CN100403293C - Network system method and protocol for hierarchical service and content distribution - Google Patents

Abstract

A network system method and protocol for hierarchical service and content distribution to improve the performance of content delivery network by utilizing the basic model design of hierarchical service network. According to the network system, a method and protocol are provided to illustrate how to design a network of multiple data centers under the implementation of existing Internet protocols and user-specific protocols, and the network allows users to implement in a scalable and fault-tolerant manner. Access to various Internet services, especially content delivery services. The network system is constructed into 4 layers, each layer is represented and managed by a service administrator, and the service administrator has a backup mirror administrator.

Description

Network system method and protocol for hierarchical service and content distribution

technical field

The present invention relates to a method and system for exchanging service routing information, and in particular to a method and system for hierarchical service and content distribution via a directory enabled network by agreement. A method and system for content distribution management to improve the performance of a content delivery network by using the basic design of a hierarchical service network.

Background technique

The web has now become one of the most effective and dominant media for B2B (business-to-business), B2C (business-to-consumer), and C2C (consumer-to-consumer) communications. The web model is based on centralized servers that deliver content or services to all points on the Internet. The surge in web traffic has thus caused congestion of many web servers and blocking of network traffic. Accordingly, content delivery networks are designed to require a plurality of cooperative, content-aware network devices that work with each other to cause distribution of content closer to the user and, on demand, to find the location of the content closest to the user. location.

Internet routing protocols, such as Border Gateway Protocol (BGP), are designed to exchange large Internet routes between routers. The Border Gateway Protocol, specified in external routing, is directional for a connection and is implemented on top of TCP, and will maintain adjacent connections through keep-alive messages and enable Consistent routing information synchronization. However, the Border Gateway Protocol will not exchange information within the central Internet of web servers. Therefore, it would be useful to have a routing protocol for services (in LLDAP as a directory format) to exchange service information in a hierarchical manner for decentralized management of services and content via a directory-enabled network, resulting in improved content Efficiency of delivery network and service provisioning and management.

Contents of the invention

The purpose of the present invention is to provide a network system with multi-layered nature, which can improve the performance of content delivery network through the following mode design of hierarchical service network.

Another object of the present invention is to provide a method and protocol so that a hop from a server to a client with an inversion can be delivered via a streaming notification to deliver quality content when the next hop is not available. In accordance with the aforementioned and other objectives, the present invention proposes a new network system and method for hierarchical service and decentralized management of content via a directory-enabled network.

To achieve the above purpose, the present invention provides a network system method and agreement for hierarchical service and content distribution, including:

at least one stratum 4 service administrator responsible for managing the multiple content delivery network;

at least one Tier 3 service administrator responsible for managing one of the content delivery networks having multiple data centers;

At least one Tier 2 service administrator responsible for managing one of the data centers with multiple server farms or service engine farms; and

At least one Tier 1 service administrator who, together with at least one Tier 2 service administrator, establishes a directory information routing protocol.

Each server of the server farm is connected by a local area network (LAN) Ethernet switching network, and the LAN Ethernet switching network supports layer 2 multicast operation.

Wherein each server of the server farm is connected with the wireless frequency band.

Among them, the data passing through the data center travels through the IPSEC channel to ensure privacy and security, and even form a virtual private network (VPN) in the data center.

wherein at least one stratum 1 service administrator is managed to establish a directory gateway agreement, and the directory gateway agreement is linked with at least one of the at least one stratum 2 service administrator to manage the at least one stratum 2 service administrator to establish a directory gateway agreement, and the directory gateway agreement is associated with at least one of the at least one layer 3 service administrator, the at least one layer 3 service administrator is administered to perform as domain name system (DNS) server, which directs user requests to different data centers as a regional load balancer, and also manages the service administrator at the origin of the server farm to establish a relationship with its parent service administrator. Linked directory gateway protocol.

The network system method and agreement for hierarchical service and content distribution provided by the present invention also include:

at least one Tier 4 service administrator responsible for managing multiple content delivery networks and storing content location information for the at least one content delivery network;

at least one Tier 3 service administrator responsible for managing one of the content delivery networks having multiple data centers, wherein each of the at least one Tier 3 service administrator stores content location information for responsive content delivery networks, and content information of the data center;

at least one Tier 2 service administrator responsible for managing one of the data centers having multiple server farms or service engine farms, wherein each of the at least one Tier 2 service administrators of the data center stores only the response data Center content location information; and

at least one tier 1 service administrator for establishing directory information routing protocols with at least one tier 2 service admin to manage each server farm, wherein the at least one tier 1 service admin and the at least one tier 2 service Managers are generated via LAN multicast and opaque link-state packets of link-state routing protocols filled with service information.

