculate the angular coordinate of the non-central node coordinate calculating unit, if not, most of the non-central node coordinate calculating unit only has one path leading to the central node, and the angular coordinate of the non-central node coordinate calculating unit directly copies the highest degree in the neighborhood.

9. The coordinate mapping based multimode identification network addressing system of claim 8, wherein the radial coordinate mapping comprises

And the network dividing module is used for dividing the global network into a plurality of subgraphs and carrying out independent path coordinate operation on each subgraph.

10. The coordinate-mapping-based multimodal identification network addressing system of claim 9, wherein the centermost node in each subgraph in the network partitioning module has a radial coordinate r₀And only minor correction is carried out on the non-central node with the original smaller radial coordinate.

11. The coordinate mapping based multimode identification network addressing system of claim 10, wherein the maximum likelihood estimation of the path coordinates comprises:

a priori the hypothesis units. According to the statistical rule of the network, the degree of a node is assumed to satisfy power distribution rho (kappa) -kappa^-γThe lowest degree is k₀Average degree ofThe relationship between the degree and the radial coordinate of the node is as follows:wherein R is the radius of the sphere; according to the statistical rule of the network, the connection probability of two nodes is assumed as follows:

x is the hyperbolic distance between two points, where R can be given by the following integral:

wherein the parameter T is temperature, and the aggregation degree of the nodes is controlled; zeta is the curvature of the hyperbolic space, x' isAnd (r,0,0)Hyperbolic distance therebetween;

a subgraph division and a computation unit. Taking m nodes i with the highest central degree in the network₁,i₂…i_mThe other nodes measure and calculate the self and each i_*Time delay of if i_kIf the time delay is the minimum, the node belongs to the subgraph G_k. For degree of possession κ_iThe maximum likelihood estimate of the radial coordinates of the node of (2) is:

if node i ∈ G_j，r_j ^*Is i_jThe radial coordinate of the maximum likelihood estimation value of (1) is:

Technical Field

The invention belongs to the field of network technology improvement, and particularly relates to a multi-mode identification network addressing method and system based on coordinate mapping.

Background

The multimode identification network is a novel open network architecture provided aiming at the essential defects that the existing internet control capability is too concentrated and international multilateral co-management is lacked, and the like, and particularly refers to a network environment in which networks with different system structures jointly deploy multipath identification and cooperative routing addressing. For example, if a content network is deployed in a conventional network, a multi-mode network environment consisting of two network architectures and addressed by a content identifier and an address identifier is formed if data can be shuttled between the two networks. By utilizing the advantages of different networks to work cooperatively, the multimode identification network can improve the basic transmission capability of the current internet, enhance the utilization rate of network resources and enrich the functions of network layers. More importantly, the multi-mode identification network reduces the dependency and limitation of the existing internet system on address identification, and provides possibility for multilateral co-management of the internet.

In the design process of a network protocol, the adaptability of the routing overhead of the network protocol to a large-scale network must be considered, and the routing overhead can be measured by the scale of a forwarding table (FIB) and the number of control messages required when the network generates topology change. In next generation networks, Information Centric Networking (ICN) forwards directly content-oriented names, while internet of things (IoT) has an extremely large number of nodes, which results in an extremely large addressing space required for them and must be highly dynamic to cope with the actual demands of the network, which has to make people think again how to design routing mechanisms that match future networks.

Greedy Geometric Routing (GGR) maps a network into a metric space and assigns an address, also known as coordinates, to each node therein. Each segment of network message transmitted in the network attaches the coordinates of the destination to the network message, and when the router forwards the message, the router respectively calculates the geometric distance between each adjacent node and the destination, and selects the minimum distance as the next hop for forwarding. Since each node only needs to know coordinate information of its neighbors, the GGR can minimize the scale of FIB as much as possible, thereby providing a basis for designing routing protocols for large-scale networks.

