WO2021047678A1 - 算力路由方法及装置 - Google Patents

算力路由方法及装置 Download PDF

Info

Publication number
WO2021047678A1
WO2021047678A1 PCT/CN2020/115154 CN2020115154W WO2021047678A1 WO 2021047678 A1 WO2021047678 A1 WO 2021047678A1 CN 2020115154 W CN2020115154 W CN 2020115154W WO 2021047678 A1 WO2021047678 A1 WO 2021047678A1
Authority
WO
WIPO (PCT)
Prior art keywords
computing power
computing
router
code
fid
Prior art date
Application number
PCT/CN2020/115154
Other languages
English (en)
French (fr)
Inventor
郑若滨
Original Assignee
华为技术有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 华为技术有限公司 filed Critical 华为技术有限公司
Priority to EP20862333.0A priority Critical patent/EP4024785A4/en
Publication of WO2021047678A1 publication Critical patent/WO2021047678A1/zh
Priority to US17/692,868 priority patent/US20220200898A1/en

Links

Images

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L45/00Routing or path finding of packets in data switching networks
    • H04L45/44Distributed routing
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L45/00Routing or path finding of packets in data switching networks
    • H04L45/30Routing of multiclass traffic
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L45/00Routing or path finding of packets in data switching networks
    • H04L45/74Address processing for routing
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L63/00Network architectures or network communication protocols for network security
    • H04L63/08Network architectures or network communication protocols for network security for authentication of entities
    • H04L63/0807Network architectures or network communication protocols for network security for authentication of entities using tickets, e.g. Kerberos
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L63/00Network architectures or network communication protocols for network security
    • H04L63/12Applying verification of the received information
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L9/00Cryptographic mechanisms or cryptographic arrangements for secret or secure communications; Network security protocols
    • H04L9/32Cryptographic mechanisms or cryptographic arrangements for secret or secure communications; Network security protocols including means for verifying the identity or authority of a user of the system or for message authentication, e.g. authorization, entity authentication, data integrity or data verification, non-repudiation, key authentication or verification of credentials
    • H04L9/3247Cryptographic mechanisms or cryptographic arrangements for secret or secure communications; Network security protocols including means for verifying the identity or authority of a user of the system or for message authentication, e.g. authorization, entity authentication, data integrity or data verification, non-repudiation, key authentication or verification of credentials involving digital signatures
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L9/00Cryptographic mechanisms or cryptographic arrangements for secret or secure communications; Network security protocols
    • H04L9/50Cryptographic mechanisms or cryptographic arrangements for secret or secure communications; Network security protocols using hash chains, e.g. blockchains or hash trees
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L45/00Routing or path finding of packets in data switching networks
    • H04L45/12Shortest path evaluation
    • H04L45/124Shortest path evaluation using a combination of metrics

Definitions

  • This application relates to the field of communication technology, and in particular to a computing power routing method and device.
  • This application provides a computing power routing method and device to solve the problem of low-cost sharing of traffic.
  • an embodiment of the present application provides a computing power routing method, including:
  • the computing power router receives computing power routing information of at least one computing container, where the computing power routing information includes a computing power code or a calculated metric value;
  • the computing power router determines a routing information base RIB according to the computing power routing information, and the RIB includes one of the function identification FID of the function that the computing container can provide or the IP address prefix of the computing container where the function is located, and the computing power One of the force code or the calculated metric value, and the function includes any one of a softwareized calculation program, a functional program, business software, or a microservice deployed in the calculation container;
  • the computing power router receives the message, forwards the message according to the routing information database FIB, the FIB is determined according to the RIB, and the FIB includes the FID or the IP address prefix.
  • the computing power routing information of at least one computing container is received through the computing power router.
  • the computing power routing information includes the computing power code or the calculated metric value, and then the computing power routing information is sent to the network Other computing power routers then determine the RIB based on the computing power routing information.
  • the computing power router receives the message, it forwards the message according to the FIB and finally routes it to the corresponding computing container.
  • the FIB is determined based on the RIB, and the computing power routing information represents The computing power that the computing container can provide, so that the packets related to computing tasks can be routed to the corresponding computing container to achieve the best user experience, computing resource utilization, and network efficiency.
  • the routing process is equivalent to computing power bidding and traffic Users gain computing power in bidding, and traffic owners release computing power, which dynamically and quickly matches the traffic supply and demand relationship between operators and Internet companies, realizing low-cost sharing of traffic.
  • the method further includes:
  • the computing power router sends the computing power routing information to other computing power routers in the network.
  • the method further includes:
  • the computing power router determines the FIB item corresponding to the message according to the computing power code in the RIB or the calculated metric value.
  • the RIB includes FID, calculated metric value, IP address prefix, and outbound port; or,
  • IP address prefix calculated metric value and outgoing port
  • IP address prefix computing power code and outgoing port
  • the FIB includes: FID, IP address prefix, and outbound port; or,
  • the forwarding the message according to the routing information base FIB includes:
  • the computing power router searches the FIB for the IP address prefix and the outgoing port corresponding to the FID;
  • the computing power router adds the FID to the IP address prefix corresponding to the FID to obtain the destination IP address, and forwards the message according to the destination IP address; or,
  • the computing power router finds out the IP address and outgoing port of the computing power router corresponding to the FID from the FIB;
  • the computing power router uses the computing power router IP address as the tunnel destination address to add tunnel encapsulation to the message, or uses the computing power router IP address as the segment identifier of the segment routing SR to add SR encapsulation;
  • the computing power router forwards the packet of adding tunnel encapsulation or adding SR encapsulation; or,
  • the computing power router finds out the outgoing port corresponding to the FID or IP address prefix from the FIB;
  • the computing power router forwards the message from the outgoing port.
  • the forwarding of the message according to the routing information base FIB includes:
  • the computing power router forwards the message according to the IP address.
  • the computing power encoding includes computing latency evaluation value and/or computing capacity level, machine type or weighted value ⁇ 2 related to the machine type, and server load
  • the computing power routing information includes computing power
  • the method further includes:
  • the computing power router calculates the calculation metric value according to the computing power code and the following formula:
  • ⁇ 2 and ⁇ 2 are weighted values.
  • the method further includes:
  • the computing power router determines the FIB item as a valid FIB item
  • an embodiment of the present application provides a computing power routing method, including:
  • a leaf router obtains a computing power code of a computing container, and the leaf router is deployed in the computing container;
  • the leaf router sends the computing power routing information of the computing container to the computing power router connected to the leaf router, so that the computing power router sends the computing power routing information to other computing power routers in the network,
  • the computing power routing information includes the computing power code or the calculated metric value;
  • the leaf router receives the message and forwards the message.
  • the computing power code of the computing container is obtained through the leaf router, and computing power routing information including the computing power code or the calculated metric value is sent to the computing power router connected to the leaf router, so that the computing power
  • the power router sends the computing power routing information to other computing power routers on the network, and finally the leaf router receives the message, forwards the message, and specifically forwards the message to the computing container.
  • the computing power routing information represents the computing power that the computing container can provide, so that the packets related to computing tasks can be routed to the corresponding computing container to achieve the best user experience, computing resource utilization, network efficiency, and routing process equivalent
  • the users of the traffic obtain the computing power in the bidding, and the traffic owner releases the computing power, which dynamically and quickly matches the traffic supply and demand relationship between the operators and the Internet companies, and realizes the low-cost sharing of traffic.
  • the leaf router acquiring the computing power code of the computing container includes:
  • the leaf router obtains the computing power code from the calculation container and/or the measurement tool, and the computing power code includes a calculation delay evaluation value and/or a calculation capacity level, a machine type, or a weighted value ⁇ 2 related to the machine type. And server load, the calculation delay evaluation value is measured by the measurement tool.
  • the method further includes:
  • the leaf router calculates the calculated metric value according to the computing power code and the following formula:
  • ⁇ 2 and ⁇ 2 are weighted values.
  • the method further includes:
  • the leaf router calculates hashrate tokens according to the hashrate code and the following formula:
  • Hash power token ⁇ 1 ⁇ [ ⁇ 1-(calculated metric value)]
  • ⁇ 1 and ⁇ 1 are the power pricing coefficients.
  • the computing power routing information further includes:
  • the function identifier FID of the function that the computing container can provide the IP address prefix or IP address of the computing container where the function is located, and the function is a software-based computing program, functional program, and business software deployed in the computing container Or microservices.
  • the method further includes:
  • the leaf router triggers the contract platform to transfer the computing power token to the computing power user.
  • the method further includes:
  • the leaf router sends the response message to the client through the computing power router connected to the leaf router, and the response message is used by the client to change the FID carried in the message to the Calculate the IP address of the container.
  • an embodiment of the present application provides a computing power routing device, including:
  • a receiving module configured to receive computing power routing information of at least one computing container, where the computing power routing information includes a computing power code or a calculated metric value;
  • the determining module is configured to determine a routing information base RIB according to the computing power routing information, where the RIB includes one of the function identification FID of the function that the computing container can provide or the IP address prefix of the computing container where the function is located, and the computing power One of the force code or the calculated measurement value, and the function includes any one of a softwareized calculation program, a functional program, business software, or a microservice deployed in the calculation container;
  • the receiving module is also used to receive messages
  • the sending module is configured to forward the message according to the routing information database FIB, the FIB is determined according to the RIB, and the FIB includes the FID or the IP address prefix.
  • the sending module is also used to:
  • the determining module is further configured to:
  • the RIB includes FID, calculated metric value, IP address prefix, and outbound port; or,
  • IP address prefix calculated metric value and outgoing port
  • IP address prefix computing power code and outgoing port
  • the FIB includes: FID, IP address prefix, and outbound port; or,
  • the sending module is used to:
  • the computing power encoding includes computing latency evaluation value and/or computing capacity level, machine type or weighted value ⁇ 2 related to the machine type, and server load
  • the computing power routing information includes computing power
  • the device further includes:
  • the calculation module is configured to calculate the calculation metric value according to the calculation power code and the following formula:
  • ⁇ 2 and ⁇ 2 are weighted values.
  • the device further includes:
  • a processing module configured to determine the FIB item as a valid FIB item if the receiving module receives a response message
  • an embodiment of the present application provides a computing power routing device, including:
  • An obtaining module configured to obtain a computing power code of a computing container, and the leaf router is deployed in the computing container;
  • the sending module is used to send the computing power routing information of the computing container to the computing power router connected to the leaf router, so that the computing power router sends the computing power routing information to other computing power routers in the network ,
  • the computing power routing information includes the computing power code or the calculated metric value;
  • the receiving module is used to receive messages
  • the sending module is also used to forward the message.
  • the acquisition module is used to:
  • the computing power code from the calculation container and/or measurement tool, where the computing power code includes a calculation delay evaluation value and/or a calculation capacity level, a machine type or a weighted value ⁇ 2 related to the machine type, and server load,
  • the calculation delay evaluation value is measured by the measurement tool.
  • the device further includes:
  • the calculation module is used to calculate the calculation metric value according to the calculation power code and the following formula:
  • ⁇ 2 and ⁇ 2 are weighted values.
  • the calculation module is also used to:
  • Hash power token ⁇ 1 ⁇ [ ⁇ 1-(calculated metric value)]
  • ⁇ 1 and ⁇ 1 are the power pricing coefficients.
  • the computing power routing information further includes:
  • the function identifier FID of the function that the computing container can provide the IP address prefix or IP address of the computing container where the function is located, and the function is a software-based computing program, functional program, and business software deployed in the computing container Or microservices.
  • an embodiment of the present application provides a device, including: a memory and a processor;
  • a memory for storing executable instructions of the processor
  • the processor is configured to execute the computing power routing in any possible design of the first aspect and the first aspect or any possible design of the second aspect and the second aspect by executing the executable instruction method.
  • the above-mentioned device may be a routing device or a chip.
  • the present application provides a computer-readable storage medium having instructions stored in the computer-readable storage medium, which when run on a computer, cause the computer to execute the methods described in the foregoing aspects.
  • the present application provides a computer program product containing instructions, which when run on a computer, causes the computer to execute the methods described in the above aspects.
  • Figure 1 is a schematic diagram of a computing power network architecture provided by this application.
  • FIG. 2 is a schematic diagram of another computing power network architecture provided by this application.
  • FIG. 3 is a schematic diagram of another computing power network architecture provided by this application.
  • FIG. 4 is a schematic diagram of an application scenario of the computing power routing method and device provided by this application.
  • FIG. 5 is a flowchart of an embodiment of a computing power routing method provided by this application.
  • FIG. 6 is a flowchart of an embodiment of a computing power routing method provided by this application.
  • FIG. 7 is a flowchart of an embodiment of a computing power routing method provided by this application.
  • FIG. 8 is a flowchart of an embodiment of a computing power routing method provided by this application.
  • FIG. 9 is a flowchart of an embodiment of a computing power routing method provided by this application.
  • FIG. 10 is a flowchart of an embodiment of a computing power routing method provided by this application.
  • FIG. 11 is a schematic diagram of the functional structure of a computing power router provided by this application.
  • FIG. 12 is a schematic diagram of the functional structure of a leaf router provided by this application.
  • FIG. 13 is a schematic structural diagram of an embodiment of a computing power routing device provided by this application.
  • FIG. 14 is a schematic structural diagram of an embodiment of a computing power routing device provided by this application.
  • 15 is a schematic structural diagram of an embodiment of a computing power routing device provided by this application.
  • FIG. 16 is a schematic structural diagram of an embodiment of a computing power routing device provided by this application.
  • FIG. 17 is a schematic structural diagram of an embodiment of a computing power routing device provided by this application.
  • FIG. 18 is a schematic diagram of the structure of a device provided by this application.
  • words such as “exemplary” or “for example” are used to represent examples, illustrations or illustrations, and any embodiment or solution described as “exemplary” or “for example” in the embodiments of this application It should not be construed as being more preferable or advantageous than other embodiments or solutions. To be precise, words such as “exemplary” or “for example” are used to present related concepts in a specific manner.
  • “at least one” refers to one or more, and “multiple” refers to two or more.
  • “And/or” describes the association relationship of the associated objects, indicating that there can be three relationships, for example, A and/or B, which can mean: A alone exists, A and B exist at the same time, and B exists alone, where A, B can be singular or plural.
  • the character “/” generally indicates that the associated objects before and after are in an "or” relationship.
  • Function is a software-based computing program, functional program, business software or microservice deployed in a computing site.
  • the computing container can be a computing site, server, virtual machine or container, or a smaller unit that provides computing resources.
  • the computing site can be a customer-premises equipment (CPE) on the user side.
  • CPE customer-premises equipment
  • Site such as the site where the corporate network is located (the corporate network has gateway equipment (Gateway, GW) and servers, such as network management equipment and servers can be computing sites) or the site where the home network is located.
  • GW gateway equipment
  • servers such as network management equipment and servers can be computing sites
  • Computing power coding is a technical abstraction of the computing power that computing containers can provide. It is a computing metric that spreads and propagates between routers. Computing power coding includes computing delay evaluation values and/or calculations. Capacity level, machine type or weighted value ⁇ 2 related to machine type and server load.
  • the computing power token is a commercial abstraction of the computing power that the computing container can provide. It is equivalent to the computing power that the operator computing container can provide.
  • the computing power token is in the form of computing power coding technology. Pass in the network.
  • this application provides a computing power routing method and device, which are deployed in computing containers.
  • the leaf router inside sends the computing power routing information of the computing container to the computing power router connected to the leaf router, and the computing power router advertises it to other computing power routers on the network, so that other computing power routers can obtain the computing power of all computing containers on the network. Power routing information.
  • the computing power router determines the routing information base (RIB) according to the computing power routing information. After the computing power router receives the message, it uses the forwarding information base (FIB) and the information carried in the message.
  • RIB routing information base
  • FIB forwarding information base
  • the FID routes the message to the computing container where the FID is located. In this way, packets related to computing tasks can be routed to the corresponding computing container.
  • the routing process is equivalent to computing power bidding.
  • the traffic users bid for computing power, and the traffic owner releases computing power, which dynamically and quickly matches the traffic of operators and Internet companies.
  • the relationship between supply and demand has realized low-cost sharing of traffic.
  • ICT Information and Communications Technology
  • FIG 1 is a schematic diagram of a computing power network architecture provided by this application.
  • the network consists of a computing container and a computing power network (usually a carrier network, Carrier Network), which can be divided into two layers: Underlay and Overlay.
  • the leaf router/(also called leaf node) is the external extension of the computing power network, and it is the antenna that perceives the external computing power.
  • the leaf router is deployed in the computing container.
  • the computing container is taken as the computing site (Site) as an example for illustration.
  • the computing site can be a CPE site on the user side, such as a site where an enterprise network is located, or a site where a home network is located.
  • the home network has devices such as home gateways, computers, and mobile phones, and these devices in the CPE site are underlay devices.
  • the computing site can also be a site on the network side: such as a site where a router, PoP, CO, or base station is located. There are also network devices or gateway devices and servers in the site, and these devices in the site belong to Underlay devices.
  • the sites on the network side are usually sites that have been transformed into small data centers or are equipped with computing resources such as servers.
  • the computing site can be deployed, usually on the Underlay equipment, deploy software-based computing programs, functional programs, business software or microservices, etc., and use Function to represent the deployment of software-based computing programs, functional programs, business software or microservices , Namely FX1...FXk shown in Figure 1, or FY1...FYm, the leaf routers are deployed in the computing site; these software or devices belong to the overlay layer, which means the layer on top of the Underlay.
  • the computing sites are interconnected by a computing power network, which is composed of computing power routers (also called computing power nodes), and can also include ordinary routers.
  • Hashrate routers are used to support hashrate routing, and ordinary routers do not have hashrate routing capabilities.