Each server of the server farm is connected by a LAN Ethernet switching network, and the LAN Ethernet switching network supports layer 2 multicast operation.

Wherein each server of the server farm is connected by an infinite frequency band.

Among them, the data passing through the data center travels through the IPSEC channel to ensure privacy and security, and even form a virtual private network (VPN) between the data centers.

wherein at least one stratum 1 service administrator is managed to establish a directory gateway agreement, and the directory gateway agreement is linked with at least one of the at least one stratum 2 service administrator to manage the at least one stratum 2 service administrator to establish a directory gateway agreement, and the directory gateway agreement is associated with at least one of the at least one layer 3 service administrator, the at least one layer 3 service administrator is administered to perform as domain name system (DNS) server, which directs user requests to different data centers as a regional load balancer, and also manages the service administrator at the origin of the server farm to establish a relationship with its parent service administrator. Linked directory gateway protocol.

The present invention provides a network system method and protocol for hierarchical service and content distribution, including at least one stratum 4 service administrator, at least one stratum 3 service administrator, at least one stratum 2 service administrator and at least one stratum 1 service manager members, and the method includes the following steps:

managing at least one content delivery network with multiple data centers and storing content location information for the at least one content delivery network;

manage data centers with multiple server farms or service engine farms; and

A directory information routing protocol is established between the at least one Tier 1 service administrator and the at least one Tier 2 service administrator and manages each server farm.

The method further comprises the step of establishing a directory gateway protocol link between the at least one layer 2 service administrator and the at least one layer 3 service administrator.

Description of drawings

The foregoing and other objects and advantages of the present invention will become apparent upon consideration of the following detailed description in conjunction with the accompanying drawings, in which all references are cited to similar parts, wherein:

Fig. 1 is to show the system according to the present invention, and is used for the content peer figure (Content Peering) of multiple CDN network;

Fig. 2 a is the integrated service network (Integrated Service Network) diagram showing multiple data centers (Multiple Data Centers) of the system according to the present invention;

FIG. 2b is another integrated service network diagram showing multiple data centers in the system according to the present invention;

Fig. 3 is a diagram showing a service administrator (ServiceManager) and an cache proxy server farm (Caching Proxy Server Farm) in the data center according to the system of the present invention;

Fig. 4 shows the directory information multicast update diagram (Directory Information Multicast Update) in the service administrator field in the system according to the present invention;

Figure 5a is a diagram showing a local area network (LAN) in the integrated service (Integrated Service) in the system according to the present invention;

Figure 5b is a program diagram showing the reliable multicast transport protocol sequence (Reliable Multicast Transport Protocol Sequence) in the system and method of the present invention;

Fig. 6 is a procedure diagram showing the transport multicast abort operation sequence (Transport Multicast abort operation sequence) in the system and method according to the present invention;

FIG. 7 is a sequence diagram showing the Reliable Multicast Directory Update Protocol Sequence (Reliable Multicast Directory Update Protocol Sequence) in the system and method according to the present invention;

FIG. 8 is a sequence diagram showing the Reliable Multicast Management Protocol Sequence (Reliable Multicast Management Protocol Sequence) in the system and method according to the present invention.

Detailed ways

network system layer

Please refer to FIG. 1 , FIG. 2 a and FIG. 2 b to illustrate the embodiment of the network system layer of the present invention. FIG. 1 is a diagram illustrating content peering for multiple Content Delivery Network (CDN) networks in accordance with the system of the present invention. This hierarchical directory enabled network (hierachical directory enabled network) provides secure content information and other forms of services.

Services from this network

web and streaming content decentralized services,

web and stream content hosting services,

Internet Communications Encryption Protocol (IPSEC) Virtual Private Network (VPN) service,

Managed Firewall Service

and any other new Internet Protocol services in the future.

This class can adjust the elements of Hierarchical Scalable Integrated Service Networks (HSISN)

a. Device

Integrated Service Switch (Integrated Service Switch, ISS):

Internet Protocol (IP) switch, IP traffic based on service and flow specifications.

Service engine (server):

A server system (or with special hardware attached) that handles Hyperfile Transfer Protocol (HTTP), cache memory, Internet communication encryption protocol, firewall or proxy, etc.

Service administrator:

A system designed to execute as an administrative agent and likewise as a Link Control Protocol (LDAP) server for the Link Control Protocol search service, and also with the system's parent service administrator and child The service administrator implements the directory gateway protocol to exchange directory information.

Link Control Protocol Description Language:

Definition of directory information exchanged by service managers and searched by LCP clients.