The proposal of Hyperbolic Routing (HR) is based on the non-scalability of the network, i.e. the property that the degree of nodes in the network obeys a power distribution. By means of a mapping algorithm, the network is mapped into a space with negative curvature, i.e. with double curvature, in a two-dimensional case, each node is mapped into a circular disk with a radius R, and is given a polar coordinate (R, θ), wherein an angular coordinate θ represents the relative position of the node in the network, and a radial coordinate R represents the degree of the center of the node, and the more the node is centered, the smaller the radial coordinate is, i.e. the closer the center of the circular disk is. When the angular coordinates of two nodes are constant, the hyperbolic distance between the two nodes is reduced along with the reduction of the radial coordinates of the two nodes, so that the greedy routing based on the hyperbolic distance tends to select the more centralized node as a forwarding object.

Because most networks in reality, such as the internet, have non-standard performance, on the basis of adopting a proper mapping algorithm, a simple greedy strategy based on the hyperbolic distance has a high probability of forwarding a message to a destination node, and for a few situations in which forwarding fails, auxiliary intelligent forwarding strategies can be adopted, so that the success rate of the hyperbolic route reaches 100%.

However, hyperbolic routing also has its drawbacks: compared with the traditional routing protocol based on the shortest path algorithm, the forwarding path selected by the hyperbolic routing greedy algorithm has larger transmission delay, which is not only due to the inherent defect of the greedy strategy, but also due to the fact that most of the existing hyperbolic mapping algorithms do not consider network delay. Since reducing latency is also one of the core goals of routing protocol design, we must address this issue.

Kleinberg, et al, "Geographic routing using hyperbaric space" in 2007, proposes the earliest mapping algorithm for hyperbolic routing, where any network can be mapped into a hyperbolic space by constructing a minimum spanning tree for the network, and greedy routing based on this mapping has 100% success rate.

The algorithm must know the global topology information, and meanwhile, when any change occurs to the minimum spanning tree of the network, the coordinates of all nodes must be recalculated, so that the adaptability to the dynamic network is poor.

Bogun, Marin, Fragkiskos Papadopoulos, Ditri Krioukov in the "stabilizing the internet with hyper-carboxylic mapping" article in 2010, and Papadopoulos, Fragkiskos equals the "Green forwarding in dynamic scalefree network embedded in hyper-carboxylic metallic spaces" article in 2010, going back to the second, no longer striving to embed any network into the hyperbolic space, but only to those whose mapping has an inordinate degree. The method maps the network into a hyperbolic space by a maximum likelihood estimation method based on prior assumptions (such as degree distribution, connection probability and the like) about the network and an actual connection state. When a dynamic network is faced, the coordinates obtained by the algorithm have better stability in a longer time (from month to year), and a node newly entering the network only needs to know the periphery of the node, but not the global topological information, so that the coordinate calculation can be completed.

The algorithm only considers the adjacency state of the network and all connections are treated as equivalent regardless of their delays, resulting in routes based on the algorithm often choosing less optimal paths with larger delays.

L ehman is equal to the word "An experimental interpretation of hyperbolically routing with a smart forwarding plane in NDN" of 2015, trying to utilize a forwarding plane with NDN having adaptivity and intelligence to perform optimization on delay of a suboptimal path selected from a hyperbolically routing.

Although this approach is significant for delay reduction, the number of delay samples probed is not sufficient to optimize the forwarding path for short-term transmission scenarios. At the same time, this approach fails to optimize the worst case scenario (i.e., those paths with delays much greater than the theoretical optimum).

Disclosure of Invention

The invention aims to provide a multi-mode identification network addressing method based on coordinate mapping, and aims to solve the technical problem.

The invention is realized in such a way that a multimode identification network addressing method based on coordinate mapping comprises the following steps:

s1, mapping a multi-mode identification network with non-standard property into a three-dimensional hyperbolic space;

s2, giving a three-dimensional spherical coordinate to each node in the multi-mode identification network in a three-dimensional hyperbolic space;

and S3, when any node in the three-dimensional hyperbolic space is addressed, the node coordinate of the destination of the node is attached to each segment of the message.

