RU2321956C2 - Multi-layered content delivery network and method for multi-layered content delivery - Google Patents

Abstract

FIELD: multi-layered content delivery.

SUBSTANCE: in accordance to the invention the network contains a content release layer, content delivery layer with at least one level of delivery sub-layer and a layer of boundary servicing, where the service release layer contains at least one node of Internet service provider (ICP), and is connected to first level delivery sub-layer in the content delivery layer, each delivery sub-layer contains at least one delivery unit, and the delivery sub-layer of the lowest level is connected to boundary service layer, the boundary service layer contains at least one boundary service area, and content, which is subject to release in a content release layer, is distributed through delivery node of first level delivery sub-layer, which is connected to content release layer, and further downwards level after level until the boundary service layer is reached.

EFFECT: distributed provision of services, optimized dispatching mode and local balancing of servicing load.

2 cl, 6 dwg

Description

FIELD OF THE INVENTION

The present invention relates to a technology for a content delivery network (informationally meaningful content), and more particularly, it relates to a multi-layer content delivery network and a multi-layer content delivery method.

State of the art

As the number of Internet users grows rapidly and the demand for improved quality of Internet services grows, the traditional mode of direct provision of information services by the Internet content provider (ICP) is facing serious problems. Typically, an ICP application site contains physical sites organized in one or more locations and provides an information service for Internet users nationally or even globally. This Internet service delivery mode makes accessing global Web sites slow and inefficient due to the structural characteristics of the Internet and problems such as bandwidth bottlenecks.

Information distribution is an effective way to solve these problems, i.e. the creation of a completely new content delivery network (CDN) based on the existing Internet network, which is specially used for the efficient transfer of diverse multimedia content over the Internet. Using CDNs for Internet users, it is possible to enjoy Internet services by accessing only nearby network sites, and intelligent distribution of network data streams can be achieved. For CDN, the network response speed has been significantly increased and a well-developed technical solution has been provided for such problems as a narrow network bandwidth, high user access intensity and an unbalanced distribution of network sites, which leads to a negative impact on user access. The principle of functioning of the CDN is to transfer the content of Web sites from the network backbone to the borders of the network closer to users in order to reduce the influence of factors such as transmission delays and network instability, which affects access efficiency and, as a result, provides high-quality services to users. As a result, it becomes possible for users to receive the requested information from nearby locations at the highest speed. Since this technology can reduce the load on the Internet, Web sites can provide a greater number of content-related services with large data streams, such as video programs, music on demand, etc. At the same time, the reliability of a variety of services, such as online transactions and network-based banking, is guaranteed in an efficient manner. As a summary, a CDN can pre-release content information from ICP Web sites to network edge devices so that users can receive quick responses from nearby edge devices when accessing Web sites, thereby improving the quality of service to users while reducing the load on central nodes.

The CDN network structures of the prior art are all relatively simple, and a single-level structure is adopted regardless of whether the CDN corresponds to an enterprise scale, a city scale, or a country scale, as shown in FIG. Key components of a CDN include devices such as a global load balancer, central server (CS), and edge server (ES). The CDN network architecture is mainly composed of a central part and a boundary part. The central portion typically comprises devices similar to the global load balancer and CS, while the boundary portion typically comprises a plurality of ES servers that act as boundary devices to provide access information to users. Information from ICP sites is pre-issued to each ES through CS, and each user request for access to the ICP site is dispatched by the global load balancer. As can be seen from the CDN network structure shown in FIG. 1, although ES CDN servers are distributed across the network boundary, global load balancing devices that act as dispatchers are located in the center of the network.

A typical user access scheduling procedure in the above CDN structure shown in FIG. 2 comprises the following steps:

Step 201: the user accesses the World Wide Web site through a browser and activates a universal index of the information resource (URL) of the content of this Web site.

Step 202: the user terminal browser requests a domain name server (DNS) to perform a domain name analysis of the Web site.