Simple Network Management Protocol (SNMP):

Definition of management messages, used between SNMP network managers and agents.

agreement

standard agreement

Existing routing protocols (Open Shortest Path First (OSPF), Border Gateway Protocol (BGP)) are implemented on the ISS to interoperate with other routers in the network.

Each server implements the LCP as a client; the service manager similarly implements the LCP to provide the services required by the service engine's LCP search.

invention agreement

Service Information Agreement [Related applications under review are applied separately]

Please refer to Figure 5a, which is implemented in a LAN or InfiniBand (a new I/O specification for servers) environment between the ISS, Service Engine, and Tier 1 Service Manager, so that

1. Register/unregister/update services and service attributes

2. Service Control Notification - Service Engine Congestion, Redirection, etc.

Can support an unlimited number of service engines (that is, the use of multiple boxes can have extremely high adjustability). Service Control Advertisements can be dynamically load balanced among Service Engines, since the ISS will transmit messages based on these advertisements to available (less congested) Service Engines. Keeping open communication between the ISS and the service administrator can assist in the detection of faulty devices, which in turn can remove the ISS from its list of available service engines.

Flow advertisement protocol (Flow advertisement protocol) [Related applications under review are applied separately]

Generated by service engine for ISS (Application Driven Flow or Chat Layer)

1. Establish flow in ISS to allow flow exchange

2. The flow has its flow attributes; one of the attributes is Quality of Service (QoS), other flow attributes are equally possible.

The fluidity attribute of the quality of service can enforce the requirements of data flow content quality delivery. The flow will be mapped by the ISS to the external network to existing or future standards such as MPLS, DiffServ, 802.1p, cable modem SID.

Assigned Numbers Authority protocol (Assigned Numbers Authority protocol) [Related applications under review are applied separately]

The protocol controls any kind of value assigned to the sub-network or area network or unlimited band if required globally. These values are IP address pools, MPLS label ranges, global interface values, HTTP data (kookies) and so on. In each sub-network, the designed service manager will be selected (representing the service engine farm including ISS). The service type will be expressed in a packet-shaped manner, so that different types of service engines can be mixed in the same subnet or area network, and all different types of service engines can be represented by the same service manager.

Directory Gateway Protocol (DGP)

Referring to Figure 1 and Figures 2a and 2b showing content peering for multiple content delivery network networks, and Figures 2a and 2b showing an integrated services network for multiple data centers, Directory Gateway Protocol is defined as Directory Information Routing Protocol . Directory Gateway Protocol utilizes similar concepts from the external routing protocol Border Gateway Protocol, except that the directory information is exchanged between the Directory Gateway Protocol parent and child, rather than between Border Gateway Protocol adjacent Internet Protocol Routing Switching. Similar to the Border Gateway Protocol, the Directory Gateway Protocol is link-oriented and implemented on top of the Transmission Control Protocol, and will maintain adjacent links by keeping the message open and, during the life of the link, make consistent Sexual directory information synchronization. But if the parent and child service managers try to initiate a directory interpreter protocol link at the same time, the directory interpreter protocol link is initiated from the parent service manager to the child service manager to avoid any link conflicts. To avoid any transmission loops, connections between service managers at the same level are not allowed. Although for duplication reasons it is possible for multiple backup parent service administrators to link to the same child service administrator to provide child service administrator LDA search services, only if the parent service administrator Connections with subservice administrators are only allowed.

A stratum 1 service administrator (representing a service subnetwork) will establish a directory explorer protocol link with its parent service administrator (stratum 2 service administrator). Typically a Tier 2 service administrator will act on behalf of the entire data center.

The stratum 2 service administrator will also establish a directory explorer protocol link with its parent service administrator (stratum 3 service administrator). Usually the service administrator of the origin server farm will also establish a directory protocol link with its parent service administrator (stratum 2 or stratum 3 service administrator).

A Tier 3 service manager typically performs as a domain name system server, and the service manager directs user demand to different data centers as a regional load balancer. The DNS redirection decision may be based on the availability of service load attributes that may be updated by the service data center via incremental updates of the directory resolver protocol, and may be based on other attributes such as proximity to the user Determines domain name service redirection.

The initial directory interpreter protocol link will exchange directory information based on the individual other directory information transfer principles; after the initial exchange, each service manager will only incrementally update (add or drop) its Directory information services and service attributes, content and content attributes, etc. to the other side. One of the service attributes is the load factor (response time) of the service domain, and one of the content attributes is the content location including the cached content location. Directory interpreter protocol packets are in the form of open, lightweight directory access protocol (LDAP) append, LDAP_delete, LDAP_modify_attach, LDAP_modify_replace, LDAP_modify_delete, notify and keep open.