  • Ordinary routers are underlay devices, and the software or hardware of the computing power routing function belongs to the superimposed layer, that is, the computing power router in the figure.
  • the node with the software or hardware superimposed with the computing power routing function is the computing power node, and the computing power routing function is superimposed.
  • a software or hardware router is a computing power router.
  • the overlay and the underlay device can be integrated together, or they can be separated into separate devices, such as overlay nodes/devices.
  • computing power refers to computing resources.
  • FIG 2 is a schematic diagram of another computing power network architecture provided by this application. As shown in Figure 2, it is different from Figure 1 in that the computing power router is moved from the intermediate network to the computing site, so the computing power network is an overlay.
  • the overlay can be integrated with the Underlay device in the computing site, or it can be separated into separate devices, such as Overlay computing nodes/devices.
  • Figure 3 is a schematic diagram of another computing power network architecture provided by this application. As shown in Figure 3, it is a hybrid architecture of Figures 1 and 2.
  • the computing power router can be located in the middle network or at the computing site. Inside.
  • the computing power network implements three main functions:
  • the leaf router advertises the computing resources and/or available functions of the site and the IP address prefix/address of the computing container where the functions are located to the computing power network, so that the computing power network can discover the functions and/or computing power.
  • computing resources Expressed in the form of "computing power code”.
  • the computing power routing information may include the site computing resources and/or the functions that can be provided and the IP address prefix/address of the computing container where the function is located.
  • This notification can be implemented by extending the DC internal protocol, or by extending a routing protocol (such as Interior Gateway Protocol (IGP) IGP/Border Gateway Protocol (BGP).
  • IGP Interior Gateway Protocol
  • BGP Border Gateway Protocol
  • the hashrate routing information can also include the IP address prefix/address of the leaf router or the IP address prefix/address of the hashrate router connected to the leaf router.
  • Each computing power router of the computing power network can discover the function and/or computing power. This kind of notification can also be implemented by extending the DC internal protocol, or by extending the routing protocol (such as IGP/BGP).
  • each computing power router can communicate directly, forming a logical full mesh (full mesh) peer (peer) relationship.
  • the computing power routers (nodes) can be logically interconnected by tunneling technologies such as Multiprotocol Label Switching (MPLS) and IP, or by segment routing (SR) technology.
  • MPLS Multiprotocol Label Switching
  • SR segment routing
  • FIG. 4 is a schematic diagram of an application scenario of the computing power routing method and device provided by this application.
  • the computing container takes a computing site as an example.
  • the method of this embodiment includes the following processes:
  • the Internet company and the operator agree on the computing power rent through a smart contract.
  • the Internet company transfers the corresponding amount of "computing power tokens" to the computing power contract to be rented according to the computing power that the operator can calculate on the site.
  • the leaf router deployed in the computing site sends the "computing power code” or the computing metric value calculated from the “computing power code” to the computing power network through the routing protocol (IGP/BGP), and sends it first To the entrance router (for the hashrate router).
  • IGP/BGP routing protocol
  • the entrance router in the computing power network recognizes the "two-dimensional code (ie computing power code or calculated metric value)", and verifies the authenticity of the signature to the contract and verifies the legality of the computing power code or calculated metric value signature with the public key ,
  • the so-called computing power routing lock unlocking is to allow computing power coding or calculation metric values to spread in the computing power network
  • the computing power code or the calculated metric value is a computing metric value that is propagated in the computing power network through the router (the computing power router).
  • the routing process is equivalent to "computing power bidding". Based on the computing power provided by the computing site represented by the computing power code or the calculated metric value, the router selects the most suitable computing container through comprehensive consideration of reasonable algorithms. Get the best computing power routing, so that the traffic is routed to the "highest bid" Internet company leased site for calculation.
  • the computing power owner (such as an operator) obtains the agreed corresponding "computing power token" from the smart contract.
  • computing power tokens are equivalent to the computing power provided by operators’ computing sites, that is, “computing power tokens” are a commercial abstraction of the computing power that computing sites can provide, while “computing power tokens” are “Power coding” is a technical abstraction of the computing power that a computing site can provide. "Hashing power tokens” are transmitted in the computing power network in the form of "Hashing power coding” technology.
  • the leaf router/leaf node can be set up for each Internet or for each type of business, usually set in a data center or computing container, and can exist in the form of hardware equipment or software, such as a software-based virtual router.
  • Smart Contract Smart Contract
  • Smart Contract can be maintained by a platform, it can be a decentralized blockchain platform, or a centralized agency platform, used to realize computing power users (such as Internet companies or third parties) and computing power Payment or transfer of hashrate tokens between owners (such as operators or enterprises).
  • the network elements involved in this application include computing power routers and leaf routers.
  • FIG. 5 is a flowchart of an embodiment of a computing power routing method provided by this application.
  • the execution subject of this embodiment may be a computing power router or a computing power node.
  • the method of this embodiment can include:
  • the computing power router receives computing power routing information of at least one computing container, where the computing power routing information includes a computing power code or a calculated metric value.
  • the computing power router receives computing power routing information of at least one computing container from at least one leaf router connected to it, the leaf router is deployed in the computing container, and at least one leaf router is deployed in one computing container.
  • the method in this embodiment may further include:
  • the computing power router sends the computing power routing information to other computing power routers in the network.
  • the computing power routing information includes the computing power code or the calculated metric value. If the computing power code is the computing power code signed by the private key, before S104, it can also include: after the computing power router verifies that the signature is legal, unlock the computing power routing lock, And trigger the calculation power transaction, get the validity period of the calculation power code, that is, the valid time that can be used, for example, the signature can be verified to the contract platform or the relevant department as legal. When the validity period of the hashrate code expires, the hashrate routing lock will re-lock the hashrate code, that is, the corresponding hashrate code or the calculated metric value calculated by it will not be allowed to spread in the network.
  • the computing power router can send computing power routing information to other computing power routers in the network through IGP/BGP.
  • the computing power router determines the RIB according to the computing power routing information.
  • the RIB includes the function ID (FID) of the function that the computing container can provide or one of the IP address prefix of the computing container where the function is located, and the computing power code or the calculated metric value.
  • FID function ID
  • One of the functions includes any one of software-based computing programs, functional programs, business software, or microservices deployed in a computing container.
  • the RIB includes the FID and computing power code of the function that the computing container can provide, or the RIB includes the FID and the calculated metric value of the function that the computing container can provide, or the RIB includes the IP address prefix and computing power of the computing container where the function is located.
  • Encoding, or RIB includes the IP address prefix of the computing container where the function is located and the calculated metric value.
  • the computing power router receives the message, forwards the message according to the FIB, the FIB determines according to the RIB, and the FIB includes at least the FID or the IP address prefix.
  • the computing power routing information includes computing power codes or calculated metric values
  • the computing power router determines the RIB according to the computing power routing information, which can be specifically: the computing power router obtains the “computing power code” according to the “computing power code” or the calculated metric value. Either calculate the mapping relationship between the metric value and the output port, or obtain the FID, the "hash code” or calculate the mapping relationship between the metric value and the output port according to the "hash code” or the calculated metric value, or according to the "hash code” Or calculate the metric value to get the IP address prefix, "computing power code” or calculate the mapping relationship between the metric value and the outgoing port, etc.
  • RIB includes FID, calculated metric value, IP address prefix, and outbound port; or,
  • IP address prefix calculated metric value and outgoing port
  • IP address prefix computing power code and outgoing port
  • FIB includes: FID, IP address prefix and outgoing port, or FID, hashrate router IP address and outgoing port, or FID and outgoing port.
  • the method of this embodiment may further include:
  • the computing power router determines the FIB item corresponding to the message according to the computing power code or calculated metric value in the RIB. Specifically, it can be to determine the FIB item corresponding to the FID or IP address prefix carried in the message. For example, the computing power that the computing power router can provide based on the computing site represented by the computing power code or the computing metric value is comprehensively considered through a reasonable algorithm.
  • the RIB includes at least one of the FID of the function that the computing container can provide or the IP address prefix of the computing container where the function is located, and one of the computing power code or the computing metric value. If the RIB includes the computing power code, the computing power router compares The size of all the hashrate codes in RIB. If there are a total of k hashrate codes, compare the size of hashrate codes 1, hashrate codes 2,... hashrate codes k to calculate the route, and get the destination IP address and/or output Port to obtain the FIB item, the FIB item includes the FID, the destination IP address and the outgoing port.
  • the computing power router compares the size of all the calculated metric values in the RIB. If there are a total of k calculated metric values, compare the calculated metric value 1, the calculated metric value 2, ... the size of the calculated metric value k, To calculate the route, obtain the destination IP address and/or the outgoing port, thereby obtaining the FIB item, the FIB item includes the FID, the destination IP address and/or the outgoing port.
  • Method 1 The computing power router searches the FIB for the IP address prefix and the outgoing port corresponding to the FID.
  • the computing power router adds the FID to the IP address prefix corresponding to the FID to obtain the destination IP address, and forwards the message according to the destination IP address.
  • Method 2 The computing power router finds the IP address and outgoing port of the computing power router corresponding to the FID from the FIB.
  • the computing power router uses the IP address of the computing power router as the tunnel destination address to add tunnel encapsulation to the message, or uses the IP address of the computing power router as the segment identifier of the segment routing SR to add SR encapsulation.
  • the computing power router will add tunnel encapsulation or add SR encapsulation for packet forwarding.
  • Method 3 The hashrate router finds the outgoing port corresponding to the FID or IP address prefix from the FIB, and the hashrate router forwards the packet from the outgoing port.
  • the message is forwarded according to the routing information database FIB, which can be: the computing power router forwards the message according to the IP address.
  • the routing information database FIB which can be: the computing power router forwards the message according to the IP address.
  • the computing power encoding includes computing delay evaluation values and/or computing capacity levels, machine types or weighted values ⁇ 2 related to machine types, and server load, that is, computing power encoding includes computing delay evaluation values, or computing Power coding includes computing capacity level, machine type or weighted value ⁇ 2 and server load related to machine type, or computing power coding includes computing latency evaluation value and computing capacity level, machine type or weighted value ⁇ 2 and server load related to machine type
  • the method of this embodiment may further include:
  • the computing power router calculates the metric value according to the computing power code and the following formula:
  • ⁇ 2 and ⁇ 2 are weighted values.
  • ⁇ 2, ⁇ 2, and ⁇ 2 can be given according to the contract or service level agreement (Service Level Agreement, SLA) between the computing power owner and the user.
  • SLA Service Level Agreement
  • the calculated metric value is calculated by the leaf router through the above calculation formula.
  • the above algorithm is only an embodiment, and other methods can also be adopted. The algorithm is obtained.
  • the method in this embodiment may further include:
  • the computing power router If the computing power router receives the response message, it will determine the determined FIB project as a valid FIB project;
  • the computing power router If the computing power router does not receive the response message, it deletes the determined FIB item.
  • the beneficial effect of this processing is: for session-based services, there is usually a session request and session response message message interaction during the session establishment phase. When the response message is not received, the session establishment is unsuccessful. The deletion of the FIB item determined when the session request message is received is equivalent to leaving garbage items that will no longer be used, resulting in unnecessary waste of storage space.
  • the computing power router After receiving the message in S104, the computing power router forwards the message according to the FIB and routes it to the corresponding computing container.
  • the computing power routing information of at least one computing container is received through the computing power router.
  • the computing power routing information includes the computing power code or the calculated metric value, and then the computing power routing information is sent to other networks in the network.
  • the computing power router determines the RIB based on the computing power routing information.
  • the computing power router receives the message, it forwards the message according to the FIB and finally routes it to the corresponding computing container.
  • the FIB is determined based on the RIB, and the computing power routing information represents the calculation.
  • the routing process is equivalent to computing power bidding and traffic usage
  • the computing power is obtained by bidding, and the traffic owner releases computing power, which dynamically and quickly matches the traffic supply and demand relationship between operators and Internet companies, and realizes low-cost sharing of traffic.
  • FIG. 6 is a flowchart of an embodiment of a computing power routing method provided by this application.
  • the execution subject of this embodiment may be a leaf router or a leaf node.
  • the method of this embodiment may include:
  • the leaf router obtains the computing power code of the computing container, and the leaf router is deployed in the computing container.
  • the leaf router obtains the computing power code of the computing container, which may specifically be:
  • the leaf router obtains the computing power code from the calculation container and/or measurement tool.
  • the computing power code includes the calculation delay evaluation value and/or the calculation capacity level, the machine type or the weighted value ⁇ 2 related to the machine type and the server load, and the calculation delay evaluation
  • the value is measured by measurement tools such as Openstack, K8S/Kubernetes, or IT software such as operating systems.
  • the computing capacity level for example, the computing capacity level corresponding to 1 server is 255, the computing capacity level corresponding to 2 servers is 127, the computing capacity level corresponding to 3-10 servers is 63, and the computing capacity corresponding to 10-1000 servers is The level is 15, the computing capacity level corresponding to 1000-10000 servers is 7, and the computing capacity level corresponding to more than 10,000 servers is 3.
  • the leaf router sends the computing power routing information of the computing container to the computing power router connected to the leaf router, so that the computing power router sends the computing power routing information to other computing power routers on the network, and the computing power routing information includes the computing power code. Or calculate the metric value.
  • the computing power routing information may also include:
  • the FID of the function that the computing container can provide the IP address prefix or IP address of the computing container where the function is located, and the function is a software-based computing program, functional program, business software, or microservice deployed in the computing container.
  • RT Route-Target
  • RD Route-Distinguisher
  • the method of this embodiment may further include:
  • the leaf router calculates the metric value according to the calculation power code and the following formula:
  • ⁇ 2, ⁇ 2 are weighted values, and ⁇ 2, ⁇ 2, and ⁇ 2 can be given according to the contract or service level agreement (SLA) between the computing power owner and the user.
  • SLA service level agreement
  • the leaf router calculates the hashrate tokens according to the hashrate code and the following formula:
  • Hash power token ⁇ 1 ⁇ [ ⁇ 1-(calculated metric value)]
  • ⁇ 1 and ⁇ 1 are computing power pricing coefficients, which can be determined by the computing power owner (such as an operator), or jointly determined by the computing power owner and the computing power user, or determined by the platform where the contract is located.
  • the leaf router receives the message and forwards the message.
  • the method of this embodiment may further include: the leaf router triggers the contract platform to transfer the computing power token to the computing power user.
  • the method in this embodiment may further include:
  • the leaf router receives the response message sent by the computing container, and the response message carries the IP address of the computing container;
  • the leaf router sends the response message to the client through the computing power router connected to the leaf router.
  • the response message is used for the client to change the FID carried in the message to the IP address of the computing container when the client performs subsequent communication, thereby solving the problem of stickiness. (Flow Affinity) issues.
  • the computing power code of the computing container is obtained through the leaf router, and the computing power routing information including the computing power code or the calculated metric value is sent to the computing power router connected to the leaf router, so that the computing power is
  • the router sends the computing power routing information to other computing power routers on the network, and finally the leaf router receives the message, forwards the message, and specifically forwards the message to the computing container.
  • the computing power routing information represents the computing power that the computing container can provide, so that the packets related to computing tasks can be routed to the corresponding computing container to achieve the best user experience, computing resource utilization, network efficiency, and routing process equivalent
  • the users of the traffic obtain the computing power in the bidding, and the traffic owner releases the computing power, which dynamically and quickly matches the traffic supply and demand relationship between the operators and the Internet companies, and realizes the low-cost sharing of traffic.
  • FIG. 5 and FIG. 6 will be described in detail through four specific embodiments.
  • FIG. 7 is a flowchart of an embodiment of a computing power routing method provided by this application. As shown in FIG. 5, the method in this embodiment may include:
  • the leaf router obtains the computing power code of the computing container, and the leaf router is deployed in the computing container.
  • leaf routers there are k leaf routers.
  • the leaf router 1 obtains the private key signed computing power code 1 from the computing container and/or IT measurement tool
  • the leaf router 2 obtains the private key signed from the computing container and/or IT measurement tool.
  • the computing power code 2 etc.
  • the leaf router k obtains the computing power code k signed by the private key from the computing container and/or IT measurement tool.
  • the leaf router obtains the calculation capacity level, the machine type or the weighted value ⁇ 2 related to the machine type and the server load from the calculation container, obtains the calculation delay evaluation value from the IT measurement tool, and the IT measurement tool measures the calculation delay evaluation value.
  • IT measurement tools such as Openstack, K8S/Kubernetes or operating system and other IT software
  • computing power coding includes calculation delay evaluation value and/or calculation capacity level, machine type or the weighted value ⁇ 2 related to the machine type and server load.
  • the leaf router calculates the metric value according to the calculation power code and the following formula:
  • ⁇ 2 and ⁇ 2 are weighted values, and ⁇ 2, ⁇ 2 and ⁇ 2 can be given according to the contract/service level agreement (SLA) between the computing power owner and the user.
  • SLA contract/service level agreement
  • the leaf router calculates the hashrate tokens according to the hashrate code and the following formula:
  • Hash power token ⁇ 1 ⁇ [ ⁇ 1-(calculated metric value)]
  • ⁇ 1 and ⁇ 1 are computing power pricing coefficients, which can be determined by the computing power owner (such as an operator), or jointly determined by the computing power owner and the computing power user, or determined by the platform where the contract is located.
  • the leaf router 1, and the leaf router k send the computing power routing information to a first computing power router connected to it through IGP/BGP.
  • the computing power routing information includes the FID, the IP address prefix of the computing container where the leaf router is located, and the computing power code k. If the leaf router in S101 calculates the computing metric value according to the computing power code, the computing power routing information includes FID, where the leaf router is located Calculate the IP address prefix of the container and calculate the metric value k. If the computing power routing information does not include the calculated metric value, the computing power router that receives the computing power routing information calculates the calculated metric value according to the computing power code.