S4, selecting hyperbolic distance according to the calculated distance between the nodes

d₁₂＝cosh^-1(cosh r₁cosh r₂-sinh r₁sinh r₂cosΔθ₁₂)

The minimum adjacent node forwards the message;

where r, θ are derived from the coordinates of the nodes, Δ θ₁₂The central angle of a connecting line between two points and the origin:

the further technical scheme of the invention is as follows: various routing identifiers exist in the network, such as content identifiers, identity identifiers, geospatial location identifiers, IP address identifiers, and the like, and the real-time requirements of the network for various requirements are met through a multi-identifier dynamic adaptation and conversion technology, wherein the step S3 is used for addressing routing of various novel network identifiers including content in a large-scale network.

The further technical scheme of the invention is as follows: the mapping in step S1 includes an angular coordinate mapping and a radial coordinate mapping, and the angular coordinate mapping includes the steps of:

s111, taking longitude and latitude information of the geographical position of the central node angular coordinate with high connectivity;

and S112, judging whether the degree of the non-central node coordinate with low connectivity is larger than or equal to a set value, if so, measuring and calculating the average time delay from the non-central node to each central node based on an IP protocol, and selecting the set value central node with the minimum time delay to calculate the angular coordinate of the non-central node, otherwise, mostly only having one path to the central node, and directly copying the angular coordinate of the non-central node with the highest neighbor degree.

The further technical scheme of the invention is as follows: the radial coordinate mapping comprises the following steps:

and S121, dividing the global network into a plurality of subgraphs and carrying out independent polar coordinate operation on each subgraph.

The further technical scheme of the invention is as follows: the most central node in each sub-graph in the step S121 has a radial coordinate r₀And only minor correction is carried out on the non-central node with the original smaller radial coordinate.

The further technical scheme of the invention is as follows: the maximum likelihood estimation of the radial coordinates comprises the following steps:

s1211, according to the statistical rule of the network, the degrees of the nodes are assumed to meet power distribution rho (kappa) -kappa^-γThe lowest degree is k₀Average degree ofThe relationship between the degree and the radial coordinate of the node is as follows:

wherein R is the radius of the sphere;

s1212, according to the statistical rule of the network, assuming that the connection probability of the two nodes is:

x is the hyperbolic distance between two points, where R can be given by the following integral:

wherein the parameter T is temperatureDegree, the degree of aggregation of the control nodes; zeta is the curvature of the hyperbolic space, x' isHyperbolic distance between (r,0, 0);

s1213, taking m nodes i with highest central degree in the network₁,i₂...i_mThe other nodes measure and calculate the self and each i_*Time delay of if i_kIf the time delay is the minimum, the node belongs to the subgraph G_k. For degree of possession κ_iThe maximum likelihood estimate of the radial coordinates of the node of (2) is:

if node i ∈ G_j，Is i_jThe radial coordinate of the maximum likelihood estimation value of (1) is:

it is another object of the present invention to provide a coordinate mapping based multimode identification network addressing system, which comprises

The network mapping module is used for mapping a multi-mode identification network with non-standard property into a three-dimensional hyperbolic space and endowing each node in the network with a three-dimensional spherical coordinate;

a forwarding module for selecting hyperbolic distance according to the calculated distance between nodes

d₁₂＝cosh^-1(cosh r₁cosh r₂-sinh r₁sinh r₂cosΔθ₁₂)

The minimum adjacent node is used as a forwarding object;

where r, θ are derived from the coordinates of the nodes, Δ θ₁₂Is two points ofCenter angle of origin line:

the further technical scheme of the invention is as follows: various routing identifications exist in the network, such as content identifications, identity identifications, geographic spatial position identifications, IP address identifications and the like, the real-time requirements of the network on various requirements are met through multi-identification dynamic adaptation and conversion technology, and double-area coordinates are used for addressing routing of various novel network identifications including contents in a large-scale network.

