CN103475719A - Content distribution method for minimizing cross-domain flows in CDN-P2P fusion network - Google Patents

Abstract

The present invention proposes a content distribution method for minimizing cross-domain traffic in a CDN-P2P fusion network. The network scenarios involved in the method include a CDN replica server, a CDN content cache server, and a local tracker server set up by an ISP, namely, Loc- tracker server, and user nodes. This method is based on the topology information collected by the Loc-tracker server, such as link delay, link cost, number of cross-operators, etc., calculates and selects the path and node with the smallest transmission cost, and controls the total cost of content distribution, thereby The server is urged to select nodes in the same ISP domain with a lower transmission cost, so as to achieve the purpose of optimizing node selection. The Loc-tracker server is used to manage nodes and CDN content cache servers in the ISP domain. Loc-tracker servers exchange information in a P2P manner to form an overlay network. The invention can effectively reduce the generation of cross-domain traffic, reduce the operating cost of the network, relieve the waste of backbone network bandwidth resources, and improve the efficiency of content distribution.

Description

A content distribution method for minimizing cross-domain traffic in a CDN-P2P fusion network

technical field

The invention relates to the technical field of communication, and in particular to a method for optimizing and controlling cross-domain traffic in a CDN-P2P integrated content distribution network.

Background technique

CDN (Content Distribution Network) network and P2P (peer-to-peer) network are currently the most popular content distribution technologies, but both have many defects. CDN has high reliability, but poor scalability, making it difficult to fundamentally improve distribution Efficiency, while P2P has strong expansion ability, but poor stability, which is easy to cause backbone network congestion. Therefore, the integration of CDN and P2P technology can complement its own defects on the one hand and transform the traditional structure of CDN or P2P. On the other hand, it can also perfectly combine the scalability of P2P with the reliability and manageability of CDN. Therefore, the CDN-P2P integrated content distribution network is a content distribution platform with manageability, high scalability and high reliability. It not only combines the advantages of both CDN network and P2P network, but also overcomes the shortcomings of both. . Therefore, it has become a more feasible content distribution technology at present, and it is a hot spot in the field of content distribution technology research in recent years. The application of this technology can also make the distribution efficiency of the content distribution network to a higher level.

FIG. 1 is a schematic diagram of an existing CDN-P2P network architecture. The network architecture is mainly composed of three parts, including: a central server 1 , several edge servers 2 , and several P2P user domains 3 . Among them, the central server 1 and several edge servers 2 can directly exchange information, and each edge server 2 can directly exchange information with the peer nodes in the adjacent P2P user domain 3, and the peer nodes in the same user P2P domain 3 Nodes can exchange information directly. The central server 1 stores all the content and is responsible for the content distribution and management of the entire network. The edge server 2 is distributed on the edge of the network and caches some content. It is mainly responsible for local load balancing and responding to user requests in the adjacent user P2P domain.

In the current CDN-P2P fusion network, the selection of the optimal node is based on a random algorithm. The range of node selection is very wide without any regionalization, and the peer selection of P2P applications is based on the layered network, which cannot consider the connection in the physical network. Utilization of resources such as relationships and bandwidth. Therefore, unreasonable cross-domain traffic floods the entire network, consumes huge network bandwidth, overwhelms the network infrastructure, and continuously increases ISP cross-domain costs. Aiming at the consumption of ISP network infrastructure by a large amount of ISP cross-domain traffic, operators have adopted various blocking measures such as blocking ports, traffic shaping, and bandwidth control, which caused problems such as excessive download time and a significant decline in user satisfaction. CDN-P2P Converged network users have also taken corresponding measures to carry out anti-blocking, presenting a state of confrontation between users and ISPs, seriously affecting the development of the Internet.

In order to better improve the distribution efficiency of the converged content distribution network, the present invention proposes a converged network content distribution method based on a Loc-tracker server. The main idea is to reduce the generation of cross-domain traffic by controlling transmission costs.

Contents of the invention

Aiming at the above-mentioned technical defects, the present invention provides an ISP-friendly CDN-P2P fusion network content distribution method, which is used to reduce the generation of cross-domain traffic between ISPs, alleviate the waste of bandwidth resources in the backbone network, improve the efficiency of content distribution, reduce Access delays reduce operating costs.