Step 203: The local DNS requests the authorized DNS of the Web site to anonymize the domain name of the Web site.

Step 204: the authorized DNS of the Web site notifies the local DNS that the global load balancer acts as an authorized DNS for the subdomain name of the Web site’s domain name and returns the Internet Protocol address (IP address) of the global load balancer.

Step 205: the local DNS sends an analysis request to the global load balancer.

Step 206: the global load balancer selects the optimal ES based on a predetermined predetermined strategy, typically through serial polling, and returns the IP address of the selected ES to the local DNS.

Step 207: The local DNS returns the IP address of the user terminal selected by the ES browser.

Step 208: the user accesses the selected ES in accordance with the returned IP address and sends a content request to the selected ES.

Step 209: if there is no content requested by the user on the selected ES, the selected ES will receive the CS content required by the user, after which step 210 will be performed, otherwise step 210 will be executed directly.

Step 210: the selected ES provides the required content according to the request of said user.

Although the quality of access by the user is improved to some extent by adopting the above scheme, there are still some problems due to which the quality of service and efficient use of the CDN network are far from satisfactory. For example, firstly, working on a large scale CDN is difficult; secondly, since a unified and centralized dispatch mode is adopted, global load balancing devices acting as dispatchers are centralized together, however their service area can be quite wide, which leads to a large load on dispatchers and slow response from them and requires dispatchers high parallelism when processing load; thirdly, it is impossible to detect the proximity of the user according to the IP address of the user and it is impossible to accurately dispatch to nearby servers; fourthly, in terms of releasing content and managing it, only the PULL (push) function is supported, while the PUSH (push) function is not supported, as a result of which the active push ability is weak; fifthly, when ES is selected, content is not detected, which leads to a low percentage of successful access attempts, moreover, since it is not possible to detect how busy the network is, it is impossible to select a network with a low data flow rate to provide the service.

Currently, there is another circuit in which the Layer 4 (L4) switch is adopted, wherein the L4 switch has the functions of a global load balancing device according to the above scheme. The typical user access scheduling procedure shown in FIG. 3 comprises the following steps:

Step 301: The user accesses the World Wide Web site and activates the content URL of this Web site.

Step 302: the user terminal browser queries the local DNS to parse the domain name of the Web site.

Step 303: The local DNS forwards the analysis request to the L4 switch.

Step 304: The L4 switch requests an authorized DNS Web site for domain name analysis.

Step 305: The authorized DNS of the Web site, through analysis, obtains the IP addresses of multiple ESs as requested, and then returns these IP addresses to the L4 switch.

Step 306: The L4 switch selects the optimal ES based on a predetermined predetermined strategy, typically through sequential polling, and then returns the IP address of the selected ES to the local DNS.

Step 307: The local DNS returns the IP address of the user terminal selected by the ES browser.

Step 308: the user accesses the selected ES according to the IP address returned by the local DNS, and sends a content request to the selected ES.

Step 309: if there is no content requested by the user on the selected ES, the selected ES will receive the content required by the user from the CS, and then step 310 will be performed, otherwise step 310 will be executed directly.

Step 310: the selected ES provides the required content as requested by said user.

The CDN, which adopted the L4 switch scheme, still has some problems, for example, the load on the L4 switch is very high, since it processes almost all the accesses, and the L4 switch can only process the layer 4 network protocol information, while being unable to process the information of the network protocol of layer 7, i.e. unable to detect content.

As can be seen from the above presentation of these two schemes, with an increase in the number of users, the current widespread single-level CDN becomes unsuitable for environments with large-scale and high-density applications and is unable to implement distributed services and local load balancing during maintenance.

SUMMARY OF THE INVENTION

The present invention provides a multi-layer content delivery network as well as a multi-layer content delivery method such that the CDN network structure can be more flexible and the traditional CDN scheduling mode can be optimized, thereby improving CDN QoS, realizing distributed services as well as local load balancing of service and increasing the share of successful access attempts.