Considering a content change as a content attribute (content time) change for the content, the content change will be transmitted to the caching server with cached content (see the section on caching content invalidation sequence for details). For frequently changing content, the (similar to Border Gateway Protocol) Directory Gateway Protocol supports directory information damping, which dampens the delivery of frequently changing directory information. Similar to the Border Gateway Protocol, the Directory Gateway Protocol also supports policy-based transfers between its parent and child service managers. The best way to aggregate directory information is to apply an aggregation policy prior to transfer. Also similar to BGP, TCP-mini-disc 5 can also be used for authentication.

Calculation of proximity

As previously mentioned, this technique is used with service load attributes that are updated by each data center so that the DNS server directs the user's needs to the best service data center for regional load balancing. Every IP destination (IP route, address, and mask) will be assigned an (x, y) attribute, where x represents longitude (between -180 and +180, but -180 and +180 are actually is the same location, this is because the earth is round), and y represents the latitude of the earth (between -90 and +90), and the destination of this IP can be physically positioned.

Assume that the source address of the user matches the longest prefix of an IP destination with the attribute (x1, y1), and the attribute of the IP address prefix of the data center is (x2, y2).

If |x1-x2|<=180, then the distance between the user and the data center is

((x1-x2) ² +(y1-y2) ² ) ^1/2

If |x1-x2|>180, then the distance between the user and the data center is

((360-|x1-x2|) ² +(y1-y2) ² ) ^1/2

The (x, y) routing attribute can be suggested to the Internet engineering task promotion team as an extension of the border gateway protocol routing attribute.

Reliable Multicast Transport Protocol

Referring to Figure 4, the multicast update of directory information within the service management field, and Figure 5b, the reliable multicast transport protocol sequence, in order to simultaneously update information to service devices in a multicast-capable network, and to improve performance, the reliable A multicast transport protocol is used for this purpose. The protocol is similar to Transmission Control Protocol, but it has a two-way (send and reply handshake) feature instead of a three-way handshake defined between sender and receiver to establish a connection. After that, the service administrator is responsible for specifying the size of the window (in the packet) so that the sender can send the message without confirmation. The size of the viewport is one of the service attributes that individual service engines register with the service administrator. The service administrator will choose the lowest value among the service attributes of the registered window size for each receiver. At the end of each window, the Service Manager is also responsible for acknowledging the receipt on behalf of all other recipients. A better way is that the service manager will wait for a short silent time (can be a configurable value) before sending an acknowledgment of receipt. If any unprogrammed packet reception is detected, or if no packet is received within a certain period of time, causing the waiting time to be too long and stopping, the receiver should send a resend request from the initial program value (used by the window) . The sender can choose to resend from the specified resend program value, or abort the connection and restart it again. Unless the connection is broken, the receiver will simply drop the received packet. The final packet should be acknowledged not only by the service manager, but also by all receivers, so as to indicate a normal disconnection of the link. If the service manager checks out that any receiver did not acknowledge the last packet during the pause, the service manager will request to resend the last packet to the receiver (unicast packet). If more than three resends have been attempted, the device is considered dead and removed from the service engine list by the service administrator. If only one packet is to be sent, the protocol becomes a reliable data gram protocol. Defines the window size for unacknowledged unprocessed packets. Both the acknowledgment and the resend request are multicast packets that allow the service administrator to monitor.

Reliable Multicast Directory Update Protocol

As shown in Figure 7, it is a reliable multicast directory update protocol. It is implemented on top of the reliable multicast transport protocol. This protocol is similar to Lightweight Directory Access Protocol implemented on top of Transmission Control Protocol, except that the transport layer is a reliable multicast transport protocol.

Reliable Multicast Management Protocol

Referring to FIG. 8 , shown is an RMM protocol sequence implemented on top of an RMM transport protocol. Since there is only one packet to send, this protocol becomes a reliable multicast segment protocol. The protocol is similar to the Simple Network Management Protocol implemented on Ethernet, except that it has a transport layer to provide multicast and reliability services.

Hierarchical management information and management methods

The management agent is formed as part of the service administrator. For policy-based service management, management information is defined at different levels. The collection of management information is passed from one level to another. For example, the value of web page hits can be recorded with a counter on each cache service engine, or can be recorded with a full counter in the entire stratum 1 service engine farm, or can be recorded with a full counter in the overall data center .