In order to achieve the above object, the present invention provides a content distribution method for minimizing cross-domain traffic in a CDN-P2P fusion network, the method comprising the following steps:

(1) Set up a network scenario to realize the ISP-friendly CDN-P2P content distribution method;

(2) Set the list structure and functions of the Loc-tracker server;

(3) Set the optimization model and selection method for candidate node selection;

(4) Set the cooperation mode between the CDN content cache server and the Loc-tracker server; among them,

In step 1, the network scene includes a CDN copy server, a CDN content cache server, a local tracker server set up by an ISP, that is, a Loc-tracker server, and a user node;

In step 2, the Loc-tracker server operates as a Tracker agent, builds and maintains a candidate list. Each Loc-tracker server maintains two lists, which are used to maintain nodes in the local ISP and other external nodes. , the Loc-tracker server can perceive the change of the network topology by obtaining the topology information in the border route, the Loc-tracker server will actively collect the local node information of the same ISP, and the information of other external nodes is provided by other Loc-tracker servers;

In step 3, the optimized model needs to meet the following conditions:

a. The number of cross-ISPs should be less than or equal to the maximum limit in the network;

b. The link delay should be less than or equal to the maximum waiting time of users in the network;

c. The number of hops must be less than or equal to the maximum value specified by the network protocol;

The selection method is as follows: when a user in a certain ISP domain has a request for content data, the user will send an inquiry request to the Loc-tracker server in the ISP domain to ask which user nodes can provide the requested content data. Finally, the Loc-tracker server will select nodes that satisfy the optimization model from the two lists according to the request content to form a candidate list; when these returned nodes cannot meet the user's content request, the user will initiate the first The second request, after the Loc-tracker server receives the second request, the server will ask other Loc-tracker servers to find nodes in other ISP areas; The Loc-tracker server in this ISP domain sends inquiries and requests data to Loc-tracker servers in other ISP domains;

In step 4, the CDN content cache server located in each ISP domain will be used as a special user node to store copy files and be available for other users to download. Like ordinary user nodes, the CDN content cache server will also store the information contained in the content Send to the same ISP or the Loc-tracker server in the nearest ISP. When the user request arrives, the Loc-tracker server will guide the user to the CDN content cache server for content request according to the current network status. When requesting the path of the user to the CDN content cache server When the state is congested, the Loc-tracker server will not guide the user to the content cache server for content request, and will return to the user some nodes with shorter delay and lower total transmission cost. If the request user goes to the CDN content cache server If the path is not congested, and the total transmission cost and delay satisfy the optimization model, the Loc-tracker server will guide the user to the CDN content cache server for data requests.

Further, the CDN copy server is connected to the CDN content cache server in a tree form, the Loc-tracker server is set by the ISP of the area to which it belongs, and the Loc-tracker server manages users and CDN content cache server, Loc-tracker servers in different ISPs exchange data in P2P mode, forming a P2P overlay network composed of Loc-tracker servers.

Description of drawings

Figure 1 Schematic diagram of CDN-P2P network architecture;

Figure 2 is a network scene diagram of an ISP-friendly CDN-P2P converged network content distribution method;

Figure 3 is a schematic diagram of the composition of the Loc-tracker server list;

Detailed ways

In order to better illustrate the working process of the present invention, its specific implementation process will be described in detail below in conjunction with the accompanying drawings.

(1) Design an ISP-friendly CDN-P2P content distribution method scenario diagram: as shown in Figure 1, the network scenario mainly includes a CDN replica server, a CDN content cache server, and a local tracker server (referred to as the Loc-tracker server) set up by the ISP. ), and user nodes. The CDN replica server is connected to the CDN content cache server in a tree form. The Loc-tracker server is set by the ISP in the area to which it belongs. The Loc-tracker server manages the users and the CDN content cache server in the area to which it belongs. Loc-tracker servers in different ISPs exchange data in P2P mode, forming a P2P overlay network composed of Loc-tracker servers.

(2) Set the structure and function of the Loc-tracker server list: the Loc-tracker server operates as a Tracker agent, builds and maintains the candidate list. Two lists are maintained in each Loc-tracker server. As shown in Figure 2, the first list is used to maintain nodes in the local ISP, and the second list is used to maintain other external nodes. The Loc-tracker server can sense changes in network topology by obtaining topology information in border routes. The Loc-tracker server will take the initiative to collect the local node information of the same ISP, record the number of cross-ISPs between nodes, link cost, link delay and hop count, etc. The information of other external nodes is provided by other Loc-tracker servers . Different Loc-tracker servers form an overlay network, and these Loc-tracker servers will regularly exchange data in a P2P manner in the overlay network. When a user initiates a request for content data, the Loc-tracker server will filter out a group of candidate nodes and feed back to the user, and these candidate nodes must satisfy the optimization model.