The technical scheme of the present invention is implemented as follows.

The multilayer content delivery network includes a content release layer including at least one Internet Content Provider (ICP) node, a content delivery layer including at least one delivery sublayer level, and an edge service layer, the layer the content release layer is connected to the first layer delivery layer of the content delivery layer, each level of the delivery layer of the content delivery layer contains at least one delivery node, and the delivery sublayer of the lowest level is connected to the boundary layer service, the boundary service layer includes at least one boundary service area, while the content release layer releases the content, and the content is distributed through the delivery node of the first level delivery sublayer, decreasing level by level to the boundary service layer.

Each of these ICP nodes includes at least one working ICP terminal and at least one ICP.

Each of these delivery nodes includes at least one media management tool (MM) and at least one CS.

Each of these border service areas is deployed in the local center of each border service area in a distributed manner.

The ICP node on the content release layer is the upstream node for the delivery sublayer delivery node that is directly connected to the content release layer, and the boundary service area on the boundary service layer is the downstream node for the lowest level delivery sublayer delivery node that is directly connected with a boundary service layer.

A multi-layer CDN delivery method comprises the following steps:

a) the ICP node distributes the content to be released to at least one content delivery layer delivery node that is directly connected to the ICP node;

b) the delivery node receives a content delivery command to determine whether it is the delivery node of the lowest level delivery sublayer, and if so, performs step c), otherwise it reads the content to be released from the upstream node that is directly connected to the node the delivery, having accepted the content delivery command, stores the content in the CS of the delivery node, distributes the content to one or more downstream delivery nodes that are directly connected to the delivery node, which received a content delivery manda, and repeats the step of determining whether it is a delivery sublayer delivery node of the lowest level;

c) in accordance with the accepted content delivery team, the delivery node receives content to be released from the upstream delivery node that is directly connected to said delivery node, stores the content in the CS of the delivery node, distributes the content to one or more border service areas that are directly connected to mentioned delivery node.

Step a) further comprises the steps of scheduling the user's access to the ICP via the MRB in the border service area where the user is located, and providing the desired service to the user via CS or ES in the border service area where the user is located.

The step of scheduling the user's access to the ICP includes at least the steps of analyzing the ICP domain name the user is accessing, discovering the ES servers or selecting the optimal ES.

The ES server discovery step comprises at least a step of detecting how close the ES servers are to the user, whether they have the content that the user needs or whether the network is busy or inactive.

The step of selecting the optimal ES comprises the step of determining the optimal ES according to how close the ES servers are to the user, whether the ES content required by the user on these servers, or whether the network is busy or inactive.

Compared with prior art techniques, the present invention provides the ability to build networks of various sizes, including networks of the scale of a city, region or country. At the same time, the phased expansion of the network is also supported by the present invention, for example, a city-wide network can be expanded to a regional-scale network and even to a country-wide network with a more flexible CDN network structure. According to the present invention, the conventional scheduling mode in the CDN is optimized so that the distributed scheduling is adapted so that the MRB service areas become smaller, the load is reduced and the response time is reduced, thereby improving the quality of service of the CDN. Since the present invention employs multi-layer and multi-level content delivery management, it is guaranteed that ICP can implement network-wide release and service for single-point access so that distributed service delivery and local balancing of service load can be achieved. At the same time, according to the present invention, user accesses are dispensed based on how busy the ES servers are and how close the ES servers are to the user, so that ES servers that are relatively inactive can be selected to provide services in order to improve the quality of service. Moreover, since user accesses can be scheduled based on content, the success rate of access attempts is increasing.

List of drawings

1 is a schematic diagram illustrating a network structure according to prior art schemes.

2 is a flowchart illustrating a user access scheduling procedure according to scheme 1 according to the prior art.

FIG. 3 is a flowchart illustrating a user access scheduling procedure according to scheme 2 according to the prior art.

4 is a schematic diagram illustrating a network structure according to an embodiment of the present invention.