  • the first computing power router verifies that the signature is legal to the contract platform or related departments, unlocks the "power routing lock”, triggers the "power transaction”, and obtains the validity period of the computing power code k.
  • the validity period of the hashrate code k is the effective use time of the hashrate code k.
  • the first computing power router sends the computing power routing information to other computing power routers on the network through IGP/BGP.
  • all computing power routers in the network obtain the IP address prefix and FID corresponding to each computing power code k and computing power code k, and the validity period of the computing power code k, and generate RIBs based on these information.
  • RIB includes computing power code k, Correspondence between IP address prefix, FID, and outbound port.
  • the first computing power router obtains the RIB item including the FID, the computing power code k, the IP address prefix, and the outgoing port.
  • the second computing power router obtains the RIB item including FID, computing power code k, IP address prefix and outgoing port.
  • the client sends the first packet carrying the FID.
  • FIDx is carried in the first packet.
  • the second computing power router that received the first packet checks the FIB according to the FID, and if there is no item corresponding to the FID in the FIB, it checks the RIB according to the FID, and compares the size of the "computing power code 1...k" to calculate By routing, the destination IP address prefix and the outgoing port corresponding to the FID are obtained, and the FIB item forwarded based on the FID is generated.
  • the FIB item includes the corresponding relationship between the FID, the IP address prefix and the outgoing port.
  • the second computing power router adds the destination IP address prefix corresponding to the FID obtained by the FID to obtain the destination IP address, and forwards the first packet to the first computing power router according to the destination IP address.
  • the packet message carries the destination IP address.
  • the destination IP address is the IP address of the computing container where the FID is located.
  • the first computing power router forwards the first packet carrying the destination IP address to the leaf router K according to the destination IP address.
  • the leaf router K here is a leaf router connected to the first hashrate router.
  • the leaf router K forwards the first packet to the computing container where the FID is located according to the destination IP address, and triggers the transfer of "computing power tokens".
  • the client sends a follow-up message carrying FIDX to the second computing power router.
  • the second computing power router searches the FIB according to the FID carried in the message, obtains the destination IP address prefix corresponding to the FID, adds the FID to the destination IP address prefix to obtain the destination IP address, and bases the message according to the destination IP
  • the address is forwarded to the first computing power router, and the first packet forwarded carries the destination IP address.
  • the first computing power router forwards the packet carrying the destination IP address to the leaf router according to the destination IP address.
  • the leaf router forwards the message to the computing container where the FID is located according to the destination IP address.
  • the conditions for triggering the transfer of hashrate tokens or hashrate transactions can be, for example, when the hashrate code is successfully injected into the hashrate network, or the hashrate network generates RIB/FIB, or if the hashrate network successfully routes the message to the Internet company/ The leaf router or computing site to which the third party belongs.
  • the message carries the FID, and the FID is not an IP address. It needs to be carried by a specific domain, and it can be placed in the destination address domain of the IP message.
  • Only the entrance of the computing power network (such as the entrance computing power router, the second computing power router in Figure 7) requires routing based on the calculated metric value and forwarding based on FID. The rest of the routers can continue to use traditional IP forwarding, so it can protect the current There is investment in IP routers.
  • FIG 8 is a flowchart of an embodiment of a computing power routing method provided by this application.
  • the computing power network such as the entrance computing power router, the second computing power router in Figure 8
  • the computing power network outlet such as the export computing power router, the first computing power router in Figure 8
  • the method of this embodiment may include:
  • S401-S403 are the same as S301-S303, please refer to the description of S301-S303, which will not be repeated here.
  • the first computing power router sends the computing power routing information to other computing power routers on the network through IGP/BGP.
  • the computing power routing information sent by the first computing power router to other computing power routers in the network through IGP/BGP includes computing power code k and the first computing power router IP address.
  • the first computing power router obtains the RIB item including the FID, the computing power code k, the IP address prefix, and the outgoing port.
  • the second computing power router obtains the RIB item including the FID, the computing power code k, and the IP address and outgoing port of the first computing power router.
  • S406 The client sends the first packet carrying the FID.
  • FIDX is carried in the first packet.
  • the second computing power router that received the first packet checks the FIB according to the FID, and if the FIB does not have an item corresponding to the FID, it checks the RIB according to the FID, and compares the size of the "computing power code 1...k" to calculate Route, obtain the IP address and output port of the best computing power router (such as the first computing power router) corresponding to the FID, and generate the FIB project forwarded based on the FID.
  • the FIB project includes the FID, the IP address of the first computing power router and the output port. Correspondence of ports.
  • the second computing power router uses the IP address of the computing power router as the tunnel destination address to add tunnel encapsulation to the packet, or uses the IP address of the computing power router as the segment identifier of the segment route (SR) to add SR. Encapsulation, forwarding the packet with tunnel encapsulation or SR encapsulation added to the first computing power router.
  • SR segment route
  • S409 The first computing power router forwards the first packet sent by the second computing power router to the leaf router K.
  • the leaf router K here is a leaf router connected to the first hashrate router.
  • the leaf router K forwards the first packet to the computing container where FIDX is located, and triggers the "computing power token" transfer.
  • S411 The client sends a follow-up message carrying FIDX to the second computing power router.
  • the second computing power router searches the FIB according to the FID carried in the message to obtain the best computing power router IP address and outgoing port corresponding to the FID, and uses the computing power router IP address as the tunnel destination address for the message Add tunnel encapsulation, or add SR encapsulation using the IP address of the computing power router as the segment identifier of the segment routing (SR), and forward the packet of adding tunnel encapsulation or adding SR encapsulation to the first computing power router.
  • SR segment identifier of the segment routing
  • S413 The first computing power router forwards the message sent by the second computing power router to the leaf router K.
  • S414 The leaf router forwards the message to the computing container where the FID is located.
  • FIG. 9 is a flowchart of an embodiment of a computing power routing method provided by this application. As shown in FIG. 9, each computing power router in the computing power network is not limited to the first and second computing power routers in FIG. Both require routing based on calculated metric values and forwarding based on FID.
  • the method of this embodiment may include:
  • S501 is the same as S301, and will not be repeated here.
  • leaf router 1 leaf router k sends the computing power routing information to a first computing power router connected to it through IGP/BGP.
  • the computing power routing information includes the FID and the computing power code k. If the leaf router in S501 calculates the calculated metric value according to the computing power code, the computing power routing information includes the FID and the computing metric value k. If the computing power routing information does not include the calculated metric value, the computing power router that receives the computing power routing information calculates the calculated metric value according to the computing power code.
  • the first computing power router verifies that the signature is legal to the contract platform or related departments, unlocks the "power routing lock”, triggers the "power transaction”, and obtains the validity period of the computing power code k.
  • the validity period of the hashrate code k is the effective use time of the hashrate code k.
  • the first computing power router sends the computing power routing information including the computing power code k to other computing power routers on the network through IGP/BGP.
  • the first computing power router obtains the RIB item including the FID, the computing power code k, and the outgoing port.
  • the second computing power router obtains the RIB item including FID, computing power code k, and outgoing port.
  • S506 The client sends the first packet carrying the FID.
  • FIDx is carried in the first packet.
  • the second computing power router that received the first packet checks the FIB according to the FID. If the FIB does not have an item corresponding to the FID, it checks the RIB according to the FID, and compares the size of the "computing power code 1...k" to calculate Route, obtain the outgoing port corresponding to the FID, and generate the FIB item forwarded based on the FID.
  • the FIB item includes the corresponding relationship between the FID and the outgoing port.
  • the second computing power router forwards the first packet according to the outgoing port, and forwards it through several similar computing power routers, and then reaches the first computing power router (the exit of the computing power network).
  • the first computing power router checks the FIB according to the FID. If the FIB does not have an item corresponding to the FID, it checks the RIB according to the FID and compares the size of the "computing power code 1...k" to calculate the route and obtain the corresponding FID
  • the FIB project forwarded based on the FID is generated, and the FIB project includes the corresponding relationship between the FID and the exit port.
  • S510 The first computing power router forwards the first packet to the leaf router K according to the outgoing port.
  • the leaf router K here is a leaf router connected to the first hashrate router.
  • the leaf router K forwards the first packet to the computing container where the FID is located, and triggers the transfer of "computing power tokens".
  • the client sends a follow-up message carrying FIDX to the second computing power router.
  • the second computing power router checks the FIB according to the FID carried in the message, obtains the outgoing port corresponding to the FID, and forwards the message to the first computing power router.
  • the first computing power router checks the FIB according to the FID carried in the message, obtains the outgoing port corresponding to the FID, and forwards the message to the leaf router K.
  • the leaf router K here is a leaf router connected to the first hashrate router.
  • the leaf router K forwards the message to the computing container where the FID is located.
  • FIG. 10 is a flowchart of an embodiment of a computing power routing method provided by this application. As shown in FIG. 10, this embodiment is based on the method shown in FIG. 7 when the first packet is a Transmission Control Protocol (Transmission Control Protocol).
  • TCP Transmission Control Protocol
  • SYN TCP synchronization
  • the method of this embodiment may include:
  • S601 is the same as S301, and will not be repeated here.
  • leaf router 1 leaf router k sends the computing power routing information to a first computing power router connected to it through IGP/BGP.
  • the computing power routing information includes the FID, the IP address prefix of the computing container where the leaf router is located, and the computing power code k. If the leaf router in S101 calculates the computing metric value according to the computing power code, the computing power routing information includes FID, where the leaf router is located Calculate the IP address prefix of the container and calculate the metric value k. If the computing power routing information does not include the calculated metric value, the computing power router that receives the computing power routing information calculates the calculated metric value according to the computing power code.
  • the first computing power router verifies that the signature is legal to the contract platform or related departments, and unlocks the "power routing lock” and triggers the "power transaction” to obtain the validity period of the computing power code k.
  • the validity period of the hashrate code k is the effective use time of the hashrate code k.
  • the first computing power router sends the computing power routing information to other computing power routers on the network through IGP/BGP.
  • all computing power routers in the network obtain the IP address prefix and FID corresponding to each computing power code k and computing power code k, and the validity period of the computing power code k, and generate RIBs based on these information.
  • RIB includes computing power code k, Correspondence between IP address prefix, FID, and outgoing port.
  • the first computing power router obtains the RIB item including the FID, the computing power code k, the IP address prefix, and the outgoing port.
  • the second computing power router obtains the RIB item including FID, computing power code k, IP address prefix and outgoing port.
  • the client sends the first packet of the TCP SYN type, and the first packet carries FIDx.
  • the second computing power router that received the first packet checks the FIB according to the FID. If the FIB does not have an item corresponding to the FID, it checks the RIB according to the FID, and compares the size of the "computing power code 1...k" to calculate By routing, the destination IP address prefix and the outgoing port corresponding to the FID are obtained, and the FIB item forwarded based on the FID is generated.
  • the FIB item includes the corresponding relationship between the FID, the IP address prefix and the outgoing port.
  • the second computing power router adds the destination IP address prefix corresponding to the FID obtained by the FID to obtain the destination IP address, and forwards the TCP SYN message to the first computing power router according to the destination IP address, and the forwarded TCP
  • the SYN message carries the destination IP address.
  • the destination IP address is the IP address of the computing container where the FID is located.
  • the first computing power router forwards the TCP SYN message carrying the destination IP address to the leaf router K according to the destination IP address.
  • the leaf router K here is a leaf router connected to the first hashrate router.
  • the leaf router K forwards the TCP SYN message to the computing container where the FID is located according to the destination IP address.
  • the computing container where the FIDX is located sends a TCP SYN response message to the leaf router K, and the response message carries the IP address of the computing container as the source address.
  • the leaf router K triggers the transfer of "computing power tokens".
  • TCP SYN ACK message is received to trigger the transfer of hashrate tokens.
  • the purpose is to ensure that the transfer of hashrate tokens is triggered only when the TCP connection or other session connections are correctly established.
  • the leaf router K sends a TCP SYN response message to the first computing power router, and the response message carries the source address of the IP address of the computing container.
  • the first computing power router sends a TCP SYN response message to the second computing power router, and the response message carries the source address of the IP address of the computing container.
  • the second computing power router After receiving the TCP SYN response message, the second computing power router determines the FIB item generated in S607 as a valid FIB item. If the TCP SYN response message is not received, the FIB item generated in S607 is deleted, indicating that the TCP session is not successfully established. If the TCP session or user session connection is unsuccessful (usually indicating that the computing site or related computing path is faulty), subsequent packets cannot receive the computing power routing service.
  • the second computing power router sends a TCP SYN response message to the client, and the response message carries the IP address of the computing container whose source address is the source address.
  • S617 The client changes the FID carried in the message to the IP address of the computing container when performing subsequent communication.
  • the client sends a follow-up message to the second computing power router, where the follow-up message carries the IP address of the computing container.
  • FIG. 11 is a schematic diagram of the functional structure of a computing power router provided by this application.
  • the computing power router includes a control plane processing module 11 and a data plane processing module 12, and the control plane processing module 11 includes a routing protocol processing unit 110 , RIB generation and maintenance unit 111, routing or calculation unit 112, FIB generation and maintenance unit 113, signature verification unit 114, and computing power transaction trigger unit 115, data plane processing module 12 includes forwarding processing unit 121 and multiple interface processing units .
  • the routing protocol processing unit 110 is used for function and/or computing power notification, specifically for computing the computing resources of the computing container, and/or the functions that can be provided, and the IP address prefix/address of the computing container where the function is located, to other computing power routers Announcement is made so that other computing power routers in the computing power network can discover the function and/or computing power.
  • the IP address prefix is advertised.
  • the computing resource is expressed in the form of "computing power code" in this embodiment of the application.
  • the computing power routing information may include the computing resources and/or functions that can be provided by the computing container and the IP address prefix/address of the computing container where the function is located.
  • the computing power routing information may also include the leaf router IP address prefix/address Or the IP address prefix/address of the hashrate router connected to the leaf router.
  • This notification can be achieved by extending the internal protocol of the DC or by extending the routing protocol (such as IGP/BGP).
  • the signature verification unit 114 is used to interactively verify the authenticity of the signature of the "computing power code” with the contract platform, and optionally obtain the validity period of the hashrate lease. If the authenticity of the signature is verified, the hashrate routing "lock” is unlocked, allowing the "computing power” The power code is passed to other computing power routers.
  • the computing power transaction trigger unit 115 is used to trigger “power transactions”, such as triggering the transfer of “power tokens”, for example, triggering the contract platform to transfer “power tokens” to computing power users, because computing power routing is used for traffic guidance The service has been started.
  • the RIB generation and maintenance unit 111 is used to generate RIB entries according to the computing power routing information.
  • the RIB entries include F ID, corresponding calculated metric values, corresponding IP address prefixes and outbound ports; or, RIB entries include FID and corresponding Calculate the metric value, the corresponding export hashrate router IP address or leaf router IP address and outgoing port.
  • the route selection or calculation unit 112 is used to select or calculate a route according to the "calculated metric value" of the RIB to obtain the destination IP address prefix and the outgoing port corresponding to the FID; or to obtain the best egress computing power router IP address corresponding to the FID Or the validity period of the leaf router's IP address, outgoing port, and computing power code, that is, to achieve load balancing between sites.
  • the FIB generation and maintenance unit 113 is used to form a corresponding FIB entry (entry) according to the processing result of the routing or calculation unit 112.
  • the FIB entry includes the FID, the corresponding IP address prefix and the outgoing port; or, the FIB entry includes the FID and the corresponding The IP address of the export hashrate router or the IP address and outgoing port of the leaf router.
  • the forwarding processing unit 121 is used to directly perform IP forwarding when the message carries a valid IP destination address; or, to check the FIB according to the FID carried in the message to obtain the outgoing port, and forward it from the outgoing port; or The FID carried in the message checks the FIB to obtain the IP address prefix, adds the FID to the IP address prefix to obtain a destination IP address, and then forwards it according to the destination IP address.
  • the interface processing unit is used for interface input or output processing.
  • FIG 12 is a schematic diagram of the functional structure of a leaf router provided by this application.
  • the leaf router can exist in the form of software (such as a virtual router) or a hardware device.
  • the leaf router includes a control plane processing module 21 and a data plane processing module.
  • Module 22 The control plane processing module 21 includes a routing protocol processing unit 210, a computing power acquisition unit 214, and a computing power transaction trigger unit 215.
  • the data plane processing module 22 includes a forwarding processing unit 221, multiple interface processing units, and multiple functional processing units .
  • the routing protocol processing unit 210 is used for function and/or computing power notification, specifically by calculating the site computing resources, and/or available Functions and Function to calculate the IP address prefix/address of the container to advertise other computing power routers, Enable other computing power routers in the computing power network to discover the function and/or computing power; preferably advertise the IP address prefix, and the computing resource is expressed in the form of "computing power code" in this case.
  • site computing resources, and/or available Functions and Function calculation container IP address prefixes/addresses become hash power routing information, and hash power routing information may also include leaf router IP address prefixes/addresses.
  • This notification can be implemented by extending the DC internal protocol, or by extending the routing protocol (such as IGP/BGP).
  • the computing power acquisition unit 214 is used to acquire the "computing power code (Code)” signed by the private key and the list of possible serviceable Function IDs, that is, for function discovery and computing resource discovery; "computing power code (Code)” Usually provided by IT software, “Code” will be discussed in detail below;
  • the computing power transaction trigger unit 215 is used to trigger “power transactions”, such as triggering the transfer of “power tokens”, for example, triggering the contract platform to transfer “power tokens” to computing power users, because the traffic has been directed to the corresponding The leaf router of the site.