The further technical scheme of the invention is as follows: the mapping in the network mapping module comprises an angular coordinate mapping and a radial coordinate mapping, the angular coordinate mapping comprising

The central node coordinate acquisition unit is used for acquiring longitude and latitude information of the geographic position of the central node angular coordinate with high connectivity;

and the non-central node coordinate acquisition unit is used for judging whether the degree of the non-central node coordinate with lower connectivity is greater than or equal to a set value or not, if so, based on an IP protocol, calculating the average time delay from each central node, and selecting the set value central node with the minimum time delay to calculate the own angular coordinate, if not, most of the non-central node coordinates only have one path to the central node, and the angular coordinate of the non-central node coordinates directly copies the highest value in the neighborhood.

The further technical scheme of the invention is as follows: the radial coordinate mapping includes

And the network dividing module is used for dividing the global network into a plurality of subgraphs and carrying out independent polar coordinate operation on each subgraph.

The further technical scheme of the invention is as follows: the most central node in each subgraph in the network partitioning module has a radial coordinate r₀And only minor correction is carried out on the non-central node with the original smaller radial coordinate.

The further technical scheme of the invention is as follows: the maximum likelihood estimation of the radial coordinate comprises

The maximum likelihood estimation of the radial coordinate comprises:

a priori the hypothesis units. According to the statistical rule of the network, the degree of a node is assumed to satisfy power distribution rho (kappa) -kappa^-γThe lowest degree is k₀Average degree ofThe relationship between the degree and the radial coordinate of the node is as follows:wherein R is the radius of the sphere; according to the statistical rule of the network, the connection probability of two nodes is assumed as follows:

x is the hyperbolic distance between two points, where R can be given by the following integral:

wherein the parameter T is temperature, and the aggregation degree of the nodes is controlled; zeta is the curvature of the hyperbolic space, x' isHyperbolic distance between (r,0, 0);

a subgraph division and a computation unit. Taking m nodes i with the highest central degree in the network₁,i₂...i_mThe other nodes measure and calculate the self and each i_*Time delay of if i_kIf the time delay is the minimum, the node belongs to the subgraph G_k. For degree of possession κ_iThe maximum likelihood estimate of the radial coordinates of the node of (2) is:

if node i ∈ G_j，Is i_jThe radial coordinate of the maximum likelihood estimation value of (1) is:

the invention has the beneficial effects that: for the traditional hyperbolic routing algorithm, the algorithm can reduce the transmission delay by about 30% at most, and meanwhile, the inherent high forwarding success rate is kept. The algorithm has the advantages of less network information, simple calculation and easy local completion, thereby improving the expansibility of the scheme.

Drawings

Fig. 1 is a schematic diagram of a 3-node to center point according to an embodiment of the present invention.

Fig. 2 is a schematic diagram of a mapping algorithm provided in the embodiment of the present invention.

Fig. 3 is a schematic diagram of a forwarding process provided in an embodiment of the present invention.

Fig. 4 is a schematic view of a hyperbolic disk provided by an embodiment of the present invention.

Fig. 5 is a schematic diagram illustrating a DS comparison between a recalculated angular coordinate and an original angular coordinate according to an embodiment of the present invention.

Fig. 6 is a schematic diagram illustrating a DS comparison between the diameter coordinates and original diameter coordinates for re-calculation according to an embodiment of the present invention.

Detailed Description

Referring to fig. 1-4, the multi-mode identification network addressing method based on coordinate mapping provided by the present invention is described in detail as follows

The core design idea of the scheme is as follows:

1. the actual delay of the network is taken into account in the coordinate mapping algorithm to reduce the average delay of the routes.

Specifically, the time delay participates in the mapping process in two ways:

1) in the angular coordinate mapping algorithm, the non-central node measures the average time delay between itself and the peripheral central node, and completes the calculation of the angular coordinate of itself based on the angular coordinate of the central node by taking the time delay as the spherical distance.

2) In the path coordinate mapping algorithm, the global network is divided into a plurality of subgraphs based on time delay from each central node, and each subgraph can independently perform path coordinate operation, so that the locality of a route is improved, and the delay is reduced.