5 is a flowchart illustrating a content release procedure according to an embodiment of the present invention.

6 is a flowchart illustrating a user access scheduling procedure according to an embodiment of the present invention.

DETAILED DESCRIPTION OF THE INVENTION

In order to make the objective, technical scheme and advantages of the present invention more clear, the present invention will be described in further detail below with reference to the accompanying drawings and specific embodiments.

A key objective of an embodiment of the present invention is to build a network configured for distributed scheduling and multi-layer delivery; while dispatch devices are not located in the center of the network, but are distributed in the local center of each service area responsible for access scheduling in the local area; content delivery management is performed in multi-layer and multi-level mode, thereby ensuring that ICP can implement release and service throughout the network for single-point access.

A multilayer CDN according to an embodiment of the present invention comprises a content release layer, a content delivery layer with at least one delivery sublayer level, and an edge service layer. Assume that the content delivery layer contains two levels of the delivery sublayer, as shown in FIG. 4, the CDN contains such basic devices as ICP, ICP work terminal, CS, media management tools (MM), media request brokers (MRBs) and ES. MM is responsible for signaling processing, and the main function of the MM is to receive release commands related to the release of content sent by the ICP working terminal, send content delivery commands to the ES or MM of the downstream level or CS of each level in accordance with the release commands and receive feedback information from ES or MM downstream level or CS of each level. Each level CS is responsible for the storage and distribution of the center’s content, as well as the provision of services. The MRB in each boundary service area of the boundary service layer is distributed in the local center of the boundary service area and is mainly responsible for scheduling accesses within the network of the local area and implementing local load balancing. The solid line in FIG. 4 indicates content release and delivery procedure, while the dashed line indicates content release commands.

The above-described CDN components can be arranged in ICP nodes, delivery nodes, and border service areas. Each ICP node contains at least one working ICP terminal and one or more ICPs. Each delivery node contains at least one MM and one CS. Each boundary service area contains at least one CS, one MM, one MRB, and one or more ES.

The content release layer contains one or more ICP nodes. The content delivery layer contains one or more delivery sublayers, and each delivery sublayer contains at least one delivery node. The boundary service layer contains one or more boundary service areas. The content release layer is directly connected to the content delivery layer, and the content delivery layer directly to the boundary service layer, and the content to be released by the content release layer extends to the boundary service layer through the content delivery layer.

The delivery sublayer and the delivery sublayer delivery nodes on the aforementioned content delivery layer can be divided into different levels according to the connection path: the delivery nodes directly connected to the content release layer act as first level delivery nodes; the set of all first-level delivery nodes is called the first-level delivery sublayer; ICP nodes in the content release layer are upstream nodes for first level delivery nodes directly connected to ICP nodes; delivery nodes directly connected to first level delivery nodes and having the same role as first level delivery nodes are called second level delivery nodes; the set of all second-level delivery nodes is called a second-level delivery sublayer; first-level delivery nodes are upstream nodes for second-level nodes directly connected to first-level delivery nodes, and the rest can be inferred by analogy. The boundary service areas of the boundary service layer are directly connected to the lowest level delivery nodes; the set of all delivery nodes of the lowest level is called the sublayer of delivery of the lowest level; the lowest level delivery nodes are upstream nodes for frontier service areas directly connected to the lowest level delivery nodes. There is at least one level of delivery nodes in the content delivery layer.

The content to be released by the content release layer first spreads to one or more first level delivery nodes directly connected to the first level release layer, and said first level delivery nodes distribute the content to be released to one or more second level delivery nodes directly connected to the nodes first-level delivery, after which said second-level delivery nodes distribute content to be released to one or more third-level delivery nodes, directly connected to second level delivery nodes; the rest can be inferred by analogy, until the content is distributed level by level down to the boundary service layer, thereby realizing multilayer content delivery.

Based on the CDN network structure shown in FIG. 4, FIG. 5 shows a typical content release procedure comprising the following steps:

Step 501: the ICP work terminal of the specific ICP node on the content release layer sends a content release command to the MM of one or more first level delivery nodes directly connected to the content delivery layer.