As far as configuration management information is concerned, the same definitions are used for configurations at different levels. For example, the default routing configuration can only be used for the same subnet, while the DNS server can be used for the entire data center. The stratum 1 service administrator is responsible for multicasting the default routing configuration to the entire subnetwork, while the stratum 2 service administrator has the domain name system (DNS) server configuration containing its data center hierarchy configuration sent to the stratum 1 service administrator. Then, the stratum 1 service manager needs to multicast its components in its subnetwork. Lower-level configurations or policies cannot conflict with higher-level policies; if they conflict, higher-level policies should take precedence over lower-level policies.

Directory Description Language and SNMP MIB

Several directory information description languages and SNMP MIBs need to be defined to support Hierarchical Scalable Integrated Service Networks (HSISN).

Web Site object

Web Content object

Service Engine object

Integrated Service Switch object

User object

and other objects

Take the following consistent resource locator as an example

Site object (raw or cached site)

original website

DN (Domain Name System, Distinguished Name): http, vision, yahoo, com Attributes:

Service website Internet Protocol (IP) address:

caching service site

DN (Domain Name System, Distinguished Name): subnet, Datacenter2, CDN3 properties:

Service website Internet Protocol (IP) address:

Generation of new entries for website objects

The origin site will send DGP with additional DN: http.vision.yahoo.com to the stratum 3 service administrator (also a DNS server) in order to add a new entry .

Modification of website object entry

According to the service level agreement, the level 3 service administrator will send the DGP Lightweight Directory Access Protocol of the service website address_modify_to add the entry attribute of the website object, and these Internet Protocol (IP) addresses will be appended to vision.yahoo .com's DNS entry list.

Yahoo's DNS server responsible for vision.yahoo.com should refer DNS requests for vision.yahoo.com to the DNS in the stratum 3 service manager. The DNS of the stratum 3 service administrator will respond to the user with the IP address of the service website with the lowest service standard, or according to other policies. A cached site selected based on the best response from the cached site to the user

An example of a Yahoo site with a video-based financial page:

The ISP's DNS server will refer to Yahoo's DNS server for vision.yahoo.com. Yahoo's DNS server will refer to the Tier 3 service administrator of the content distribution service provider.

Each data center may have one or more service websites, and each service website may be served by a server with a virtual IP address. If you find that many cached sites for vision.yahoo.com are available (for example, one of them has a site at 216.136.131.74 and another site at 216.136.131.99), and designate all of them to serve vision. yahoo.com. The stratum 3 service administrator's DNS will have multiple entries for vision.yahoo.com. It will select one of the sites as the policy-based DNS answer (weighted round-robin or service metric from those sites to the user). Assume that DNS picks 216.136.131.74 to respond to the user.

The user sends an http request such as

Service metric

The service standard from user 1 to website 1 is the current average server service response time of website 1 + weighted proportion * current proximity from user 1 to website 1 . The weighting is assembled on a policy basis. Website 1 calculates the current proximity by the formula described above. The website 1 of the layer 1 service administrator will receive the response time of each server in the keep-alive message from the server received by the service engine, so as to calculate the current average service response time from the server as the load factor of this website.

Site content object (in the original or quick access site)

Domain Name System (DN): fv.html, ie, web, http, vision, yahoo, com Attributes:

Original Content Location: The Internet Protocol (IP) address of the original server

Cache content location: DN of cache service site 1, number of cache service engines with this content in site 1, DN of cache service site 2, number of cache service engines with this content in site 2 , the DN of the cache service website 31, the number of cache service engines that have this content in the website 31, the DN of the cache service website 41...

To cache the media access control (MAC) address of the content serving engine in the stratum 1 service manager:

Service Engine 1MAC (applicable only to stratum 1 service administrator),

Service Engine 2MAC (Applicable to Stratum 1 Service Manager only),

The number of caching engines that have this cached content

Date and time the content was last modified:

Date and time the content expires:

service engine object

DN: Internet Protocol (IP) address, Subnet 1, Data Center 2, Content Delivery Network (CDN)

Attributes:

Service form:

Service engine name:

Service engine subnet mask:

Service Engine Media Access Control (MAC) address:

Service Engine Security Policy: If it is a different data center, use a secure socket layer

Service administrator IP address:

Service engine certificate:

Integration Services Exchange Objects

DN: IP address on the server farm interface, Subnet 1, Data Center 2, CDN3 Properties:

Exchange form:

Swap IP addresses:

Exchange MAC addresses:

Service administrator IP address:

Exchange certificates:

user object

DN: name, organization, country

Attributes:

Mailing address:

email address:

User certificate:

Account records:

Creation and modification of new entries for website content objects

According to the service agreement, the original website will send DGP LDAP_MODIFY_ADD DN: fv.html, ie, web, http, vision.yahoo.com to the stratum 3 service administrator. When the DNS sends 216.136.131.74 as a response, the user will send http requests such as