  • control plane processing module 21 further includes a RIB generation and maintenance unit 211, a routing or calculation unit 212, and a FIB generation and maintenance unit 213, wherein the RIB generation and maintenance unit 211 is used to generate RIB items based on computing power routing information (entry), the RIB item includes the Function ID, the corresponding calculated metric value, the corresponding IP address prefix, and the outgoing port.
  • RIB generation and maintenance unit 211 is used to generate RIB items based on computing power routing information (entry)
  • the RIB item includes the Function ID, the corresponding calculated metric value, the corresponding IP address prefix, and the outgoing port.
  • the route selection or calculation unit 212 is used to select or calculate the route according to the "calculated metric value" of the RIB to obtain the best destination IP address prefix and virtual outgoing port corresponding to the Function ID; it can be used as a site or DC internal Computing resource selection, that is, load balancing within the site or server.
  • the FIB generation and maintenance unit 213 is configured to form a corresponding FIB entry (entry) according to the processing result of the routing or calculation unit.
  • the FIB entry includes the Function ID, the corresponding IP address prefix, and the virtual out port.
  • the forwarding processing unit 221 is used to directly perform IP forwarding when the message carries a valid IP destination address; or, to check the FIB according to the Function ID carried in the message to obtain a virtual outgoing port, and forward it to the corresponding Function from the virtual outgoing port Processing unit.
  • the interface processing unit is used for interface input or output processing.
  • the function processing unit is used to perform calculation, service or program processing corresponding to the function.
  • the leaf router runs in the form of virtual router software, and it is connected to the function through a virtual port.
  • FIG. 13 is a schematic structural diagram of an embodiment of a computing power routing device provided by this application. As shown in FIG. 13, the device of this embodiment may include: a receiving module 31, a determining module 32, and a sending module 33, where:
  • the receiving module 31 is configured to receive computing power routing information of at least one calculation container, where the computing power routing information includes a computing power code or a calculated metric value;
  • the determining module 32 is configured to determine a routing information base RIB according to the computing power routing information, and the RIB includes one of the function identification FID of the function that the computing container can provide or the IP address prefix of the computing container where the function is located, and the computing power One of the force code or the calculated metric value, and the function includes any one of a softwareized calculation program, a functional program, business software, or a microservice deployed in the calculation container;
  • the receiving module 31 is also used to receive messages
  • the sending module 33 is configured to forward the message according to the routing information database FIB, the FIB is determined according to the RIB, and the FIB includes the FID or the IP address prefix.
  • sending module 33 is also used for:
  • the determining module 32 is further configured to:
  • the RIB includes FID, calculated metric value, IP address prefix and outgoing port; or,
  • IP address prefix calculated metric value and outgoing port
  • IP address prefix computing power code and outgoing port
  • the FIB includes: FID, IP address prefix, and outbound port; or,
  • the sending module is configured to:
  • the device in this embodiment can be used to implement the technical solution of the method embodiment shown in FIG. 5, and its implementation principles and technical effects are similar, and will not be repeated here.
  • FIG. 14 is a schematic structural diagram of an embodiment of a computing power routing device provided by this application.
  • the device of this embodiment is based on the device shown in FIG. 13, and computing power coding includes computing delay evaluation values and / Or calculate the capacity level, the machine type or the weighted value ⁇ 2 related to the machine type and the server load, and when the computing power routing information includes computing power coding, it may further include: a computing module 34, which is used for The calculation metric value is calculated according to the calculation power code and the following formula:
  • ⁇ 2 and ⁇ 2 are weighted values.
  • the device in this embodiment can be used to implement the technical solution of the method embodiment shown in FIG. 5, and its implementation principles and technical effects are similar, and will not be repeated here.
  • FIG. 15 is a schematic structural diagram of an embodiment of a computing power routing device provided by this application.
  • the device of this embodiment is based on the device structure shown in FIG. 13, and may further include: a processing module 35.
  • the processing module 35 is configured to determine the FIB item as a valid FIB item if the receiving module receives a response message;
  • the device in this embodiment can be used to implement the technical solution of the method embodiment shown in FIG. 5, and its implementation principles and technical effects are similar, and will not be repeated here.
  • FIG. 16 is a schematic structural diagram of an embodiment of a computing power routing device provided by this application.
  • the device of this embodiment may include: an acquisition module 41, a sending module 42, and a receiving module 43, where:
  • the obtaining module 41 is configured to obtain the computing power code of the computing container, and the leaf router is deployed in the computing container;
  • the sending module 42 is configured to send the computing power routing information of the computing container to the computing power router connected to the leaf router, so that the computing power router sends the computing power routing information to other computing power routers on the network ,
  • the computing power routing information includes the computing power code or the calculated metric value;
  • the receiving module 43 is used to receive messages
  • the sending module 42 is also used to forward the message.
  • the obtaining module 41 is used to:
  • the computing power code from the calculation container and/or measurement tool, where the computing power code includes a calculation delay evaluation value and/or a calculation capacity level, a machine type or a weighted value ⁇ 2 related to the machine type, and server load,
  • the calculation delay evaluation value is measured by the measurement tool.
  • the device in this embodiment can be used to implement the technical solution of the method embodiment shown in FIG. 6, and its implementation principles and technical effects are similar, and will not be repeated here.
  • Figure 17 is a schematic structural diagram of an embodiment of a computing power routing device provided by this application. As shown in Figure 17, the device of this embodiment is based on the device structure shown in Figure 16, if the computing power routing information includes all
  • the calculated metric value may further include: a calculation module 44, which is configured to calculate the calculated metric value according to the computing power code and the following formula:
  • ⁇ 2 and ⁇ 2 are weighted values.
  • calculation module 44 is also used to:
  • Hash power token ⁇ 1 ⁇ [ ⁇ 1-(calculated metric value)]
  • ⁇ 1 and ⁇ 1 are the power pricing coefficients.
  • the computing power routing information also includes:
  • the function identifier FID of the function that the computing container can provide the IP address prefix or IP address of the computing container where the function is located, and the function is a software-based computing program, functional program, and business software deployed in the computing container Or microservices.
  • the device in this embodiment can be used to implement the technical solution of the method embodiment shown in FIG. 6, and its implementation principles and technical effects are similar, and will not be repeated here.
  • the present application may divide the computing power routing device into functional modules according to the foregoing method examples.
  • each functional module may be divided corresponding to each function, or two or more functions may be integrated into one processing module.
  • the above-mentioned integrated modules can be implemented in the form of hardware or software function modules. It should be noted that the division of modules in the embodiments of the present application is illustrative, and is only a logical function division, and there may be other division methods in actual implementation.
  • FIG. 18 is a schematic structural diagram of an apparatus provided by this application. As shown in FIG. 18, the apparatus 300 includes: a memory 301 and a processor 302;