(3) Design the optimization model and selection method for optimal node selection:

The optimization model is as follows:

Assume that the cost incurred by a unit of traffic across an ISP in the network is A. The shortest path experienced when exchanging data from node v ₀ to node v _h is p=<v ₀ ,v ₁ ,…,,v _h >, the number of cross-ISPs in path p is x _p , and the traffic on path p is D _p , the number of hops is h _p , and the minimum bandwidth on path p is b(p).

The total cost generated on the path p from node v ₀ to node v _h is:

The path delay required when transmitting a data block of size _Dp is:

Payload Size/Path Minimum Bandwidth =

When the Loc-tracker server performs peer node screening, three conditions are required. The selected nodes must meet the three conditions, as follows:

The number of cross-ISPs should be less than or equal to the maximum limit Mn in the network (this value can be set according to the network structure, for example, the number of cross-ISPs between nodes can be limited to be less than or equal to 2);

The link delay should be less than or equal to the maximum value Mt of the user's waiting time in the network (for example, the limitable link delay must be less than or equal to 6 seconds);

The number of hops must be less than or equal to the maximum value Mh specified by the network protocol (for example, the maximum number of hops that can be limited is less than or equal to 15 hops).

Assuming that a shared file K is divided into N data blocks, the user needs to send content requests to at most N nodes in order to obtain all the blocks. When the above three conditions are combined and N links are taken into consideration, the minimum cost function of the entire content distribution process can be transformed into a solution to an optimization problem. The optimization problem (that is, the optimization model) is as follows:

$\left\{\begin{array}{c}\mathrm{<span\; class="notranslate">\; min</span>}<span\; class="notranslate">\; (</span>\underset{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; N</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}<span\; class="notranslate">\; (</span>\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\mathrm{<span\; class="notranslate">\; c</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; v</span>}}_{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; v</span>}}_{\mathrm{<span\; class="notranslate">\; ij</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; pj</span>}}{\mathrm{<span\; class="notranslate">\; D.</span>}}_{\mathrm{<span\; class="notranslate">\; pj</span>}}\mathrm{<span\; class="notranslate">\; A</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span>\\ \mathrm{<span\; class="notranslate">\; the\; s</span>}<span\; class="notranslate">\; .</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; .</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; pj</span>}}<span\; class="notranslate">\; \&le;</span>\mathrm{<span\; class="notranslate">\; mn</span>}\\ \frac{{\mathrm{<span\; class="notranslate">\; D.</span>}}_{\mathrm{<span\; class="notranslate">\; pj</span>}}}{\mathrm{<span\; class="notranslate">\; b</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; pj</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; \&le;</span>\mathrm{<span\; class="notranslate">\; Mt.</span>}\\ {\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; pj</span>}}<span\; class="notranslate">\; \&le;</span>\mathrm{<span\; class="notranslate">\; mh</span>}\end{array}\right.<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1,2,3</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; )</span>$

The node v _ij that satisfies the optimization model among all the nodes containing the file K blocks is the candidate node selected by the Loc-tracker server. Since most of the fees that the operator needs to pay during the transmission process are generated across ISP domains, the method of limiting the total fee can improve the local node hit rate within the same ISP, thereby reducing the generation of cross-ISP traffic.

The optimal node selection method is as follows:

As shown in Figure 3, when a user in an ISP domain has a content data request, the user will first send an inquiry request to the Loc-tracker server in the ISP domain to inquire which peer nodes can provide the requested content data. After receiving the request, the Loc-tracker server will select the node v _ij that satisfies the optimization model from list 1 and list 2 according to the content of the request to form a candidate list.

When these returned nodes cannot satisfy the user's content request (such as incomplete download content, etc.), the user will initiate a second request to the Loc-tracker server. After the Loc-tracker server receives the second request, the server These three conditions will be relaxed in order to search. For example, expand the number of cross-ISPs from less than or equal to 2 to less than or equal to 3, and so on.

When a data request needs to be made across ISPs, for example, when a node in ISP1 needs to download content blocks in ISP2, the selection process of candidate nodes is as follows: first, the requesting node needs to send a request to the Loc-tracker server in the domain (denoted as L1 ) to send a request, and then L1 sends an inquiry request to the Loc-tracker server of ISP2 (denoted as L2). According to the node information in the list, L2 will filter out the corresponding nodes to form a candidate list and return it to L1 through the calculation of the optimization model. Finally, the candidate list will be sent to the requesting user via L1.