According to another patent invented by us, the integrated service exchange of the virtual IP address will direct the request to the less congested cache service engine, for example, engine 1 is selected. If the content is not in cache engine 1, the integrated service switch of this virtual IP address will send the LDAP search request to its stratum 1 service manager. If the stratum 1 service administrator does not have the content, it refers to its stratum 2 service administrator. If the stratum 2 service administrator does not have the content, it refers to its stratum 3 service administrator. The layer 3 service administrator will return the attribute of the original server IP address to indicate whether it can be quickly retrieved, as well as other content attributes. If the requested content information cannot be quickly obtained, the cache engine will use http to redirect the user to the original server.

If the content information of the request is cacheable content, then the cache engine will initiate a new http session layer to the origin server on behalf of the user. And if the http response from the origin server also specifies "cacheable", then it will cache the content. Real-time streaming protocols may also support the redirection message, but other existing application protocols may not always support the redirection message. Once the content is cached, it adds the LDAP_ADD_DN:fv.html, ie, web, http, vision.yahoo.com object to the Stratum 1 service administrator. If the object is not found in the stratum 1 service manager, then add the DN: fv.html, ie, web, http, vision.yahoo.com (DN of the service engine) together with its own cached content location attribute. If the object is found in the stratum 1 service manager, the object is modified and added with the new cached content location attribute. The stratum 1 service administrator will then execute DGP LDAP_ADD or DGP LDAP_MODIFY_ADD DN:fv.html,ie,web,http,vision.yahoo.com to the stratum 2 service administrator. The stratum 2 service administrator will then execute DGP LDAP_ADD or DGP LDAP_MODIFY_ADD DN:fv.html,ie,web,http,vision.yahoo.com to the stratum 3 service administrator.

The update of the cache location directory information is a triggered update operation, and the update should be much faster than the periodic synchronization process used in the existing replication process between LDAP servers.

Content retrieval from nearest location (raw or cached)

Retrievals from neighboring caching service engines are managed by the same Tier 1 service manager on the same LAN. If another user sends an http request, and the http request is transmitted to the service engine 2 by the integrated service exchange, and the service engine 2 is under the same level 1 service administrator as the service engine 1 (also as a lightweight Level Directory Access Protocol (LDAP) server) to manage. When a service engine 2 that does not have the content searches from its stratum 1 service manager LLDAP_Search, the stratum 1 service manager will reply with its attributes and the service engine 1 as the content cache location.

Because it is fetched content, the Serving Engine 2 will initiate a new http session layer to Serving Engine 1 on behalf of the user, rather than the original server. And in addition to returning content to its users, it will also cache content. Once the content is cached, Service Engine 2 will LDAP_ADD to the same Stratum 1 Service Manager (also used as the LDAP server). If the entry already exists, the service engine 2 will LDAP_MODIFY_ADD to add another cached location (itself) to the content attribute.

Retrieval from neighboring sites is managed by the same stratum 2 service administrator for the entire data center, if another user sends an http request to the second service site, and the http request is sent by 216.136.131.99 The integrated service exchange of the service website is transmitted to the service engine 31 . When there is no service engine 31 of the content from its stratum 1 service manager LDAP_SEARCH, and the stratum 1 service manager does not have the content, and then refers to the stratum 2 service manager, the stratum 2 service manager will reply as the cache location 216.136.131.74 website, attribute of the service engine number containing the content. If there are two or more websites with the content, the website with more service engines with the content will be selected. The service engine 31 will then initiate a new http session layer to 216.136.131.74 on behalf of the user, rather than the original server. And in addition to responding to content to its users, the serving engine 31 will also cache the content. Once the content is cached, the service engine 31 ADDs the Lightweight Directory Access Protocol (LDAP) to its stratum 1 service administrator (also acting as an LDAP server). If the entry does not exist, the stratum 1 service administrator will add the DN: fv.html, ie, web, http, vision.yahoo along with its own cached content location (media access control (MAC) address) attribute .com. And the stratum 1 service administrator of the service engine 31 also sends the Directory Gateway Protocol (DGP) LDAP_ADD DN: fv.html, ie, web, http, vision.yahoo.com to the stratum 2 service administrator. If an entry is found, the Tier 2 Service Manager will modify to append another cached location (itself) to the content attribute and increment the number of websites that own the content.