2. Different mapping mechanisms are adopted for the central node and the non-central node of the network so as to adapt to the non-scale property of the network.

In a scaleless network, the degree distribution of nodes obeys the power law, in other words, the nodes in the network can be roughly divided into two categories: most of the common user nodes with a low number of connections, and very few of the central nodes with a very large number of connections. For the two types of nodes, different coordinate mapping mechanisms are respectively adopted.

Since the number of central nodes occupies a crucial position in the routing process, the central nodes are used as anchor points of coordinate mapping operation, the angular coordinates of the central nodes are from the geographic position, and the polar coordinates are from the central degree of the central nodes in the peripheral area. And the non-central node measures and calculates the coordinate of the non-central node by taking the central node as a reference.

The technical problems to be solved by the invention are as follows:

1) the average transmission delay of the hyperbolic route is reduced, and the original high success rate and high expandability of the hyperbolic route are not lost.

2) Reducing the global information required for the mapping algorithm: in order to improve the expansibility and the adaptability, the coordinate mapping calculation only needs local network knowledge as far as possible and does not depend on any complex global scheduling process.

The algorithm maps a network with non-standard property to a three-dimensional hyperbolic spaceIn which each node in the network is given three-dimensional spherical coordinatesTwo pointsAndthe hyperbolic distance of (a) is derived from the cosine theorem:

d₁₂＝cosh^-1(cosh r₁cosh r₂-sinh r₁sinh r₂cosΔθ₁₂) (1)

where Δ θ₁₂The central angle of a connecting line between two points and the origin:

when the hyperbolic routing is adopted, the address of the destination of each segment of message is attached to the hyperbolic routing, each routing node knows the coordinates of the neighbor and greedily selects the neighbor with the minimum hyperbolic distance from the destination to forward.

The algorithm is divided into an angular coordinate mapping part and a radial coordinate mapping part, and the specific flow is as follows:

1. angular coordinate mapping

Each node is first given an angular coordinateThereby being mapped to a spherical surfaceThe above. Spherical surfaceCan be viewed as a simulation of the earth's surface, and the angular position of a node reflects its actual position in the network.

For those nodes with high connectivity, their angular coordinates come directly from their geographical location, i.e. latitude and longitude information, for the following reasons: 1) according to past experience, the transmission delay between two nodes is proportional to the geographic distance between the two nodes, and mapping based on the geographic position has a good optimization effect on the delay. 2) The mapping method is very direct and convenient for calculation. 3) The mapping does not depend on the topological information of the network, so that the mapping has better stability in a dynamic network environment.

For non-central nodes with a low number of connections, we do not apply the same approach to them, since their location in the network depends more on local topology information than on geographical information. For the node i with the degree of more than or equal to 3, based on the IP protocol (the route of the node is usually based on the shortest path algorithm), the node I measures and calculates the average time delay from the node I to each central node, and selects three central nodes j with the minimum time delay₁,j₂,j₃For calculating the own angular coordinates:

if i is from j_kHas a time delay of t_k，j_kWith angular coordinatesi angular coordinateThe following optimization problem results:

s.t.λΔθ_ik＝t_k-ξ(k＝1,2,3) (4)

Δθ_ikis composed ofAndequation constraint (4) represents a direct proportional relationship between network delay and spherical distance, and a slack variable ξ is used to ensure that a feasible solution can be found, as shown in fig. 1.

The antecedent term of the objective function ensures that the value of the relaxation variable is as small as possible, the consequent (Δ θ)_i1+Δθ_i2+Δθ_i3) Then the method is used to select the one with the smallest sum of spherical distances when there are multiple feasible solutions.

For the non-central nodes with the degree less than or equal to 2, because most of the nodes have only one path to the central node, the angular coordinate of the node directly copies the node with the highest degree in the neighbor.

2. Radial coordinate mapping

The radial coordinate r reflects the central degree of the node, and in a scale-free network, r should satisfy exponential distribution.