  • the memory 301 is used to store computer programs
  • the processor 302 is configured to execute a computer program stored in the memory to implement the computing power routing method in the foregoing embodiment. For details, refer to the relevant description in the foregoing method embodiment.
  • the memory 301 may be independent or integrated with the processor 302.
  • the apparatus 300 may further include:
  • the bus 303 is used to connect the memory 301 and the processor 302.
  • this embodiment further includes: a communication interface 304, and the communication interface 304 may be connected to the processor 302 through a bus 303.
  • the processor 302 may control the communication interface 303 to implement the aforementioned receiving and sending functions of the apparatus 300.
  • the device can be used to execute various steps and/or processes corresponding to the computing power router or the leaf router in the foregoing method embodiment.
  • the present application also provides a readable storage medium.
  • the readable storage medium stores an execution instruction.
  • the device shown in FIG. 18 executes the computing power in the foregoing method embodiment. Routing method.
  • a person of ordinary skill in the art can understand that: in the foregoing embodiments, it may be implemented in whole or in part by software, hardware, firmware, or any combination thereof.
  • software it can be implemented in the form of a computer program product in whole or in part.
  • the computer program product includes one or more computer instructions.
  • the computer program instructions When the computer program instructions are loaded and executed on the computer, the processes or functions according to the embodiments of the present application are generated in whole or in part.
  • the computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices.
  • Computer instructions may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium.
  • computer instructions may be transmitted from a website, computer, server, or data center through a cable (such as Coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (such as infrared, wireless, microwave, etc.) to transmit to another website site, computer, server or data center.
  • the computer-readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server or data center integrated with one or more available media.
  • the usable medium may be a magnetic medium (for example, a floppy disk, a hard disk, and a magnetic tape), an optical medium (for example, a DVD), or a semiconductor medium (for example, a solid state disk (SSD)).

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Computer Security & Cryptography (AREA)
  • Computer Hardware Design (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

本申请提供一种算力路由方法及装置。该方法包括:算力路由器接收至少一个计算容器的算力路由信息,算力路由信息包括算力编码或计算度量值;算力路由器根据算力路由信息确定路由信息库RIB,RIB包括计算容器可提供的功能的功能标识FID或功能所在计算容器的IP地址前缀中的一个和算力编码或计算度量值中的一个,功能包括部署在计算容器内的软件化的计算程序、功能程序、业务软件或微服务中的任一个;算力路由器接收报文,根据转发信息库FIB转发报文,FIB根据RIB确定,FIB包括FID或IP地址前缀。从而可以将计算任务相关的报文路由到相应的计算容器,实现了低成本共享流量。

Description

算力路由方法及装置
本申请要求于2019年09月12日提交中国专利局、申请号为2019108668714、申请名称为“算力路由方法及装置”的中国专利申请的优先权,其全部内容通过引用结合在本申请中。
技术领域
本申请涉及通信技术领域,尤其涉及一种算力路由方法及装置。
背景技术
人工智能技术带来信息社会向着智能社会转变,智能社会的一个典型特征即物理世界和数字世界的深度融合,未来数字世界通过物联网及增强现实等技术提供的传感器和执行器与真实世界产生互动。交互方式上由传统社会人与人的交互,转变为数据与计算能力的交互,用户和应用使用互联网的首要需求可能不再是获得信息,而是获得计算资源、获得功能,核心的目的是获得计算的结果。
目前,流量仍是互联网行业的核心竞争力,传统的互联网流量获取方式包括搜索引擎、聚合网站、移动分发、门户网站等等,按照现有的流量获取方式,对于互联网企业而言获取流量成本较高,几乎占到互联网企业总成本30%以上,如何低成本高效率地获取流量始终是互联网企业的一大难题。在智能社会里,数以万计的互联网公司是流量的获取者、算力(也称为计算资源)的需求者,而运营商拥有大量的边缘机房和网络云化后边缘算力的优势,是算力的拥有者和流量的引导者。
如何动态、快速匹配上述供需关系,实现低成本共享流量,是亟需解决的问题。
发明内容
本申请提供一种算力路由方法及装置,以解决低成本共享流量的问题。
第一方面,本申请实施例提供一种算力路由方法,包括:
算力路由器接收至少一个计算容器的算力路由信息,所述算力路由信息包括算力编码或计算度量值;
所述算力路由器根据所述算力路由信息确定路由信息库RIB,所述RIB包括计算容器可提供的功能的功能标识FID或所述功能所在计算容器的IP地址前缀中的一个和所述算力编码或所述计算度量值中的一个,所述功能包括部署在所述计算容器内的软件化的计算程序、功能程序、业务软件或微服务中的任一个;
所述算力路由器接收报文,根据路由信息库FIB转发所述报文,所述FIB根据所述RIB确定,所述FIB包括FID或IP地址前缀。
通过第一方面提供的算力路由方法,通过算力路由器接收至少一个计算容器的算力路由信息,算力路由信息包括算力编码或计算度量值,然后将算力路由信息发送给 网络中的其它算力路由器,然后根据算力路由信息确定RIB,最后算力路由器接收到报文后,根据FIB转发报文,最终路由到相应的计算容器,其中的FIB根据RIB确定,算力路由信息代表计算容器可提供的算力,从而可以将计算任务相关的报文路由到相应的计算容器,实现用户体验最优、计算资源利用率最优、网络效率最优,路由过程等同算力竞价,流量使用者竞价得算力,流量拥有者释放算力,动态、快速匹配了运营商与互联网企业的流量供需关系,实现了低成本共享流量。
在一种可能的设计中,所述方法还包括:
所述算力路由器将所述算力路由信息发送给网络中的其它算力路由器。
在一种可能的设计中,所述算力路由器在所述FIB中未查找到与所述报文对应的FIB项目时,所述方法还包括:
所述算力路由器根据所述RIB中的算力编码或所述计算度量值确定与所述报文对应的FIB项目。
在一种可能的设计中,所述RIB包括FID、计算度量值、IP地址前缀和出端口;或者,
FID、算力编码、IP地址前缀和出端口;或者,
FID、计算度量值、算力路由器IP地址和出端口;或者,
FID、算力编码、算力路由器IP地址和出端口;或者,
IP地址前缀、计算度量值和出端口;或者,
IP地址前缀、算力编码和出端口;或者,
FID、计算度量值和出端口;或者,
FID、算力编码和出端口;
所述FIB包括:FID、IP地址前缀和出端口;或者,
FID、算力路由器IP地址和出端口;或者,
FID和出端口。
在一种可能的设计中,若所述报文携带FID,所述根据路由信息库FIB转发所述报文,包括:
所述算力路由器从所述FIB中查找与所述FID对应的IP地址前缀和出端口;
所述算力路由器将所述FID添加与所述FID对应的IP地址前缀得到目的地IP地址,根据所述目的地IP地址转发所述报文;或者,
所述算力路由器从所述FIB中查找出与所述FID对应的算力路由器IP地址和出端口;
所述算力路由器将所述算力路由器IP地址作为隧道目的地址对所述报文添加隧道封装,或,将所述算力路由器IP地址作为分段路由SR的段标识添加SR封装;
所述算力路由器转发添加隧道封装或添加SR封装的报文;或者,
所述算力路由器从所述FIB查找出与所述FID或IP地址前缀对应的出端口;
所述算力路由器从所述出端口转发所述报文。
在一种可能的设计中,若所述报文携带IP地址,所述根据路由信息库FIB转发所述报文,包括:
所述算力路由器根据所述IP地址转发所述报文。
在一种可能的设计中,所述算力编码包括计算时延评估值和/或计算容量等级、机器类型或与机器类型相关的加权值α2和服务器负载,所述算力路由信息包括算力编码时,所述方法还包括:
所述算力路由器根据所述算力编码和如下公式计算所述计算度量值:
Figure PCTCN2020115154-appb-000001
或者,
Figure PCTCN2020115154-appb-000002
或者,
计算度量值=γ2×计算时延评估值;
其中,β2、γ2为加权值。
在一种可能的设计中,所述方法还包括:
若所述算力路由器接收到应答报文,则将所述FIB项目确定为有效的FIB项目;
若所述算力路由器未接收到所述应答报文,则删除所述FIB项目。
通过本实施方式提供的方法,对于基于会话的业务,通常在会话建立阶段存在会话请求和会话应答的消息报文交互,当未接收到应答报文时,说明会话建立不成功,此时若不将接收到会话请求消息报文时所确定的FIB项目删除则相当于留下了不再会用到的垃圾项目,造成不必要的存储空间浪费。
第二方面,本申请实施例提供一种算力路由方法,包括:
叶子路由器获取计算容器的算力编码,所述叶子路由器部署在所述计算容器内;
所述叶子路由器将所述计算容器的算力路由信息发送到与所述叶子路由器连接的算力路由器,以使所述算力路由器将所述算力路由信息发送给网络的其它算力路由器,所述算力路由信息包括所述算力编码或计算度量值;
所述叶子路由器接收报文,转发所述报文。
通过第二方面提供的算力路由方法,通过叶子路由器获取计算容器的算力编码,将包括算力编码或计算度量值的算力路由信息发送到与叶子路由器连接的算力路由器,使得该算力路由器将算力路由信息发送给网络的其它算力路由器,最后叶子路由器接收报文,转发报文,具体是转发报文到计算容器。算力路由信息代表计算容器可提供的算力,从而可以将计算任务相关的报文路由到相应的计算容器,实现用户体验最优、计算资源利用率最优、网络效率最优,路由过程等同算力竞价,流量使用者竞价得算力,流量拥有者释放算力,动态、快速匹配了运营商与互联网企业的流量供需关系,实现了低成本共享流量。
在一种可能的设计中,所述叶子路由器获取计算容器的算力编码,包括:
所述叶子路由器从所述计算容器和/或测量工具获取所述算力编码,所述算力编码包括计算时延评估值和/或计算容量等级、机器类型或与机器类型相关的加权值α2和服务器负载,所述计算时延评估值由所述测量工具测量。
在一种可能的设计中,若所述算力路由信息包括所述计算度量值,所述方法还包括:
所述叶子路由器根据所述算力编码和如下公式计算所述计算度量值:
Figure PCTCN2020115154-appb-000003
或者,
Figure PCTCN2020115154-appb-000004
或者,
计算度量值=γ2×计算时延评估值;
其中,β2、γ2为加权值。
在一种可能的设计中,所述方法还包括:
所述叶子路由器根据所述算力编码和如下公式计算算力代币:
算力代币=α1×[β1-(计算度量值)]
其中,α1、β1为算力定价系数。
在一种可能的设计中,所述算力路由信息还包括:
所述计算容器可提供的功能的功能标识FID、所述功能所在计算容器的IP地址前缀或IP地址,所述功能为部署在所述计算容器内的软件化的计算程序、功能程序、业务软件或微服务。
在一种可能的设计中,所述叶子路由器接收报文之后,所述方法还包括:
所述叶子路由器触发合约平台将所述算力代币转账给算力使用者。
在一种可能的设计中,所述方法还包括:
所述叶子路由器接收所述计算容器发送的应答报文,所述应答报文携带所述计算容器的IP地址;
所述叶子路由器将所述应答报文通过与所述叶子路由器连接的算力路由器发送给客户端,所述应答报文用于客户端进行后续通信时将报文中携带的FID改为所述计算容器的IP地址。
第三方面,本申请实施例提供一种算力路由装置,包括:
接收模块,用于接收至少一个计算容器的算力路由信息,所述算力路由信息包括算力编码或计算度量值;
确定模块,用于根据所述算力路由信息确定路由信息库RIB,所述RIB包括计算容器可提供的功能的功能标识FID或所述功能所在计算容器的IP地址前缀中的一个和所述算力编码或所述计算度量值中的一个,所述功能包括部署在所述计算容器内的软件化的计算程序、功能程序、业务软件或微服务中的任一个;
所述接收模块还用于接收报文;
发送模块,用于根据路由信息库FIB转发所述报文,所述FIB根据所述RIB确定,所述FIB包括FID或IP地址前缀。
在一种可能的设计中,所述发送模块还用于:
将所述算力路由信息发送给网络中的其它算力路由器。
在一种可能的设计中,所述算力路由器在所述FIB中未查找到与所述报文对应的FIB项目时,所述确定模块还用于:
根据所述RIB中的算力编码或所述计算度量值确定与所述报文对应的FIB项目。
在一种可能的设计中,所述RIB包括FID、计算度量值、IP地址前缀和出端口;或者,
FID、算力编码、IP地址前缀和出端口;或者,
FID、计算度量值、算力路由器IP地址和出端口;或者,
FID、算力编码、算力路由器IP地址和出端口;或者,
IP地址前缀、计算度量值和出端口;或者,
IP地址前缀、算力编码和出端口;或者,
FID、计算度量值和出端口;或者,
FID、算力编码和出端口;
所述FIB包括:FID、IP地址前缀和出端口;或者,
FID、算力路由器IP地址和出端口;或者,
FID和出端口。
在一种可能的设计中,若所述报文携带FID,所述发送模块用于:
从所述FIB中查找与所述FID对应的IP地址前缀和出端口;
将所述FID添加与所述FID对应的IP地址前缀得到目的地IP地址,根据所述目的地IP地址转发所述报文;或者,
从所述FIB中查找出与所述FID对应的算力路由器IP地址和出端口;
将所述算力路由器IP地址作为隧道目的地址对所述报文添加隧道封装,或,将所述算力路由器IP地址作为分段路由SR的段标识添加SR封装;
转发添加隧道封装或添加SR封装的报文;或者,
从所述FIB查找出与所述FID或IP地址前缀对应的出端口;
从所述出端口转发所述报文。
在一种可能的设计中,所述算力编码包括计算时延评估值和/或计算容量等级、机器类型或与机器类型相关的加权值α2和服务器负载,所述算力路由信息包括算力编码时,所述装置还包括:
计算模块,用于根据所述算力编码和如下公式计算所述计算度量值:
Figure PCTCN2020115154-appb-000005
或者,
Figure PCTCN2020115154-appb-000006
或者,
计算度量值=γ2×计算时延评估值;
其中,β2、γ2为加权值。
在一种可能的设计中,所述装置还包括:
处理模块,用于若所述接收模块接收到应答报文,将所述FIB项目确定为有效的FIB项目;
若所述接收模块未接收到所述应答报文,则删除所述FIB项目。
第四方面,本申请实施例提供一种算力路由装置,包括:
获取模块,用于获取计算容器的算力编码,所述叶子路由器部署在所述计算容器内;
发送模块,用于将所述计算容器的算力路由信息发送到与所述叶子路由器连接的算力路由器,以使所述算力路由器将所述算力路由信息发送给网络的其它算力路由器,所述算力路由信息包括所述算力编码或计算度量值;
接收模块,用于接收报文;
所述发送模块还用于转发所述报文。
在一种可能的设计中,所述获取模块用于:
从所述计算容器和/或测量工具获取所述算力编码,所述算力编码包括计算时延评估值和/或计算容量等级、机器类型或与机器类型相关的加权值α2和服务器负载,所述计算时延评估值由所述测量工具测量。
在一种可能的设计中,若所述算力路由信息包括所述计算度量值,所述装置还包括:
计算模块,用于根据所述算力编码和如下公式计算所述计算度量值:
Figure PCTCN2020115154-appb-000007
或者,
Figure PCTCN2020115154-appb-000008
或者,
计算度量值=γ2×计算时延评估值;
其中,β2、γ2为加权值。
在一种可能的设计中,所述计算模块还用于:
根据所述算力编码和如下公式计算算力代币:
算力代币=α1×[β1-(计算度量值)]
其中,α1、β1为算力定价系数。
在一种可能的设计中,所述算力路由信息还包括:
所述计算容器可提供的功能的功能标识FID、所述功能所在计算容器的IP地址前缀或IP地址,所述功能为部署在所述计算容器内的软件化的计算程序、功能程序、业务软件或微服务。
第五方面,本申请实施例提供一种装置,包括:存储器和处理器;
处理器;以及
存储器,用于存储所述处理器的可执行指令;
其中,所述处理器配置为经由执行所述可执行指令来执行第一方面及第一方面任一种可能的设计中或第二方面及第二方面任一种可能的设计中的算力路由方法。
在一种可能的设计中,上述装置可以是路由设备,也可以是芯片。
第六方面,本申请提供了一种计算机可读存储介质,所述计算机可读存储介质中存储有指令,当其在计算机上运行时,使得计算机执行上述各方面所述的方法。
第七方面,本申请提供了一种包含指令的计算机程序产品,当其在计算机上运行时,使得计算机执行上述各方面所述的方法。
附图说明
图1为本申请提供的一种算力网络架构示意图;
图2为本申请提供的另一种算力网络架构示意图;
图3为本申请提供的又一种算力网络架构示意图;
图4为本申请提供的算力路由方法及装置的应用场景示意图;
图5为本申请提供的一种算力路由方法实施例的流程图;
图6为本申请提供的一种算力路由方法实施例的流程图;
图7为本申请提供的一种算力路由方法实施例的流程图;
图8为本申请提供的一种算力路由方法实施例的流程图;
图9为本申请提供的一种算力路由方法实施例的流程图;