Retrieval from adjacent data centers is managed by the same Tier 3 service administrator used throughout the CDN (Content Delivery Network). If there is a second service site set up in another data center, and if that data center does not already have such cached content, LDAP_SEARCH will eventually refer to the stratum 3 service administrator to find the cached data center location. If a neighboring data center has the cached content, then the http proxy will initiate to its neighboring data center instead of the origin server on behalf of a user from a data center's caching engine. In case several data centers have the cached content, the number of caching engines (in the data center) that have the cached content determines the preference.

The service engine can dynamically discover the LDAP server it refers to as its stratum 1 service administrator. Depending on whether a link-state routing protocol (such as OSF) is running or not, a stratum 1 service manager may or may not need a static configuration to find its stratum 2 service manager. If link-state routing protocols are being implemented, opaque link-state packets can be used to carry service manager information and be used to flood the routing domain. LDAP search results are also affected by policy configuration. It is also possible to add policy management related to content attributes such as proxies or redirects, if content information can be quickly obtained, life time can be quickly retrieved, etc.

cached content invalid

When the original server modifies the content of DN: fv.html, ie, web, http, vision.yahoo.com, it will LDAP_MODIFY_DELETE to remove all cached content from stratum 3 service administrators Location. Otherwise, it can be specified to direct the scheduled content update, or through DGP, to change the expiration date attribute of the content to perform the update action. The Tier 3 service administrator then LDAP_MODIFY_DELETE to remove all cached content locations from the Tier 2 service administrator it manages, or to change the expiration date.

The stratum 2 service administrator will then LDAP_MODIFY_DELETE to remove all cached content locations from the stratum 1 service administrator it manages, or change the expiration date. And the stratum 1 service manager will notify (multicast) all its caching service engines to remove the cached content from their storage areas.

When the content is scheduled to be changed by the original server, the original server will also send LDAP_MODIFY_SUBSITE to modify the last modified date and time attribute of the content in the stratum 3 service manager, and then transmit it down to the lower Low level service manager and cache service engine. Based on the date and time of the last modification, the server decides when to discard old content.

Dynamic discovery in service engine (LDAP client), stratum 1 service administrator and stratum 2 service administrator In (layer 2) area network environment, layer 2 multicast can be used from all services The engine passes service information to the stratum 1 service manager. Well-known Ethernet multicast URLs will be defined for the stratum 1 service administrators (primary and backup stratum 1 service administrators).

In the field of link-state routing, opaque link-state packet flooding will be used by all stratum 1 service administrators and stratum 2 service administrators to transport the service engine and the services it provides within a domain or an autonomous system.

Tier 2 service administrators should always flood the entire autonomous system. If the entire autonomous system has only one stratum 2 service manager, then the opaque link state packets from the stratum 1 service manager should be flooded throughout the autonomous system. If each zone has a stratum 2 service manager, then opaque link state packets from the stratum 1 service manager should only be flooded to that zone. Although DGP links to other same-level service managers are not allowed, the level-2 service manager can first refer to the level-2 service manager before referencing the level-3 service manager for date information.

More than one autonomous system may utilize Internet Protocol (IP) multicast to transport services within an IP multicast tree between layer 2, layer 3 or layer 4 service managers. Static configuration can be used to transfer, find and update services between service administrators.

The hop flow notification protocol for quality content delivery IP flows from caching service engines to clients with inverses via hops is described according to pattern-matching rules. After checking or initiating the ID and account, flow notification will start from the caching service engine to its upstream aggregation service exchange, and if the flow notification protocol is supported, the flow notification will continue to advertise the flow to Its upstream adjacent integration service switches and jumps to the end user. But the end user does not need to be included in the mobile notification protocol. In case the flow advertisement protocol is not supported, each hop will map flows and flow attributes into its (possibly different) upstream traffic characteristics via static configuration or signaling protocols. For example, IP flows can be mapped to Asynchronous Transfer Mode (ATM) switched virtual circuits (SVCs) or permanent virtual circuits (PVCs), and ATM SVCs or PVCs can also be mapped to Internet protocol flows via the hop flow advertisement. If IP MPLS is also available, IP flow notifications can be mapped to MPLS via the MPLS signaling protocol. If the upstream hop does not support any streaming signals, streaming notifications will stop.

Flow exchanges need to cover every hop as long as flows can be mapped and defined, and should try to cover all network elements from layer 2 to layer 7. If only traffic classes are defined, then downstream hops should still attempt to map the appropriate traffic class upstream. Typical examples of quality of service can map any useful part to the upstream network such as DiffServ, single-line encapsulation of cable modems and 802.1p.