For example, shanghai city has an extremely large number of internet users, so there are several super nodes with high centrality, and for a message sent from north to south of korea, because the population of korea is small compared to shanghai, the hyperbolic route forwarding path of the message may be attracted by the high centrality of shanghai, that is: north korea-shanghai-south korea, thereby causing additional delay.

We adopt the way of dividing the global network into subgraphs to deal with this problem, and the most central node in each subgraph has similar path coordinates, so the forwarding process will tend to select the central node in this subgraph more, thereby improving the locality of the route and reducing the transmission delay. Taking m nodes i with the highest central degree in the network₁,i₂...i_mThe other nodes measure and calculate the self and each i_*Time delay of if i_kIf the time delay is the minimum, the node belongs to the subgraph G_k。

The path coordinates are first derived from the maximum likelihood estimation, we have the following prior conditions:

1) the degrees k of the nodes satisfy power distributions ρ (κ) to κ^-γThe lowest degree is k₀Average degree ofThe degree and the radial coordinate satisfy the following relation:

wherein R is the radius of the sphere.

2) The hyperbolic distance is two nodes of x, and the connection probability is as follows:

t is temperature, and the aggregation degree of the nodes is controlled; ζ is the curvature of the hyperbolic space; in this case, R can be obtained by the following integral equation:

wherein x isAnd (r,0, 0).

Given the above prior conditions, the degree k is possessed_iThe maximum likelihood estimate of the path coordinates of node i of (2) is:

if node i ∈ G_jThen, the radial coordinate is:

β are used to adjust the relative weights of the path and angular coordinates during routing, by the above formula, the centermost node in each sub-graph has a path coordinate r₀And only minor correction is carried out on the non-central node with the original smaller radial coordinate.

To evaluate the relative advantages of the above algorithm, we generated a series of topological graphs satisfying the non-scale with a random network generator for simulation testing of the algorithm. Wherein the degrees of the nodes obey power distribution, and the longitudes and latitudes obey uniform distribution on the spherical surface. The algorithm parameters of the generator are derived from a fit to the real internet, and the propagation delay of two nodes in the graph is proportional to the spherical distance between them.

Two nodes in a network are randomly selected as a starting point and a stopping point of a route, and a hyperbolic route is tried to be used for forwarding a message from the starting point to the stopping point. We use Delay spread (DS, Delay Stretch) as an indicator to measure the effect of the algorithm: the DS of a route test is (actual transmission delay/theoretical minimum transmission delay). We generated 50 random graphs and for each set of experimental parameters, one hundred thousand random routing tests were performed in each graph.

In the experiment, the routing performance based on the original coordinates (namely, the coordinates based on the degree and longitude and latitude information completely) and the coordinates recalculated by the algorithm is compared. Since the mapping algorithms for the angular and radial coordinates are relatively independent, we will evaluate them separately.

As shown in fig. 5, the recalculated angular coordinate vs is compared with the DS of the original angular coordinate, and the 75 th percentile (representing a worse case) and the 95 th percentile (representing an almost worst case) are listed. It can be seen that the recalculation of coordinates can better reduce the worst-case delay when selecting the appropriate center node ratio. While the 75 percentile is already close to 1 by itself, the degree of optimization is relatively small.

As shown in fig. 6, for comparing the recalculated radius coordinate vs with the DS of the original radius coordinate, where the abscissa is the number of divided subgraphs, dividing the graph and recalculating the radius coordinate can reduce the latency of the network better than the angular coordinate, and as shown in fig. 6, the algorithm can reduce the latency by up to 30% for the worse case under proper parameter selection.

The hyperbolic mapping algorithm maps the network into a hyperbolic space. For visual intuition, the network is mapped into a 2-dimensional disk, and it can be seen that the greater the degree of a node (which can be used to represent the popularity of the node), the closer the node is to the center of the disk, i.e., the smaller the radial coordinate; and the angular position of a node represents its relative position in the network. As shown in fig. 2.