Step 502: in accordance with the received content release command, the MM of said first level delivery node sends a content release command to the CS of the delivery node where the MM is located.

Step 503: in accordance with the received content release command, the CS of said first level delivery node reads the content to be released by ICP from the content storage equipment of the ICP node and stores the content in itself.

Step 504: The CS of the first level delivery node notifies the MM of the delivery node where the CS is located that the content to be released by the ICP is already stored in the CS of the first level delivery node.

Step 505: The first level delivery node MM sends a content delivery command to the second level delivery node MM directly connected to the first level delivery node.

Step 506: in accordance with the received content delivery command, the second-level delivery node MM sends a content delivery command to the CS of the delivery node where the MM is located.

Step 507: in accordance with the received CS content delivery command, the second level delivery node reads the content to be released from the upstream CS, namely the CS of the first level delivery node, and stores the content in itself.

Step 508: The CS of the second level delivery node notifies the MM of the delivery node where the CS is located that the content to be released is already stored in the CS of the second level delivery node.

Step 509: The second level delivery node MM sends a content release command to the MM of the border service area directly connected to the delivery node.

Step 510: in accordance with the received content release command MM of said boundary service area sends a content release command to the CS boundary service area.

Step 511: in accordance with the received CS content release command of said border service area, reads the content to be released from the upstream CS, namely the CS of the second level delivery node, and stores the content in the local CS, namely the CS of the border service area.

Step 512: The CS of the border service area notifies the MM of the node where the CS is located that the content to be released is already stored in the CS of the border service area.

Step 513: MM of said border service area sends a content delivery command to an ES of said border service area.

Step 514: in accordance with the received ES content delivery command of said border service area, reads the content to be released from the CS border service area where the ES is located and stores the content in itself.

Following the above content release procedure, the content to be released by the ICP is pushed to the network boundary in advance through the aforementioned multi-layer CDN, namely to the ES at the network boundary. The above content release procedure is applicable not only to the regional level CDN, but also the city-wide CDN and the country-wide CDN.

Based on the CDN network structure according to an embodiment of the present invention, the typical user access scheduling procedure shown in FIG. 6 comprises the following steps:

Step 601: the user accesses the World Wide Web site and activates the content URL of this site.

Step 602: the user terminal browser requests the local DNS to parse the domain name of the Web site.

Step 603: The local DNS queries the authorized DNS of the Web site to parse the domain name of the Web site.

Step 604: the authorized DNS of the Web site notifies the local DNS that the authorized DNS of the corresponding content of the Web site is a locally deployed MRB, and returns the IP address of the MRB.

Step 605: The local DNS returns the IP address of the MRB to the browser.

Step 606: the user terminal browser accesses the MRB.

Step 607: the MRB detects all ES servers in the local group area, selects the optimal ES according to how close the ES is or whether the content is in the ES, and returns the IP address of the selected ES to the browser.

Step 608: the user accesses the selected ES and sends a content request to the selected ES.

Step 609: if the selected ES does not have the content required by the user, the selected ES will receive content from the local CS before step 610, otherwise step 610 is performed directly.

Step 610: the selected ES provides the user-desired content according to the user's request.

As can be seen from the above procedure, after the analysis performed by the DNS, the access of local users is controlled by the local MRB and the services are provided by the local ES, which embodies the idea of implementing services in a distributed mode.

Since distributed scheduling and multi-layer delivery are adapted in the embodiments of the present invention to build a network, compared to a conventional CDN, an embodiment of the present invention has the advantage of supporting a multi-layer and multi-level network design, whereby small and large-scale CDNs can be constructed. Since content delivery management according to an embodiment of the present invention is also performed in a multi-layer and multi-layer mode, ICP can implement the release and provision of network-wide services for single-point access. Since the service areas are divided according to an embodiment of the present invention and distributed scheduling is implemented locally in each boundary service area, the MRB load is effectively reduced and the network status is detected in an efficient manner.