In the event that a connection or exchange is interrupted along the flow path, the upstream hop should interrupt the flow by sending a Flow Abort Notification to its further upstream neighbor and transmitted to the end user. On the other hand, the downstream hop should initiate another flow advertisement to other available upstream hops and further to the end user to re-establish the flow. If no upstream hop will accept the flow, the exchange should interrupt the flow and notify the flow to stop (reverse) to its downstream hops, which should find other available upstream hops so that redelivery attempts are attempted to end users. If the upstream hop is no longer available, it should continue to advertise flow aborts (reversals) to its downstream hops until a usable exchange is found, or just go back to the service engine which will drop the flow.

Virtual Private Network (VPN) with PKI

A VPN with PKI can use the same directory enabling network as a non-content-related service engine, such as an IPSEC engine. A VPN with PKI can refer to its stratum 1 service administrator for certificates and the like. It also references Tier 2 and Tier 3 service administrators for hierarchical user and account management.

Claims (10)

1. A network system for hierarchical service and content distribution, the system comprising: at least one stratum 4 service administrator responsible for managing the multiple content delivery network; at least one Tier 3 service administrator responsible for managing one of the content delivery networks having multiple data centers; At least one Tier 2 service administrator responsible for managing one of the data centers with multiple server farms or service engine farms; and at least one stratum 1 service administrator who, together with at least one stratum 2 service administrator, establishes a directory information routing protocol, Wherein, at least one layer 1 service administrator is managed to establish a directory gateway agreement, and the directory gateway agreement is linked with at least one of the at least one layer 2 service administrator, and the at least one layer 2 service administrator an administrator is administered to establish a directory gateway agreement associated with at least one of the at least one layer 3 service administrator, the at least one layer 3 service administrator is administered to perform as a domain name A system server that directs user requests to different data centers as a regional load balancer, and also manages the service administrator at the origin of the server farm to establish a link with its parent service administrator Directory Gateway Protocol. 2. The network system according to claim 1, wherein each server of the server farm is connected by a switching network of the Ethernet in the local area network, and the switching network of the Ethernet in the local area network Used to support multicast operations in Layer 2. 3. The network system according to claim 1, wherein each server in the server farm is connected to the wireless frequency band. 4. The network system according to claim 1, wherein the data passing through the data center travels through an IPSEC channel, and a virtual private network can also be formed in the data center. 5. A network system for hierarchical service and content distribution, the system comprising: at least one Tier 4 service administrator responsible for managing multiple content delivery networks and storing content location information for the at least one content delivery network; at least one stratum 3 service administrator responsible for managing one of the content delivery networks having multiple data centers, wherein each of the at least one stratum 3 service administrator stores responsive content location information for the content delivery network, and the content information of the data center; at least one stratum 2 service administrator responsible for managing one of the data centers having multiple server farms or service engine farms, wherein each service administrator of at least one stratum 2 of the data center stores only the response data Center content location information; and at least one Tier 1 service administrator for establishing directory information routing protocols with at least one Tier 2 service administrator to manage each server farm, wherein the multicast and link state routing protocols are flooded with service information via the local area network an opaque link state packet that generates the at least one stratum 1 service manager and the at least one stratum 2 service manager, Wherein, at least one layer 1 service administrator is managed to establish a directory gateway agreement, and the directory gateway agreement is linked with at least one of the at least one layer 2 service administrator, and the at least one layer 2 service administrator an administrator is administered to establish a directory gateway agreement associated with at least one of the at least one layer 3 service administrator, the at least one layer 3 service administrator is administered to perform as a domain name A system server that directs user requests to different data centers as a regional load balancer, and also manages the service administrator at the origin of the server farm to establish a link with its parent service administrator Directory Gateway Protocol. 6. The network system according to claim 5, wherein each server of the server farm is connected by an Ethernet switching network in the local area network, and the Ethernet switching network in the local area network Used to support multicast operations in Layer 2. 7. The network system as claimed in claim 5, wherein each server in the server farm is connected by wireless frequency band. 8. The network system according to claim 5, wherein the data passing through the data centers travels through IPSEC channels to ensure privacy and security, and a virtual private network can be formed between data centers. 9. A network system method for hierarchical service and content distribution comprising at least one stratum 4 service administrator, at least one stratum 3 service administrator, at least one stratum 2 service administrator and at least one stratum 1 service administrator, And this method then comprises the following steps: managing at least one content delivery network with multiple data centers and storing content location information for the at least one content delivery network; manage data centers with multiple server farms or service engine farms; and A directory information routing protocol is established between the at least one Tier 1 service administrator and the at least one Tier 2 service administrator and manages each server farm. 10. The method of claim 9, further comprising the step of establishing a DGP link between the at least one layer 2 service administrator and the at least one layer 3 service administrator.

Images