The cooperation method between the CDN content cache server and the Loc-tracker server set in (4) specifically includes the following content: the CDN content cache server located in each ISP domain will be regarded as a special peer, and this special peer is used to store Copy files and make them available for download by other users. Like ordinary peers, the CDN content cache server will also send the content information to the Loc-tracker server in the same ISP or the nearest ISP, and the information is recorded in List 1. When a user request arrives, the Loc-tracker server can guide the user to the CDN content cache server for content request according to the current network status. When the path status of the requesting user to the CDN is congested, the Loc-tracker server will not guide the user to the content cache server for content request, and will return to the user some nodes with shorter delay and lower total transmission cost. If there is no congestion in the path from the user to the CDN content cache server, and if the total transmission cost and delay satisfy the optimization model, the Loc-tracker server will guide the user to the CDN content cache server for data requests.

The present invention uses the designed integrated content distribution method based on the Loc-tracker server to control the distribution cost to increase the hit rate of nodes in the local ISP, so as to reduce the generation of cross-domain traffic and improve the distribution efficiency. In the present invention, the distribution cost is realized through the optimization model (the selected optimal node needs to meet the minimum path cost to the requesting node, less than or equal to Mn cross-ISPs, less than or equal to Mt seconds of path delay, and less than or equal to Mh hops). In order to achieve the purpose of improving the hit rate of nodes within the ISP while ensuring that user needs are met, thereby reducing the generation of cross-ISP traffic, reducing operating costs and waste of backbone network bandwidth resources, and alleviating the impact of infrastructure The pressure improves the distribution efficiency of CDN-P2P integrated content distribution network.

The above disclosures are only preferred embodiments of the present invention, and should not limit the scope of rights of the present invention. The equivalent changes made according to the patent scope of the present invention still fall within the scope of the present invention.

Claims (2)

1. A content distribution method for minimizing cross-domain traffic in a CDN-P2P fusion network, characterized in that the method comprises the following steps: (1) Set up a network scenario to realize the ISP-friendly CDN-P2P content distribution method; (2) Set the list structure and functions of the Loc-tracker server; (3) Set the optimization model and selection method for candidate node selection; (4) Set the cooperation mode between the CDN content cache server and the Loc-tracker server; among them, In step 1, the network scene includes a CDN copy server, a CDN content cache server, a local tracker server set up by an ISP, that is, a Loc-tracker server, and a user node; In step 2, the Loc-tracker server operates as a Tracker agent, builds and maintains a candidate list. Each Loc-tracker server maintains two lists, which are used to maintain nodes in the local ISP and other external nodes. , the Loc-tracker server can perceive the change of the network topology by obtaining the topology information in the border route, the Loc-tracker server will actively collect the local node information of the same ISP, and the information of other external nodes is provided by other Loc-tracker servers; In step 3, the optimized model needs to meet the following conditions: a. The number of cross-ISPs should be less than or equal to the maximum limit in the network; b. The link delay should be less than or equal to the maximum waiting time of users in the network; c. The number of hops must be less than or equal to the maximum value specified by the network protocol; The selection method is as follows: when a user in a certain ISP domain has a request for content data, the user will send an inquiry request to the Loc-tracker server in the ISP domain to ask which user nodes can provide the requested content data. Finally, the Loc-tracker server will select nodes that satisfy the optimization model from the two lists according to the request content to form a candidate list; when these returned nodes cannot meet the user's content request, the user will initiate the first The second request, after the Loc-tracker server receives the second request, the server will ask other Loc-tracker servers to find nodes in other ISP areas; The Loc-tracker server in this ISP domain sends inquiries and requests data to Loc-tracker servers in other ISP domains; In step 4, the CDN content cache server located in each ISP domain will be used as a special user node to store copy files and be available for other users to download. Like ordinary user nodes, the CDN content cache server will also store the information contained in the content Send to the same ISP or the Loc-tracker server in the nearest ISP. When the user request arrives, the Loc-tracker server will guide the user to the CDN content cache server for content request according to the current network status. When requesting the path of the user to the CDN content cache server When the state is congested, the Loc-tracker server will not guide the user to the content cache server for content request, and will return to the user some nodes with shorter delay and lower total transmission cost; if the request user goes to the CDN content cache server If the path is not congested, and the total transmission cost and delay satisfy the optimization model, the Loc-tracker server will guide the user to the CDN content cache server for data requests. 2. The method according to claim 1, wherein the CDN replica server is connected in a tree form with the CDN content cache server, the Loc-tracker server is set by the ISP of the area to which it belongs, and the The Loc-tracker server manages the users and CDN content cache servers in the area to which it belongs. The Loc-tracker servers in different ISPs exchange data in P2P mode, forming a P2P overlay network composed of Loc-tracker servers.

Images