And (4) forwarding process of the hyperbolic route. The coordinates of the destination are attached to the message, the routing node calculates the hyperbolic distance between each next hop and the destination, and selects the minimum distance to forward. As shown in fig. 3.

A triangular tessellation of hyperbolic discs (poincare discs), where each triangle is of the same size. For two points that are visually the same distance, the closer to the edge of the disk, the greater the actual distance between them. Three-dimensional hyperbolic spheres have similar properties. As shown in fig. 4.

It is another object of the present invention to provide a coordinate mapping based multimode identification network addressing system, which comprises

The network mapping module is used for mapping a multi-mode identification network with non-standard property into a three-dimensional hyperbolic space and endowing each node in the network with a three-dimensional spherical coordinate;

a forwarding module for selecting hyperbolic distance according to the calculated distance between nodes

d₁₂＝cosh^-1(cosh r₁cosh r₂-sinh r₁sinh r₂cosΔθ₁₂)

The minimum adjacent node is used as a forwarding object;

where r, θ are derived from the coordinates of the nodes, Δ θ₁₂The central angle of a connecting line between two points and the origin:

various routing identifications exist in the network, such as content identifications, identity identifications, geographic spatial position identifications, IP address identifications and the like, the real-time requirements of the network on various requirements are met through multi-identification dynamic adaptation and conversion technology, and the hyperbolic coordinate is used for addressing routing of various novel network identifications including the content in a large-scale network.

The mapping in the network mapping module comprises an angular coordinate mapping and a radial coordinate mapping, the angular coordinate mapping comprising

The central node coordinate acquisition unit is used for acquiring longitude and latitude information of the geographic position of the central node angular coordinate with high connectivity;

and the non-central node coordinate acquisition unit is used for judging whether the degree of the non-central node coordinate with lower connectivity is greater than or equal to a set value or not, if so, based on an IP protocol, calculating the average time delay from each central node, and selecting the set value central node with the minimum time delay to calculate the own angular coordinate, if not, most of the non-central node coordinates only have one path to the central node, and the angular coordinate of the non-central node coordinates directly copies the highest value in the neighborhood.

The polar mapping includes

And the network dividing module is used for dividing the global network into a plurality of subgraphs and carrying out independent polar coordinate operation on each subgraph.

The most central node in each subgraph in the network partitioning module has a radial coordinate r₀And only minor correction is carried out on the non-central node with the original smaller radial coordinate.

The maximum likelihood estimation of the radial coordinate comprises

A priori the hypothesis units. According to the statistical rule of the network, the degree of a node is assumed to satisfy power distribution rho (kappa) -kappa^-γThe lowest degree is k₀Average degree ofThe relationship between the degree and the radial coordinate of the node is as follows:wherein R is the radius of the sphere; according to the statistical rule of the network, the connection probability of two nodes is assumed as follows:

x is the hyperbolic distance between two points, where R can be given by the following integral:

wherein the parameter T is temperature, and the aggregation degree of the nodes is controlled; zeta is the curvature of the hyperbolic space, x' isHyperbolic distance between (r,0, 0);

a subgraph division and a computation unit. Taking m nodes i with the highest central degree in the network₁,i₂...i_mThe other nodes measure and calculate the self and each i_*Time delay of if i_kIf the time delay is the minimum, the node belongs to the subgraph G_k. For degree of possession κ_iThe maximum likelihood estimate of the radial coordinates of the node of (2) is:

if node i ∈ G_j，Is i_jThe radial coordinate of the maximum likelihood estimation value of (1) is:

based on simulation experiments, when proper parameters are adopted, for the traditional hyperbolic routing algorithm, the algorithm can reduce transmission delay by about 30% at most, and meanwhile, the inherent high forwarding success rate is kept.

Meanwhile, the algorithm depends on less network information, is simple in calculation and is easy to complete locally, and therefore the expansibility of the scheme is improved.

The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents and improvements made within the spirit and principle of the present invention are intended to be included within the scope of the present invention.