Although the above embodiment is illustrated by an example of a content delivery layer containing two levels of a delivery sublayer, the embodiment is also applicable to a network structure with a content delivery layer containing one or more levels of a delivery sublayer.

The above description is only a preferred embodiment of the present invention, which should not be interpreted as limiting the scope of the present invention.

Claims (12)

1. A multilayer content delivery network (CDN), wherein the nodes of this network are organized in a plurality of layers, wherein the plurality contains a content release layer including at least one Internet content provider (ICP) node, a content delivery layer including at least one level of a delivery sublayer, and a boundary service layer for interacting with a user, which includes at least one boundary service region, wherein each level of the delivery sublayer includes at least m Here is one delivery node, each of the ICP nodes in the content release layer is connected to at least one delivery node in the first level delivery sublayer, and the delivery node in the lowest level delivery sublayer is connected to nodes in the boundary service layer, with ICP nodes in the release layer content releases content, and the content is distributed through the delivery node of the first level delivery sublayer, decreasing level by level to the nodes in the boundary service layer, while the delivery node reads the content from the delivery node of the upper level I, with whom it is directly connected, stores the content and distributes it to one or more downstream delivery nodes with which it is directly connected, or distributes this content to one or more border service areas with which it is directly connected. 2. The content delivery network of claim 1, wherein each of said ICP nodes includes at least one ICP operational terminal and at least one ICP. 3. The content delivery network of claim 1, wherein each of said delivery nodes includes at least one multimedia control (MM) and at least one central server (CS). 4. The content delivery network according to claim 1, in which each of the said border service areas includes at least one CS, at least one MM, at least one multimedia request broker (MRB), and at least one border server (ES). 5. The content delivery network of claim 1, wherein the MRBs in each border service area are deployed in a local center in a distributed manner in each border service area. 6. The content delivery network of claim 1, wherein the ICP node on the content release layer is an upstream delivery node for a delivery node in a first level delivery sublayer that is directly connected to the content release layer, and the boundary service area on the boundary service layer represents is a downstream node for the delivery node in the delivery sublayer of the lowest level, which is directly connected to the boundary service layer. 7. A method for multilayer content delivery through a content delivery network, comprising the steps of: a) the ICP node distributes the content to be released to at least one delivery node on the content delivery layer, b) the delivery node receives a content delivery command that determines whether the delivery node is the delivery node of the lowest level delivery sublayer, and if so, performs step c), otherwise it reads the content to be released from the upstream node that is directly connected to the delivery node that has accepted the content delivery command, stores the content, distributes the content to one or more downstream delivery nodes that are directly connected to the delivery node that received the delivery command and content, and repeats the step of determining whether the delivery node is the delivery node of the lowest level delivery sublayer, c) in accordance with the adopted content delivery team, the delivery node receiving the content to be released from the upstream delivery node that is directly connected to the delivery node stores the content, distributes the content to one or more frontier service areas that are directly connected to the delivery node. 8. The method according to claim 7, in which step a) comprises the step of scheduling user access to ICP via MRB within the boundary service area in which the user is located, and providing the desired service to the user via CS or ES within the boundary service area where the user is located. 9. The method of claim 8, wherein the step of scheduling user access comprises at least one procedure from analyzing the ICP domain name to which the user accesses, discovering ES servers within the boundary service area and selecting an optimal ES. 10. The method of claim 9, wherein the step of detecting the ES servers includes at least one step of detecting how close the ES servers are to the user, whether the user requires the content on the ES servers, and whether the network is busy or inaction. 11. The method according to claim 9, in which the step of selecting the optimal ES comprises determining the optimal ES according to how close the ES servers are to the user, whether the user needs the content on the ES servers, or if the network is busy or inactive . 12. The method according to claim 7, in which when saving the content, the content is stored in the CS of the delivery node.

Images