图10为本申请提供的一种算力路由方法实施例的流程图;
图11为本申请提供的一种算力路由器的功能结构示意图;
图12为本申请提供的一种叶子路由器的功能结构示意图;
图13为本申请提供的一种算力路由装置实施例的结构示意图;
图14为本申请提供的一种算力路由装置实施例的结构示意图;
图15为本申请提供的一种算力路由装置实施例的结构示意图;
图16为本申请提供的一种算力路由装置实施例的结构示意图;
图17为本申请提供的一种算力路由装置实施例的结构示意图;
图18为本申请提供的一种装置结构示意图。
具体实施方式
本申请实施例中,“示例性的”或者“例如”等词用于表示作例子、例证或说明,本申请实施例中被描述为“示例性的”或者“例如”的任何实施例或方案不应被解释为比其它实施例或方案更优选或更具优势。确切而言,使用“示例性的”或者“例如”等词旨在以具体方式呈现相关概念。
本申请中,“至少一个”是指一个或者多个,“多个”是指两个或两个以上。“和/或”,描述关联对象的关联关系,表示可以存在三种关系,例如,A和/或B,可以表示:单独存在A,同时存在A和B,单独存在B的情况,其中A,B可以是单数或者复数。字符“/”一般表示前后关联对象是一种“或”的关系。
首先,下面对本申请实施例中的部分用语进行解释说明,以便于本领域技术人员理解。
1、功能(Function),为部署在计算站点内的软件化的计算程序、功能程序、业务软件或微服务。
2、计算容器,可以是计算站点、服务器、虚拟机或容器(Container),或者是更小的提供计算资源的单元,其中的计算站点可以是用户侧的客户终端设备(customer-premises equipment,CPE)站点,如企业网所在站点(企业网有网关设备(Gateway,GW)和服务器,如网管设备和服务器都可以为计算站点)或者家庭网络所在站点。
3、算力编码,是计算容器可提供的计算能力的一种技术抽象,为一种计算度量值(Computing Metric)在路由器间扩散和传播,算力编码包括计算时延评估值和/或计算容量等级、机器类型或与机器类型相关的加权值α2和服务器负载。
4、算力代币,是计算容器可提供的计算能力的一种商业抽象,与运营商计算容器可提供的计算能力是等价的,算力代币以算力编码的技术形式在算力网络中传递。
本申请实施例中,为实现低成本共享流量,动态、快速匹配运营商与互联网企业的流量供需关系,适应智能社会的发展,本申请提供一种算力路由方法及装置,通过部署在计算容器内的叶子路由器将计算容器的算力路由信息发送到与叶子路由器连接 的算力路由器,由该算力路由器通告给网络的其它算力路由器,使得其它算力路由器能获得网络所有计算容器的算力路由信息,算力路由器根据算力路由信息确定路由信息库(routing information base,RIB),算力路由器接收到报文后,根据转发信息库(forwarding information base,FIB)和报文中携带的FID将报文路由到FID所在的计算容器。从而可以将计算任务相关的报文路由到相应的计算容器,路由过程等同算力竞价,流量使用者竞价得算力,流量拥有者释放算力,动态、快速匹配了运营商与互联网企业的流量供需关系,实现了低成本共享流量,对于运营商:边缘机房、网络云化优势得释放,变现“网络连接及其相关的数字资产(计算和存储)”,从而助力运营商向信息和通信技术(Information and Communications Technology,ICT)服务转型;对于互联网公司:从运营商按需调用轻资产,流量获取来自算力网络,节省大量重资产投资与流量获取成本,助力互联网公司向轻资产模式转型。
下面结合附图详细说明本申请提供的算力路由方法及装置的具体实现过程。
本申请实施例可应用于算力网络中,图1为本申请提供的一种算力网络架构示意图,如图1所示,该网络由计算容器和算力网络(通常为运营商网络,Carrier Network)组成,可以分成Underlay和叠加层(Overlay)两层,叶子路由器/(也可以称为叶子节点)是算力网络的对外的延伸,是感知外部算力的触角,叶子路由器部署在计算容器内,图1中以计算容器为计算站点(Site)为例进行说明。
计算站点可以是用户侧的CPE站点,如企业网所在站点,或者家庭网络所在站点,家庭网有家庭网关和计算机、手机等设备,CPE站点内的这些设备属于Underlay设备。
计算站点还可以是网络侧的站点:如路由器、PoP、CO或基站所在站点,站点内也有网络设备或网关设备和服务器,站点内的这些设备属于Underlay设备。网络侧的站点通常是一些被改造为小型数据中心或放置有服务器等计算资源的站点。
计算站点内可以部署,通常在Underlay设备上面,部署软件化的计算程序、功能程序、业务软件或微服务等,以功能(Function)代表部署软件化的计算程序、功能程序、业务软件或微服务,即图1中所示的FX1…FXk,或,FY1…FYm,叶子路由器部署在计算站点内;这些软件或设备属于叠加层,也就是叠加在Underlay上面的一层的意思。
计算站点间由算力网络互联,算力网络由算力路由器(也称算力节点)组成,还可以包含普通路由器。算力路由器用于支持算力路由,普通路由器没有算力路由能力。普通路由器属于Underlay设备,算力路由功能的软件或硬件属于叠加层,即图中的算力路由器,叠加了算力路由功能的软件或硬件的节点就是算力节点,叠加了算力路由功能的软件或硬件的路由器就是算力路由器。叠加层与Underlay设备可以集成在一起,也可以分离单独成设备,如叠加节点/设备。本申请实施例中,算力即计算资源。
图2为本申请提供的另一种算力网络架构示意图,如图2所示,与图1不同在于算力路由器由中间的网络搬移到计算站点内,所以算力网络是一个叠加层,由算力路由器构成。叠加层与计算站点内的Underlay设备可以集成在一起,也可以分离单独成设备,如Overlay算力节点/设备。
图3为本申请提供的又一种算力网络架构示意图,如图3所示,它是一种图1与图2的混合架构,算力路由器既可以位于中间的网络,也可以在计算站点内。
本申请实施例中,算力网络实现三个主要功能:
(1)功能和/或算力自动发现
由叶子路由器将站点计算资源和/或可提供的功能以及功能所在计算容器的IP地址前缀/地址对算力网络通告,使算力网络可以发现功能和/或算力,计算资源本实施例中以“算力编码”的形式表达。其中,算力路由信息可以包括站点计算资源和/或可提供的功能以及功能所在计算容器的IP地址前缀/地址。这种通告可以通过扩展DC内部协议,或通过扩展路由协议(如内部网关协议(Interior Gateway Protocol,IGP)IGP/边界网关协议(Border Gateway Protocol,BGP)实现。
(2)功能和/或算力通告
在算力路由器间通告算力路由信息,算力路由信息还可以包括叶子路由器IP地址前缀/地址或与叶子路由器所连接的算力路由器的IP地址前缀/地址,算力路由器间通告,使整个算力网络的各个算力路由器都可以发现功能和/或算力。这种通告同样可以通过扩展DC内部协议,或通过扩展路由协议(如IGP/BGP)实现。
(3)基于功能的隧道技术
由于Underlay设备间是IP可达的,即underlay设备为算力路由器提供了可达性,每个算力路由器(节点)之间可以直接通信,构成逻辑上全网状(full mesh)的对等(peer)关系。算力路由器(节点)之间可以用多协议标记交换(Multiprotocol Label Switching,MPLS)、IP等隧道技术逻辑互联,或分段路由(Segment Routing,SR)技术互联。
图4为本申请提供的算力路由方法及装置的应用场景示意图,如图4所示,计算容器以计算站点为例,本实施例的方法包括以下流程:
S1、互联网公司与运营商通过智能合约约定算力租金,互联网公司根据运营商计算站点可提供的计算能力,对所要租的算力合约转账相应数量的“算力代币”。
S2、将“算力代币(数字钱币)”转换为“算力编码”,用私钥签名类似于签名的“二维码”,用私钥签名。
S3、部署在计算站点内的叶子路由器,将“算力编码”或由“算力编码”推算出来的计算度量值(Computing Metric)通过路由协议(IGP/BGP)发给算力网络,首先发送到入口路由器(为算力路由器)。
S4、算力网络中的入口路由器识别“二维码(即算力编码或计算度量值)”,并向合约验证签名的真实性以及用公钥验证算力编码或计算度量值签名的合法性,进行算力路由“锁”解锁;其中,所谓算力路由锁解锁即允许算力编码或计算度量值在算力网络中传播
S5、算力编码或计算度量值为一种计算度量值通过路由器(为算力路由器)在算力网络内传播。
S6、路由过程相当于“算力竞价”,路由器基于“算力编码”或计算度量值所代表的计算站点可提供的计算能力,通过合理的算法综合考量并选定出最合适的计算容器,得到最佳算力路由,从而流量被路由到“最高出价”的互联网公司租用站点进行计算。
S7、在成功引导计算流量后,算力拥有者(如运营商)从智能合约获得约定的相应的“算力代币”。
需要说明的是,“算力代币”与运营商计算站点可提供的计算能力是等价的,即“算力代币”是计算站点可提供的计算能力的一种商业抽象,而“算力编码”是计算站点可提供的计算能力的一种技术抽象,“算力代币”以“算力编码”的技术形式在算力网络中传递。
其中,叶子路由器/叶子节点可以每个互联网或针对每种业务设置一个,通常是设置于数据中心或计算容器内,可以以硬件设备或以软件的形态存在,例如软件化的虚拟路由器。智能合约(Smart Contract)可以用一个平台维护,可以是去中心化的区块链平台,也可以是有中心的代理平台,用于实现算力使用者(如互联网公司或第三方)与算力拥有者(如运营商或企业)之间的算力代币支付或转账。
本申请涉及的网元包括算力路由器和叶子路由器。
图5为本申请提供的一种算力路由方法实施例的流程图,本实施例的执行主体可以为算力路由器,也可以为算力节点,如图5所示,本实施例的方法可以包括:
S101、算力路由器接收至少一个计算容器的算力路由信息,算力路由信息包括算力编码或计算度量值。
具体地,算力路由器从与其连接的至少一个叶子路由器接收至少一个计算容器的算力路由信息,叶子路由器部署在计算容器内,一个计算容器内部署至少一个叶子路由器。
可选的,在S101之后,本实施例的方法还可以包括:
S104、算力路由器将算力路由信息发送给网络中的其它算力路由器。
算力路由信息包括算力编码或计算度量值,若算力编码为私钥签名过的算力编码,则S104之前,还可以包括:算力路由器验证签名为合法后,解锁算力路由锁,并触发算力交易,得到算力编码的有效期,即可以使用的有效时间,例如可以向合约平台或有关部门验证签名为合法。当算力编码的有效期到期,算力路由锁将对该算力编码重新上锁,即将不允许相应的算力编码或由其推算出来的计算度量值在网络中传播。
具体地,算力路由器可以通过IGP/BGP将算力路由信息发送给网络中的其它算力路由器。
S102、算力路由器根据算力路由信息确定RIB,RIB包括计算容器可提供的功能的功能标识(Function ID,FID)或功能所在计算容器的IP地址前缀中的一个和算力编码或计算度量值中的一个,功能包括部署在计算容器内的软件化的计算程序、功能程序、业务软件或微服务中的任一个。当算力编码的有效期到期,则算力编码相应的RIB项目将被删除。
具体地,RIB包括计算容器可提供的功能的FID和算力编码,或者,RIB包括计算容器可提供的功能的FID和计算度量值,或者,RIB包括功能所在计算容器的IP地址前缀和算力编码,或者,RIB包括功能所在计算容器的IP地址前缀和计算度量值。
S103、算力路由器接收报文,根据FIB转发报文,FIB根据RIB确定,FIB至少包括FID或IP地址前缀。
具体地,算力路由信息包括算力编码或计算度量值,算力路由器根据算力路由信息确定RIB,具体可以为:算力路由器根据“算力编码”或计算度量值得到“算力编码”或计算度量值与出端口的映射关系,或者,根据“算力编码”或计算度量值得到FID、“算 力编码”或计算度量值与出端口的映射关系,或者,根据“算力编码”或计算度量值得到IP地址前缀、“算力编码”或计算度量值与出端口的映射关系等。
还可以是:计算度量值所代表的计算站点可提供的计算能力,通过合理的算法综合考量并选定出最合适的计算容器,例如通过比较“算力编码”或计算度量值后选最小“算力编码”或计算度量值所对应的计算容器,或者,例如通过比较“算力编码”或计算度量值后选最小“算力编码”或计算度量值所对应的计算站点,从而得到相应的FIB项目。
可选的,RIB包括FID、计算度量值、IP地址前缀和出端口;或者,
FID、算力编码、IP地址前缀和出端口;或者,
FID、计算度量值、算力路由器IP地址和出端口;或者,
FID、算力编码、算力路由器IP地址和出端口;或者,
IP地址前缀、计算度量值和出端口;或者,
IP地址前缀、算力编码和出端口;或者,
FID、计算度量值和出端口;或者,
FID、算力编码和出端口;
相应地,FIB包括:FID、IP地址前缀和出端口,或者,FID、算力路由器IP地址和出端口,或者,FID和出端口。
在一种可能的实施方式中,算力路由器根据FIB转发报文时,算力路由器在FIB中未查找到与报文对应的FIB项目时,本实施例的方法还可以包括:
算力路由器根据RIB中的算力编码或计算度量值确定与报文对应的FIB项目。具体可以是确定与报文携带的FID或IP地址前缀对应的FIB项目,例如算力路由器基于“算力编码”或计算度量值所代表的计算站点可提供的计算能力,通过合理的算法综合考量并选定出最合适的计算容器,例如通过比较“算力编码”或计算度量值后选最小“算力编码”或计算度量值所对应的计算容器,或者,例如通过比较“算力编码”或计算度量值后选最小“算力编码”或计算度量值所对应的计算站点,从而得到相应的FIB项目。
具体地,RIB至少包括计算容器可提供的功能的FID或功能所在计算容器的IP地址前缀中的一个和算力编码或计算度量值中的一个,若RIB包括算力编码,则算力路由器比较RIB中所有的算力编码的大小,如一共有k个算力编码,则比较算力编码1、算力编码2、…算力编码k的大小,来计算路由,得到目的IP地址和/或出端口,从而得到FIB项目,该FIB项目包括FID、目的IP地址和出端口。若RIB包括计算度量值,则算力路由器比较RIB中所有的计算度量值的大小,如一共有k个计算度量值,则比较计算度量值1、计算度量值2、…计算度量值k的大小,来计算路由,得到目的IP地址和/或出端口,从而得到FIB项目,该FIB项目包括FID、目的IP地址和/或出端口。
具体地,若报文携带FID,根据FIB转发报文,有如下三种可实施的方式:
方式一:算力路由器从FIB中查找与FID对应的IP地址前缀和出端口。
算力路由器将FID添加与FID对应的IP地址前缀得到目的地IP地址,根据目的地IP地址转发报文。
方式二:算力路由器从FIB中查找出与FID对应的算力路由器IP地址和出端口。
算力路由器将算力路由器IP地址作为隧道目的地址对报文添加隧道封装,或,将算力路由器IP地址作为分段路由SR的段标识添加SR封装。
算力路由器将添加隧道封装或添加SR封装的报文转发。
方式三:算力路由器从FIB查找出与FID或IP地址前缀对应的出端口,算力路由器将报文从出端口转发。
若报文携带IP地址,根据路由信息库FIB转发报文,可以是:算力路由器根据IP地址转发报文。
可选的,算力编码包括计算时延评估值和/或计算容量等级、机器类型或与机器类型相关的加权值α2和服务器负载,即:算力编码包括计算时延评估值,或者,算力编码包括计算容量等级、机器类型或与机器类型相关的加权值α2和服务器负载,或者,算力编码包括计算时延评估值和计算容量等级、机器类型或与机器类型相关的加权值α2和服务器负载,算力路由信息包括算力编码时,本实施例的方法还可以包括:
算力路由器根据算力编码和如下公式计算计算度量值:
Figure PCTCN2020115154-appb-000009
或者,
Figure PCTCN2020115154-appb-000010
或者,
计算度量值=γ2×计算时延评估值;
其中,β2、γ2为加权值。α2、β2、γ2可根据算力拥有者与使用者间的合约或服务水平协议(Service Level Agreement,SLA)给出。
若算力路由信息包括计算度量值,则该计算度量值是由叶子路由器通过上述计算公式计算得到。
计算度量值(Compute Metric)越小,计算容器的计算能力越强,计算服务质量越好,所需要支付的算力代币值也越大,上面的算法仅是一种实施例,也可是通过其他的算法获得。
进一步地,本实施例方法还可以包括:
若算力路由器接收到应答报文,则将所确定的FIB项目确定为有效的FIB项目;
若算力路由器未接收到应答报文,则删除所确定的FIB项目。这样处理的有益效果在于:对于基于会话的业务,通常在会话建立阶段存在会话请求和会话应答的消息报文交互,当未接收到应答报文时,说明会话建立不成功,此时若不将接收到会话请求消息报文时所确定的FIB项目删除则相当于留下了不再会用到的垃圾项目,造成不必要的存储空间浪费。
S104中算力路由器接收到报文后,根据FIB转发报文,路由到到相应的计算容器。
本实施例提供的算力路由方法,通过算力路由器接收至少一个计算容器的算力路由信息,算力路由信息包括算力编码或计算度量值,然后将算力路由信息发送给网络中的其它算力路由器,然后根据算力路由信息确定RIB,最后算力路由器接收到报文后,根据FIB转发报文,最终路由到相应的计算容器,其中的FIB根据RIB确定,算力路由信息代表计算容器可提供的算力,从而可以将计算任务相关的报文路由到相应的计算容器,实现用户体验最优、计算资源利用率最优、网络效率最优,路由过程等同算力竞价,流量使用者竞价得算力,流量拥有者释放算力,动态、快速匹配了运营 商与互联网企业的流量供需关系,实现了低成本共享流量。
图6为本申请提供的一种算力路由方法实施例的流程图,本实施例的执行主体可以为叶子路由器,也可以为叶子节点,如图6所示,本实施例的方法可以包括:
S201、叶子路由器获取计算容器的算力编码,叶子路由器部署在计算容器内。
在一种可能的实施方式中,叶子路由器获取计算容器的算力编码,具体可以是:
叶子路由器从计算容器和/或测量工具获取算力编码,算力编码包括计算时延评估值和/或计算容量等级、机器类型或与机器类型相关的加权值α2和服务器负载,计算时延评估值由测量工具测量,测量工具如Openstack、K8S/Kubernetes、或操作系统等IT软件。
其中,计算容量等级例如1台服务器对应的计算容量等级为255,2台服务器对应的计算容量等级为127,3-10台服务器对应的计算容量等级为63,10-1000台服务器对应的计算容量等级为15,1000-10000台服务器对应的计算容量等级为7,10000台以上服务器对应的计算容量等级为3。
S202、叶子路由器将计算容器的算力路由信息发送到与叶子路由器连接的算力路由器,以使算力路由器将算力路由信息发送给网络的其它算力路由器,算力路由信息包括算力编码或计算度量值。
可选的,算力路由信息还可以包括:
计算容器可提供的功能的FID、功能所在计算容器的IP地址前缀或IP地址,功能为部署在计算容器内的软件化的计算程序、功能程序、业务软件或微服务。
可选的,算力路由信息例如可以包括:路由目标(Route-Target,RT)=6、路由区分符(Route-Distinguisher,RD)=功能路由通告(Function Advertisement Route)、功能长度(1个八位字节(octet))、FID、IP地址前缀长度(1 octet)、IP地址前缀(4 or 16 octets)(功能和其所在计算容器的IP地址前缀(IPv4或IPv6))和算力编码(8 octet)。
在一种可能的实施方式中,若算力路由信息包括计算度量值,本实施例的方法还可以包括:
叶子路由器根据算力编码和如下公式计算计算度量值:
Figure PCTCN2020115154-appb-000011
或者,
Figure PCTCN2020115154-appb-000012
或者,
计算度量值=γ2×计算时延评估值;
其中,β2、γ2为加权值,α2、β2、γ2可根据算力拥有者与使用者间的合约或服务水平协议(Service Level Agreement,SLA)给出。
进一步地,还可以包括:
叶子路由器根据算力编码和如下公式计算算力代币:
算力代币=α1×[β1-(计算度量值)]
其中,α1、β1为算力定价系数,可以由算力拥有者(如运营商)确定,也可以算力拥有者与算力使用者共同确定,或由合约所在平台确定。
S203、叶子路由器接收报文,转发报文。
进一步地,在叶子路由器接收报文之后,本实施例的方法还可以包括:叶子路由器触发合约平台将算力代币转账给算力使用者。
可选的,本实施例的方法还可以包括:
叶子路由器接收计算容器发送的应答报文,应答报文携带计算容器的IP地址;
叶子路由器将应答报文通过与叶子路由器连接的算力路由器发送给客户端,应答报文用于客户端进行后续通信时将报文中携带的FID改为计算容器的IP地址,从而解决流黏性(Flow Affinity)问题。
本实施例提供的算力路由方法,通过叶子路由器获取计算容器的算力编码,将包括算力编码或计算度量值的算力路由信息发送到与叶子路由器连接的算力路由器,使得该算力路由器将算力路由信息发送给网络的其它算力路由器,最后叶子路由器接收报文,转发报文,具体是转发报文到计算容器。算力路由信息代表计算容器可提供的算力,从而可以将计算任务相关的报文路由到相应的计算容器,实现用户体验最优、计算资源利用率最优、网络效率最优,路由过程等同算力竞价,流量使用者竞价得算力,流量拥有者释放算力,动态、快速匹配了运营商与互联网企业的流量供需关系,实现了低成本共享流量。
下面通过四个具体的实施例,对图5和图6所示的技术方案进行具体的说明。
图7为本申请提供的一种算力路由方法实施例的流程图,如图5所示,本实施例的方法可以包括:
S301、叶子路由器获取计算容器的算力编码,叶子路由器部署在计算容器内。
具体地,例如有k个叶子路由器,叶子路由器1从计算容器和/或IT测量工具获取私钥签名过的算力编码1,叶子路由器2从计算容器和/或IT测量工具获取私钥签名过的算力编码2,……,叶子路由器k从计算容器和/或IT测量工具获取私钥签名过的算力编码k。
具体地,叶子路由器从计算容器获取计算容量等级、机器类型或与机器类型相关的加权值α2和服务器负载,从IT测量工具获取计算时延评估值,由IT测量工具测量计算时延评估值,IT测量工具例如Openstack、K8S/Kubernetes或操作系统等IT软件,算力编码包括计算时延评估值和/或计算容量等级、机器类型或与机器类型相关的加权值α2和服务器负载。
叶子路由器根据算力编码和如下公式计算计算度量值:
Figure PCTCN2020115154-appb-000013
或者,
Figure PCTCN2020115154-appb-000014
或者,
计算度量值=γ2×计算时延评估值;
其中,β2、γ2为加权值,α2、β2、γ2可根据算力拥有者与使用者间的合约/服务水平协议(Service Level Agreement,SLA)给出。
叶子路由器根据算力编码和如下公式计算算力代币:
算力代币=α1×[β1-(计算度量值)]
其中,α1、β1为算力定价系数,可以由算力拥有者(如运营商)确定,也可以算力拥有者与算力使用者共同确定,或由合约所在平台确定。
S302、叶子路由器1、叶子路由器k把算力路由信息通过IGP/BGP发送到与其连接的一个第一算力路由器。
其中,算力路由信息包括FID、叶子路由器所在计算容器的IP地址前缀和算力编码k,若S101中叶子路由器根据算力编码计算出计算度量值,则算力路由信息包括FID、叶子路由器所在计算容器的IP地址前缀和计算度量值k。若算力路由信息不包括计算度量值,则由接收到算力路由信息的算力路由器根据算力编码计算计算度量值。
S303、第一算力路由器向合约平台或有关部门验证签名为合法,则解锁“算力路由锁”,并触发“算力交易”,得到算力编码k的有效期。
其中,算力编码k的有效期即算力编码k的有效使用时间。
S304、第一算力路由器通过IGP/BGP将算力路由信息发送给网络的其它算力路由器。
具体地,网络的所有算力路由器均得到每一算力编码k与算力编码k对应的IP地址前缀和FID以及算力编码k的有效期,根据这些信息生成RIB,RIB包括算力编码k、IP地址前缀、FID和出端口的对应关系。
S305、第一算力路由器得到包括FID、算力编码k、IP地址前缀和出端口的RIB项目。
第二算力路由器得到包括FID、算力编码k和IP地址前缀和出端口的RIB项目。
S306、客户端发送携带FID的首包报文。
例如,首包报文中携带FIDx。
S307、接收到首包报文的第二算力路由器根据FID查FIB,若FIB不存在与该FID对应的项目,则根据FID查RIB,比较“算力编码1…k”的大小,来计算路由,得到与该FID对应的目的地IP地址前缀和出端口,生成基于该FID转发的FIB项目,FIB项目包括FID、IP地址前缀和出端口的对应关系。
S308、第二算力路由器将FID添加所得到的与该FID对应的目的地IP地址前缀得到目的地IP地址,将首包报文根据目的地IP地址转发到第一算力路由器,转发的首包报文携带目的地IP地址。
其中,该目的地IP地址为FID所在计算容器IP地址。
S309、第一算力路由器将携带目的地IP地址的首包报文根据目的地IP地址转发到叶子路由器K。
此处的叶子路由器K是与第一算力路由器连接的叶子路由器。
S310、叶子路由器K根据目的地IP地址转发首包报文到FID所在计算容器,并触发“算力代币”转账。
S311、客户端发送携带FIDX的后续报文至第二算力路由器。
S312、第二算力路由器根据报文携带的FID查FIB,得到与该FID对应的目的地IP地址前缀,将FID添加该目的地IP地址前缀得到目的地IP地址,将报文根据目的地IP地址转发到第一算力路由器,转发的首包报文携带目的地IP地址。
S313、第一算力路由器将携带目的地IP地址的报文根据目的地IP地址转发到叶子路由器。
S314、叶子路由器根据目的地IP地址转发报文到FID所在计算容器。
其中,触发算力代币转账或算力交易的条件可以是如算力编码成功注入算力网络时,或算力网络生成RIB/FIB,或如算力网络成功将报文路由到互联网公司/第三方所属的叶子路由器或计算站点。报文携带的是FID,FID不是IP地址,需要由特定域来携带,可以是放在IP报文的目的地址域中。仅算力网络入口(如入口算力路由器,图7中的第二算力路由器)要求基于计算度量值选择路由,以及基于FID的转发,其余路由器都能延用传统IP转发,因此可以保护现有IP路由器的投资。
图8为本申请提供的一种算力路由方法实施例的流程图,如图8所示,不仅仅算力网络入口(如入口算力路由器,图8中的第二算力路由器)要求基于计算度量值选择路由,以及基于FID的转发,而且算力网络出口(如出口算力路由器,图8中的第一算力路由器),其余路由器都能延用传统IP转发,因此可以保护现有IP路由器的投资。本实施例的方法可以包括:
S401-S403与S301-S303相同,可参见S301-S303的描述,此处不再赘述。
S404、第一算力路由器通过IGP/BGP将算力路由信息发送给网络的其它算力路由器。
具体地,本实施例中第一算力路由器通过IGP/BGP发送给网络的其它算力路由器的算力路由信息包括算力编码k和第一算力路由器IP地址。
S405、第一算力路由器得到包括FID、算力编码k、IP地址前缀和出端口的RIB项目。
第二算力路由器得到包括FID、算力编码k和第一算力路由器IP地址和出端口的RIB项目。
S406、客户端发送携带FID的首包报文。
例如,首包报文中携带FIDX。
S407、接收到首包报文的第二算力路由器根据FID查FIB,若FIB不存在与该FID对应的项目,则根据FID查RIB,比较“算力编码1…k”的大小,来计算路由,得到与该FID对应的最佳算力路由器(如第一算力路由器)IP地址和出端口,生成基于该FID转发的FIB项目,FIB项目包括FID、第一算力路由器IP地址和出端口的对应关系。
S408、第二算力路由器将所述算力路由器IP地址作为隧道目的地址对所述报文添加隧道封装,或,将所述算力路由器IP地址作为分段路由(SR)的段标识添加SR封装,转发添加隧道封装或添加SR封装的报文至第一算力路由器。
S409、第一算力路由器将第二算力路由器发送的首包报文转发到叶子路由器K。
此处的叶子路由器K是与第一算力路由器连接的叶子路由器。
S410、叶子路由器K转发首包报文到FIDX所在计算容器,并触发““算力代币”转账”。
S411、客户端发送携带FIDX的后续报文至第二算力路由器。
S412、第二算力路由器根据报文携带的FID查FIB,得到与该FID对应的最佳算力路由器IP地址和出端口,将所述算力路由器IP地址作为隧道目的地址对所述报文添加隧道封装,或,将所述算力路由器IP地址作为分段路由(SR)的段标识添加SR封装,转发添加隧道封装或添加SR封装的报文至第一算力路由器。
S413、第一算力路由器将第二算力路由器发送的报文转发到叶子路由器K。
S414、叶子路由器转发报文到FID所在计算容器。
图9为本申请提供的一种算力路由方法实施例的流程图,如图9所示,算力网络中的各个算力路由器,不仅限图9中的第一和第二算力路由器,皆要求基于计算度量值选择路由,以及基于FID转发。本实施例的方法可以包括:
S501与S301相同,此处不再赘述。
S502、叶子路由器1、叶子路由器k把算力路由信息通过IGP/BGP发送到与其连接的一个第一算力路由器。
其中,算力路由信息包括FID和算力编码k,若S501中叶子路由器根据算力编码计算出计算度量值,则算力路由信息包括FID和计算度量值k。若算力路由信息不包括计算度量值,则由接收到算力路由信息的算力路由器根据算力编码计算计算度量值。
S503、第一算力路由器向合约平台或有关部门验证签名为合法,则解锁“算力路由锁”,并触发“算力交易”,得到算力编码k的有效期。
其中,算力编码k的有效期即算力编码k的有效使用时间。
S504、第一算力路由器通过IGP/BGP将包括算力编码k的算力路由信息发送给网络的其它算力路由器。
S505、第一算力路由器得到包括FID、算力编码k和出端口的RIB项目。
第二算力路由器得到包括FID、算力编码k和出端口的RIB项目。
S506、客户端发送携带FID的首包报文。
例如,首包报文中携带FIDx。
S507、接收到首包报文的第二算力路由器根据FID查FIB,若FIB不存在与该FID对应的项目,则根据FID查RIB,比较“算力编码1…k”的大小,来计算路由,得到与该FID对应的出端口,生成基于该FID转发的FIB项目,FIB项目包括FID和出端口的对应关系。
S508、第二算力路由器根据出端口转发首包报文,经过若干类似的算力路由器转发,到达第一算力路由器(算力网络的出口)。
S509、第一算力路由器根据FID查FIB,若FIB不存在与该FID对应的项目,则根据FID查RIB,比较“算力编码1…k”的大小,来计算路由,得到与该FID对应的出端口,生成基于该FID转发的FIB项目,FIB项目包括FID和出端口的对应关系。
S510、第一算力路由器根据出端口转发首包报文到到叶子路由器K。
此处的叶子路由器K是与第一算力路由器连接的叶子路由器。
S511、叶子路由器K转发首包报文到FID所在计算容器,并触发“算力代币”转账。
S512、客户端发送携带FIDX的后续报文至第二算力路由器。
S513、第二算力路由器根据报文携带的FID查FIB,得到与该FID对应的出端口,转发报文到第一算力路由器。
S514、第一算力路由器根据报文携带的FID查FIB,得到与该FID对应的出端口,转发报文到叶子路由器K。
此处的叶子路由器K是与第一算力路由器连接的叶子路由器。
S515、叶子路由器K转发报文到FID所在计算容器。
图10为本申请提供的一种算力路由方法实施例的流程图,如图10所示,本实施 例在图7所示方法的基础上,当首个报文为传输控制协议(Transmission Control Protocol,TCP)报文,如TCP同步(synchronization,SYN)时,本实施例的方法可以包括:
S601与S301相同,此处不再赘述。
S602、叶子路由器1、叶子路由器k把算力路由信息通过IGP/BGP发送到与其连接的一个第一算力路由器。
其中,算力路由信息包括FID、叶子路由器所在计算容器的IP地址前缀和算力编码k,若S101中叶子路由器根据算力编码计算出计算度量值,则算力路由信息包括FID、叶子路由器所在计算容器的IP地址前缀和计算度量值k。若算力路由信息不包括计算度量值,则由接收到算力路由信息的算力路由器根据算力编码计算计算度量值。
S603、第一算力路由器向合约平台或有关部门验证签名为合法,则解锁“算力路由锁”,并触发“算力交易”,得到算力编码k的有效期。
其中,算力编码k的有效期即算力编码k的有效使用时间。
S604、第一算力路由器通过IGP/BGP将算力路由信息发送给网络的其它算力路由器。
具体地,网络的所有算力路由器均得到每一算力编码k与算力编码k对应的IP地址前缀和FID以及算力编码k的有效期,根据这些信息生成RIB,RIB包括算力编码k、IP地址前缀、FID、出端口的对应关系。
S605、第一算力路由器得到包括FID、算力编码k、IP地址前缀和出端口的RIB项目。
第二算力路由器得到包括FID、算力编码k和IP地址前缀和出端口的RIB项目。
S606、客户端发送类型为TCP SYN的首包报文,首包报文中携带FIDx。
S607、接收到首包报文的第二算力路由器根据FID查FIB,若FIB不存在与该FID对应的项目,则根据FID查RIB,比较“算力编码1…k”的大小,来计算路由,得到与该FID对应的目的地IP地址前缀和出端口,生成基于该FID转发的FIB项目,FIB项目包括FID、IP地址前缀和出端口的对应关系。
S608、第二算力路由器将FID添加所得到的与该FID对应的目的地IP地址前缀得到目的地IP地址,将TCP SYN报文根据目的地IP地址转发到第一算力路由器,转发的TCP SYN报文携带目的地IP地址。
其中,该目的地IP地址为FID所在计算容器IP地址。
S609、第一算力路由器将携带目的地IP地址的TCP SYN报文根据目的地IP地址转发到叶子路由器K。
此处的叶子路由器K是与第一算力路由器连接的叶子路由器。
S610、叶子路由器K根据目的地IP地址转发TCP SYN报文到FID所在计算容器。
S611、FIDX所在计算容器向叶子路由器K发送TCP SYN应答报文,应答报文中携带源地址为计算容器的IP地址。
S612、叶子路由器K触发“算力代币”转账。
本实施例中,只有收到TCP SYN ACK报文,才触发算力代币转账,目的是保证只有TCP连接或其它会话连接正确建立起来的情况下才触发算力代币转账。
S613、叶子路由器K向第一算力路由器发送TCP SYN应答报文,应答报文中携带源地址为计算容器的IP地址。
S614、第一算力路由器向第二算力路由器发送TCP SYN应答报文,应答报文中携带源地址为计算容器的IP地址。
S615、第二算力路由器收到TCP SYN应答报文,则将S607生成的FIB项目确定为有效的FIB项目。若未收到TCP SYN应答报文,则删除S607生成的FIB项目,说明TCP会话建立不成功。若TCP会话或用户会话连接建立不成功(通常表明计算站点或相关计算路径有故障),后续来的报文都不能得到算力路由服务。
S616、第二算力路由器向客户端发送TCP SYN应答报文,应答报文中携带源地址为计算容器的IP地址。
S617、客户端进行后续通信时将报文中携带的FID改为所述计算容器的IP地址。
S618、客户端向第二算力路由器发送后续报文,后续报文携带计算容器的IP地址。
本申请涉及的网元包括算力路由器和叶子路由器。图11为本申请提供的一种算力路由器的功能结构示意图,如图11所示,算力路由器包括控制面处理模块11和数据面处理模块12,控制面处理模块11包括路由协议处理单元110、RIB生成与维护单元111、路由选择或计算单元112、FIB生成与维护单元113、签名验证单元114和算力交易触发单元115,数据面处理模块12包括转发处理单元121和多个接口处理单元。
其中,路由协议处理单元110用于功能和/或算力通告,具体为将计算容器的计算资源、和/或可提供的功能以及功能所在计算容器的IP地址前缀/地址,对其它算力路由器进行通告,使算力网络其它算力路由器可以发现功能和/或算力,可选的,通告IP地址前缀,计算资源本申请实施例中以“算力编码”的形式表达。其中,算力路由信息可以包括计算容器的计算资源和/或可提供的功能以及功能所在计算容器的IP地址前缀/地址,可选的,算力路由信息还可以包括叶子路由器IP地址前缀/地址或与叶子路由器所连接的算力路由器的IP地址前缀/地址。这种通告可以通过扩展DC内部协议,或通过扩展路由协议(如IGP/BGP)实现。
签名验证单元114用于与合约平台交互验证“算力编码”的签名的真实性,可选获取算力租用有效期,如果验证签名的真实性通过,则解锁算力路由“锁”,允许“算力编码”传递给其它算力路由器。
算力交易触发单元115用于触发“算力交易”,例如触发“算力代币”转账,例如触发合约平台将“算力代币”转账给算力使用者,因为算力路由做流量引导已经开始启动服务。
RIB生成与维护单元111用于根据算力路由信息生成RIB项目(entry),RIB项目包括F ID、对应的计算度量值、对应的IP地址前缀和出端口;或者,RIB项目包括FID、对应的计算度量值、对应的出口算力路由器IP地址或叶子路由器IP地址和出端口。
路由选择或计算单元112用于根据RIB的“计算度量值”,来选择或计算路由,得到FID对应的目的地IP地址前缀和出端口;或者,得到FID对应的最佳出口算力路由器IP地址或叶子路由器IP地址、出端口和算力编码的有效期,也就是实现站点间负载均衡。
FIB生成与维护单元113用于根据路由选择或计算单元112的处理结果,形成相应的FIB项目(entry),FIB项目包含FID、对应的IP地址前缀和出端口;或者,FIB项目包括FID、对应的出口算力路由器IP地址或叶子路由器IP地址和出端口。
转发处理单元121用于在报文携带有效IP目的地址的情况下,直接做IP转发;或者,用于根据报文携带的FID查FIB得到出端口,从出端口进行转发;或者,用于根据报文携带的FID查FIB得到IP地址前缀,将FID添加该IP地址前缀得到一个目的IP地址,然后根据该目的IP地址进行转发。
接口处理单元用于做接口的输入或输出处理。
图12为本申请提供的一种叶子路由器的功能结构示意图,叶子路由器可以以软件(如虚拟路由器)或硬件设备形态存在,如图12所示,叶子路由器包括控制面处理模块21和数据面处理模块22,控制面处理模块21包括路由协议处理单元210、算力获取单元214和算力交易触发单元215,数据面处理模块22包括转发处理单元221、多个接口处理单元和多个功能处理单元。
其中,路由协议处理单元210用于Function和/或算力通告,具体为由将站点计算资源、和/或可提供的Function以及Function计算容器IP地址前缀/地址,对其它算力路由器进行通告,使算力网络其它算力路由器可以发现Function和/或算力;优选通告IP地址前缀,计算资源本案中以“算力code”的形式表达。其中,站点计算资源、和/或可提供的Function以及Function计算容器IP地址前缀/地址成为算力路由信息,算力路由信息还可以包括叶子路由器IP地址前缀/地址。这种通告可以扩展DC内部协议,或扩展路由协议(如IGP/BGP)实现。
算力获取单元214用于获取经私钥签名过的“算力编码(Code)”和可能存在的可服务的Function ID列表,即做Function发现和计算资源发现;“算力编码(Code)”通常由IT软件提供,“算力编码(Code)”下文还将详细讨论;
算力交易触发单元215用于触发“算力交易”,例如触发“算力代币”转账,例如触发合约平台将“算力代币”转账给算力使用者,因为流量已引导到相应的站点的叶子路由器。
可选的,控制面处理模块21还包括RIB生成与维护单元211、路由选择或计算单元212和FIB生成与维护单元213,其中,RIB生成与维护单元211用于根据算力路由信息生成RIB项目(entry),RIB项目包括Function ID、对应的计算度量值、对应的IP地址前缀、出端口、。
路由选择或计算单元212用于根据RIB的“计算度量值”,按来选择或计算路由,得到Function ID对应的最佳目的地IP地址前缀、虚拟出端口;可选用于做站点或DC内部的计算资源选择,即站点内或服务器负载均衡。
FIB生成与维护单元213用于根据路由选择或计算单元的处理结果,形成相应的FIB项目(entry),FIB项目包含Function ID、对应的IP地址前缀、虚拟出端口。
转发处理单元221用于报文携带有效IP目的地址的情况下,直接做IP转发;或者,用于根据报文携带的Function ID查FIB得到虚拟出端口,从虚拟出端口进行转发给相应的Function处理单元。
接口处理单元用于做接口的输入或输出处理。
功能处理单元用于做功能对应的计算、业务或程序处理,本实施例中假设叶子路由器以虚拟路由器软件形态运行,它与功能之间通过虚拟端口连接。
图13为本申请提供的一种算力路由装置实施例的结构示意图,如图13所示,本实施例的装置可以包括:接收模块31、确定模块32和发送模块33,其中,
接收模块31用于接收至少一个计算容器的算力路由信息,所述算力路由信息包括算力编码或计算度量值;
确定模块32用于根据所述算力路由信息确定路由信息库RIB,所述RIB包括计算容器可提供的功能的功能标识FID或所述功能所在计算容器的IP地址前缀中的一个和所述算力编码或所述计算度量值中的一个,所述功能包括部署在所述计算容器内的软件化的计算程序、功能程序、业务软件或微服务中的任一个;
所述接收模块31还用于接收报文;
发送模块33用于根据路由信息库FIB转发所述报文,所述FIB根据所述RIB确定,所述FIB包括FID或IP地址前缀。
进一步地,发送模块33还用于:
将所述算力路由信息发送给网络中的其它算力路由器。
进一步地,所述算力路由器在所述FIB中未查找到与所述报文对应的FIB项目时,所述确定模块32还用于:
根据所述RIB中的算力编码或所述计算度量值确定与所述报文对应的FIB项目。
进一步地,所述RIB包括FID、计算度量值、IP地址前缀和出端口;或者,
FID、算力编码、IP地址前缀和出端口;或者,
FID、计算度量值、算力路由器IP地址和出端口;或者,
FID、算力编码、算力路由器IP地址和出端口;或者,
IP地址前缀、计算度量值和出端口;或者,
IP地址前缀、算力编码和出端口;或者,
FID、计算度量值和出端口;或者,
FID、算力编码和出端口;
所述FIB包括:FID、IP地址前缀和出端口;或者,
FID、算力路由器IP地址和出端口;或者,
FID和出端口。
进一步地,若所述报文携带FID,所述发送模块用于:
从所述FIB中查找与所述FID对应的IP地址前缀和出端口;
将所述FID添加与所述FID对应的IP地址前缀得到目的地IP地址,根据所述目的地IP地址转发所述报文;或者,
从所述FIB中查找出与所述FID对应的算力路由器IP地址和出端口;
将所述算力路由器IP地址作为隧道目的地址对所述报文添加隧道封装,或,将所述算力路由器IP地址作为分段路由SR的段标识添加SR封装;
转发添加隧道封装或添加SR封装的报文;或者,
从所述FIB查找出与所述FID或IP地址前缀对应的出端口;
从所述出端口转发所述报文。
本实施例的装置,可以用于执行图5所示方法实施例的技术方案,其实现原理和技术效果类似,此处不再赘述。
图14为本申请提供的一种算力路由装置实施例的结构示意图,如图14所示,本实施例的装置在图13所示装置的基础上,算力编码包括计算时延评估值和/或计算容量等级、机器类型或与机器类型相关的加权值α2和服务器负载,所述算力路由信息包括算力编码时,进一步地,还可以包括:计算模块34,该计算模块34用于根据所述算力编码和如下公式计算所述计算度量值:
Figure PCTCN2020115154-appb-000015
或者,
Figure PCTCN2020115154-appb-000016
或者,
计算度量值=γ2×计算时延评估值;
其中,β2、γ2为加权值。
本实施例的装置,可以用于执行图5所示方法实施例的技术方案,其实现原理和技术效果类似,此处不再赘述。
图15为本申请提供的一种算力路由装置实施例的结构示意图,如图15所示,本实施例的装置在图13所示装置结构的基础上,进一步地,还可以包括:处理模块35,处理模块35用于若所述接收模块接收到应答报文,将所述FIB项目确定为有效的FIB项目;
若所述接收模块未接收到所述应答报文,则删除所述FIB项目。
本实施例的装置,可以用于执行图5所示方法实施例的技术方案,其实现原理和技术效果类似,此处不再赘述。
图16为本申请提供的一种算力路由装置实施例的结构示意图,如图16所示,本实施例的装置可以包括:获取模块41、发送模块42和接收模块43,其中,
获取模块41用于获取计算容器的算力编码,所述叶子路由器部署在所述计算容器内;
发送模块42用于将所述计算容器的算力路由信息发送到与所述叶子路由器连接的算力路由器,以使所述算力路由器将所述算力路由信息发送给网络的其它算力路由器,所述算力路由信息包括所述算力编码或计算度量值;
接收模块43用于接收报文;
所述发送模块42还用于转发所述报文。
可选的,获取模块41用于:
从所述计算容器和/或测量工具获取所述算力编码,所述算力编码包括计算时延评估值和/或计算容量等级、机器类型或与机器类型相关的加权值α2和服务器负载,所述计算时延评估值由所述测量工具测量。
本实施例的装置,可以用于执行图6所示方法实施例的技术方案,其实现原理和技术效果类似,此处不再赘述。
图17为本申请提供的一种算力路由装置实施例的结构示意图,如图17所示,本实施例的装置在图16所示装置结构的基础上,若所述算力路由信息包括所述计算度量值,进一步地,还可以包括:计算模块44,计算模块44用于根据所述算力编码和如 下公式计算所述计算度量值:
Figure PCTCN2020115154-appb-000017
或者,
Figure PCTCN2020115154-appb-000018
或者,
计算度量值=γ2×计算时延评估值;
其中,β2、γ2为加权值。
可选的,计算模块44还用于:
根据所述算力编码和如下公式计算算力代币:
算力代币=α1×[β1-(计算度量值)]
其中,α1、β1为算力定价系数。
可选的,算力路由信息还包括:
所述计算容器可提供的功能的功能标识FID、所述功能所在计算容器的IP地址前缀或IP地址,所述功能为部署在所述计算容器内的软件化的计算程序、功能程序、业务软件或微服务。
本实施例的装置,可以用于执行图6所示方法实施例的技术方案,其实现原理和技术效果类似,此处不再赘述。
本申请可以根据上述方法示例对算力路由装置进行功能模块的划分,例如,可以对应各个功能划分各个功能模块,也可以将两个或两个以上的功能集成在一个处理模块中。上述集成的模块既可以采用硬件的形式实现,也可以采用软件功能模块的形式实现。需要说明的是,本申请各实施例中对模块的划分是示意性的,仅仅为一种逻辑功能划分,实际实现时可以有另外的划分方式。
图18为本申请提供的一种装置结构示意图,如图18所示,该装置300包括:存储器301和处理器302;
存储器301,用于存储计算机程序;
处理器302,用于执行存储器存储的计算机程序,以实现上述实施例中的算力路由方法。具体可以参见前述方法实施例中的相关描述。
可选地,存储器301既可以是独立的,也可以跟处理器302集成在一起。
当存储器301是独立于处理器302之外的器件时,装置300还可以包括:
总线303,用于连接存储器301和处理器302。
可选地,本实施例还包括:通信接口304,该通信接口304可以通过总线303与处理器302连接。处理器302可以控制通信接口303来实现装置300的上述的接收和发送的功能。
该装置可以用于执行上述方法实施例中算力路由器或叶子路由器对应的各个步骤和/或流程。
本申请还提供一种可读存储介质,可读存储介质中存储有执行指令,当上述装置的至少一个处理器执行该执行指令时,图18所示的装置执行上述方法实施例中的算力路由方法。
本领域普通技术人员可以理解:在上述实施例中,可以全部或部分地通过软件、硬件、固件或者其任意组合来实现。当使用软件实现时,可以全部或部分地以计算机 程序产品的形式实现。计算机程序产品包括一个或多个计算机指令。在计算机上加载和执行计算机程序指令时,全部或部分地产生按照本申请实施例的流程或功能。计算机可以是通用计算机、专用计算机、计算机网络、或者其他可编程装置。计算机指令可以存储在计算机可读存储介质中,或者从一个计算机可读存储介质向另一个计算机可读存储介质传输,例如,计算机指令可以从一个网站站点、计算机、服务器或数据中心通过有线(例如同轴电缆、光纤、数字用户线(DSL))或无线(例如红外、无线、微波等)方式向另一个网站站点、计算机、服务器或数据中心进行传输。计算机可读存储介质可以是计算机能够存取的任何可用介质或者是包含一个或多个可用介质集成的服务器、数据中心等数据存储设备。可用介质可以是磁性介质,(例如,软盘、硬盘、磁带)、光介质(例如,DVD)、或者半导体介质(例如固态硬盘Solid State Disk(SSD))等。

Claims (38)

  1. 一种算力路由方法,其特征在于,包括:
    算力路由器接收至少一个计算容器的算力路由信息,所述算力路由信息包括算力编码或计算度量值;
    所述算力路由器根据所述算力路由信息确定路由信息库RIB,所述RIB包括计算容器可提供的功能的功能标识FID或所述功能所在计算容器的IP地址前缀中的一个和所述算力编码或所述计算度量值中的一个,所述功能包括部署在所述计算容器内的软件化的计算程序、功能程序、业务软件或微服务中的任一个;
    所述算力路由器接收报文,根据路由信息库FIB转发所述报文,所述FIB根据所述RIB确定,所述FIB包括FID或IP地址前缀。
  2. 根据权利要求1所述的方法,其特征在于,所述方法还包括:
    所述算力路由器将所述算力路由信息发送给网络中的其它算力路由器。
  3. 根据权利要求1或2所述的方法,其特征在于,所述算力路由器在所述FIB中未查找到与所述报文对应的FIB项目时,所述方法还包括:
    所述算力路由器根据所述RIB中的算力编码或所述计算度量值确定与所述报文对应的FIB项目。
  4. 根据权利要求1所述的方法,其特征在于,
    所述RIB包括FID、计算度量值、IP地址前缀和出端口;或者,
    FID、算力编码、IP地址前缀和出端口;或者,
    FID、计算度量值、算力路由器IP地址和出端口;或者,
    FID、算力编码、算力路由器IP地址和出端口;或者,
    IP地址前缀、计算度量值和出端口;或者,
    IP地址前缀、算力编码和出端口;或者,
    FID、计算度量值和出端口;或者,
    FID、算力编码和出端口;
    所述FIB包括:FID、IP地址前缀和出端口;或者,
    FID、算力路由器IP地址和出端口;或者,
    FID和出端口。
  5. 根据权利要求4所述的方法,其特征在于,若所述报文携带FID,所述根据路由信息库FIB转发所述报文,包括:
    所述算力路由器从所述FIB中查找与所述FID对应的IP地址前缀和出端口;
    所述算力路由器将所述FID添加与所述FID对应的IP地址前缀得到目的地IP地址,根据所述目的地IP地址转发所述报文;或者,
    所述算力路由器从所述FIB中查找出与所述FID对应的算力路由器IP地址和出端口;
    所述算力路由器将所述算力路由器IP地址作为隧道目的地址对所述报文添加隧道封装,或,将所述算力路由器IP地址作为分段路由SR的段标识添加SR封装;
    所述算力路由器转发添加隧道封装或添加SR封装的报文;或者,
    所述算力路由器从所述FIB查找出与所述FID或IP地址前缀对应的出端口;
    所述算力路由器从所述出端口转发所述报文。
  6. 根据权利要求1所述的方法,其特征在于,所述算力编码包括计算时延评估值和/或计算容量等级、机器类型或与机器类型相关的加权值α2、服务器负载,所述算力路由信息包括算力编码时,所述方法还包括:
    所述算力路由器根据所述算力编码和如下公式计算所述计算度量值:
    Figure PCTCN2020115154-appb-100001
    或者,
    Figure PCTCN2020115154-appb-100002
    或者,
    计算度量值=γ2×计算时延评估值;
    其中,β2、γ2为加权值。
  7. 根据权利要求3所述的方法,其特征在于,所述方法还包括:
    若所述算力路由器接收到应答报文,则将所述FIB项目确定为有效的FIB项目;
    若所述算力路由器未接收到所述应答报文,则删除所述FIB项目。
  8. 一种算力路由方法,其特征在于,包括:
    叶子路由器获取计算容器的算力编码,所述叶子路由器部署在所述计算容器内;
    所述叶子路由器将所述计算容器的算力路由信息发送到与所述叶子路由器连接的算力路由器,以使所述算力路由器将所述算力路由信息发送给网络的其它算力路由器,所述算力路由信息包括所述算力编码或计算度量值;
    所述叶子路由器接收报文,转发所述报文。
  9. 根据权利要求8所述的方法,其特征在于,所述叶子路由器获取计算容器的算力编码,包括:
    所述叶子路由器从所述计算容器和/或测量工具获取所述算力编码,所述算力编码包括计算时延评估值和/或计算容量等级、机器类型或与机器类型相关的加权值α2和服务器负载,所述计算时延评估值由所述测量工具测量。
  10. 根据权利要求9所述的方法,其特征在于,若所述算力路由信息包括所述计算度量值,所述方法还包括:
    所述叶子路由器根据所述算力编码和如下公式计算所述计算度量值:
    Figure PCTCN2020115154-appb-100003
    或者,
    Figure PCTCN2020115154-appb-100004
    或者,
    计算度量值=γ2×计算时延评估值;
    其中,β2、γ2为加权值。
  11. 根据权利要求10所述的方法,其特征在于,所述方法还包括:
    所述叶子路由器根据所述算力编码和如下公式计算算力代币:
    算力代币=α1×[β1-(计算度量值)]
    其中,α1、β1为算力定价系数。
  12. 根据权利要求8-11任一项所述的方法,其特征在于,所述算力路由信息还包 括:
    所述计算容器可提供的功能的功能标识FID、所述功能所在计算容器的IP地址前缀或IP地址,所述功能为部署在所述计算容器内的软件化的计算程序、功能程序、业务软件或微服务。
  13. 一种算力路由装置,其特征在于,包括:
    接收模块,用于接收至少一个计算容器的算力路由信息,所述算力路由信息包括算力编码或计算度量值;
    确定模块,用于根据所述算力路由信息确定路由信息库RIB,所述RIB包括计算容器可提供的功能的功能标识FID或所述功能所在计算容器的IP地址前缀中的一个和所述算力编码或所述计算度量值中的一个,所述功能包括部署在所述计算容器内的软件化的计算程序、功能程序、业务软件或微服务中的任一个;
    所述接收模块还用于接收报文;
    发送模块,用于根据路由信息库FIB转发所述报文,所述FIB根据所述RIB确定,所述FIB包括FID或IP地址前缀。
  14. 根据权利要求13所述的装置,其特征在于,所述发送模块还用于:
    将所述算力路由信息发送给网络中的其它算力路由器。
  15. 根据权利要求13或14所述的装置,其特征在于,所述算力路由器在所述FIB中未查找到与所述报文对应的FIB项目时,所述确定模块还用于:
    根据所述RIB中的算力编码或所述计算度量值确定与所述报文对应的FIB项目。
  16. 根据权利要求13所述的装置,其特征在于,
    所述RIB包括FID、计算度量值、IP地址前缀和出端口;或者,
    FID、算力编码、IP地址前缀和出端口;或者,
    FID、计算度量值、算力路由器IP地址和出端口;或者,
    FID、算力编码、算力路由器IP地址和出端口;或者,
    IP地址前缀、计算度量值和出端口;或者,
    IP地址前缀、算力编码和出端口;或者,
    FID、计算度量值和出端口;或者,
    FID、算力编码和出端口;
    所述FIB包括:FID、IP地址前缀和出端口;或者,
    FID、算力路由器IP地址和出端口;或者,
    FID和出端口。
  17. 根据权利要求16所述的装置,其特征在于,若所述报文携带FID,所述发送模块用于:
    从所述FIB中查找与所述FID对应的IP地址前缀和出端口;
    将所述FID添加与所述FID对应的IP地址前缀得到目的地IP地址,根据所述目的地IP地址转发所述报文;或者,
    从所述FIB中查找出与所述FID对应的算力路由器IP地址和出端口;
    将所述算力路由器IP地址作为隧道目的地址对所述报文添加隧道封装,或,将所述算力路由器IP地址作为分段路由SR的段标识添加SR封装;
    转发添加隧道封装或添加SR封装的报文;或者,
    从所述FIB查找出与所述FID或IP地址前缀对应的出端口;
    从所述出端口转发所述报文。
  18. 根据权利要求13所述的装置,其特征在于,所述算力编码包括计算时延评估值和/或计算容量等级、机器类型或与机器类型相关的加权值α2和服务器负载,所述算力路由信息包括算力编码时,所述装置还包括:
    计算模块,用于根据所述算力编码和如下公式计算所述计算度量值:
    Figure PCTCN2020115154-appb-100005
    或者,
    Figure PCTCN2020115154-appb-100006
    或者,
    计算度量值=γ2×计算时延评估值;
    其中,β2、γ2为加权值。
  19. 根据权利要求15所述的装置,其特征在于,所述装置还包括:
    处理模块,用于若所述接收模块接收到应答报文,将所述FIB项目确定为有效的FIB项目;
    若所述接收模块未接收到所述应答报文,则删除所述FIB项目。
  20. 一种算力路由装置,其特征在于,包括:
    获取模块,用于获取计算容器的算力编码,所述叶子路由器部署在所述计算容器内;
    发送模块,用于将所述计算容器的算力路由信息发送到与所述叶子路由器连接的算力路由器,以使所述算力路由器将所述算力路由信息发送给网络的其它算力路由器,所述算力路由信息包括所述算力编码或计算度量值;
    接收模块,用于接收报文;
    所述发送模块还用于转发所述报文。
  21. 根据权利要求20所述的装置,其特征在于,所述获取模块用于:
    从所述计算容器和/或测量工具获取所述算力编码,所述算力编码包括计算时延评估值和/或计算容量等级、机器类型或与机器类型相关的加权值α2和服务器负载,所述计算时延评估值由所述测量工具测量。
  22. 根据权利要求21所述的装置,其特征在于,若所述算力路由信息包括所述计算度量值,所述装置还包括:
    计算模块,用于根据所述算力编码和如下公式计算所述计算度量值:
    Figure PCTCN2020115154-appb-100007
    或者,
    Figure PCTCN2020115154-appb-100008
    或者,
    计算度量值=γ2×计算时延评估值;
    其中,β2、γ2为加权值。
  23. 根据权利要求22所述的装置,其特征在于,所述计算模块还用于:
    根据所述算力编码和如下公式计算算力代币:
    算力代币=α1×[β1-(计算度量值)]
    其中,α1、β1为算力定价系数。
  24. 根据权利要求20-23任一项所述的装置,其特征在于,所述算力路由信息还包括:
    所述计算容器可提供的功能的功能标识FID、所述功能所在计算容器的IP地址前缀或IP地址,所述功能为部署在所述计算容器内的软件化的计算程序、功能程序、业务软件或微服务。
  25. 一种装置,其特征在于,包括:
    处理器;以及
    存储器,用于存储所述处理器的可执行指令;
    其中,所述处理器配置为经由执行所述可执行指令来执行权利要求1-7或8-12任一项所述的算力路由方法。
  26. 一种计算机可读存储介质,所述计算机可读存储介质中存储有执行指令,其特征在于,当装置的至少一个处理器执行所述执行指令时,所述装置执行权利要求1-7或8-12任一项所述的算力路由方法。
  27. 一种算力路由方法,其特征在于,包括:
    算力使用者与算力拥有者通过智能合约约定算力租金,所述算力使用者根据所述算力拥有者的计算容器所提供的计算能力,向所述智能合约转账与所述计算能力相应数量的算力代币。
  28. 根据权利要求27所述的方法,其特征在于,所述方法还包括:
    所述算力使用者根据所述算力代币生成算力编码,并使用私钥对所述算力编码签名;
    所述算力使用者向部署在所述计算容器内的叶子路由器发送所述使用私钥签名后的算力编码。
  29. 根据权利要求27所述的方法,其特征在于,所述方法还包括:
    所述算力使用者接收部署在所述计算容器内的叶子路由器发送的算力编码,并使用私钥对所述算力编码签名,其中,所述算力编码由所述叶子路由器根据所述算力代币生成;
    所述算力使用者向所述叶子路由器发送所述使用私钥签名后的算力编码。
  30. 一种算力路由方法,其特征在于,包括:
    部署在计算容器内的叶子路由器获取算力使用者发送的使用私钥签名后的算力编码;
    所述叶子路由器将所述使用私钥签名后的算力编码或由所述使用私钥签名后的算力编码推导获得的计算度量值通过IGP/BGP路由协议发送到算力网络中的第一算力路由器。
  31. 根据权利要求30所述的方法,其特征在于,所述方法还包括:
    所述叶子路由器根据算力代币生成所述算力编码,所述叶子路由器向所述算力使用者发送所述算力编码,其中,所述算力代币的数量与所述计算容器所提供的计算能力相关。
  32. 一种算力路由方法,其特征在于,包括:
    算力网络中的第一算力路由器接收使用私钥签名后的算力编码或由所述使用私钥签名后的算力编码推导获得的计算度量值,所述第一算力路由器使用公钥对所述算力编码或所述计算度量值签名的合法性进行验证;
    所述第一算力路由器将所述算力编码或所述计算度量值的签名是否合法的第一信息通知智能合约。
  33. 根据权利要求32所述的方法,其特征在于,所述方法还包括:
    所述第一算力路由器向所述算力网络中的其它算力路由器发送所述使用私钥签名后的算力编码或由所述使用私钥签名后的算力编码推导获得的所述计算度量值。
  34. 根据权利要求33所述的方法,其特征在于,所述方法还包括:
    所述第一算力路由器和所述算力网络中的其它算力路由器根据所述算力编码或所述计算度量值对应的计算容器提供的计算能力选择计算容器,获得最佳算力路由,以使流量被路由到最高出价的算力使用者进行计算。
  35. 根据权利要求31-34中任一项所述的方法,其特征在于,在第一算力路由器接收使用私钥签名后的算力编码或由所述使用私钥签名后的算力编码推导获得的计算度量值之后,所述方法还包括:
    所述第一算力路由器接收智能合约发送的所述算力编码或所述计算度量值的签名是否合法的第二信息,并根据所述第二信息确定是否引导计算流量。
  36. 一种算力路由方法,其特征在于,包括:
    智能合约从算力网络中的第一算力路由器获取使用私钥签名后的算力编码或由所述使用私钥签名后的算力编码推导获得的计算度量值的签名是否合法的信息;若所述签名合法,所述智能合约在所述第一算力路由器成功引导计算流量后,向算力拥有者发送算力代币,所述算力代币的数量与所述算力拥有者的计算容器所提供的计算能力相关。
  37. 根据权利要求36所述的方法,其特征在于,所述方法还包括:在确认签名合法之前,所述方法还包括:
    所述智能合约获取使用私钥签名后的算力编码或由所述使用私钥签名后的算力编码推导获得的计算度量值,所述智能合约使用公钥对所述算力编码或所述计算度量值的签名的合法性进行验证,并将所述算力编码或所述计算度量值的签名是否合法的第二信息通知所述第一算力路由器和/或所述算力拥有者。
  38. 一种算力路由方法,其特征在于,包括:
    算力拥有者与算力使用者通过智能合约约定算力租金,所述算力拥有者在算力网络中的第一算力路由器成功引导计算流量后,从所述智能合约获得与所述算力拥有者的计算容器所提供的计算能力相应数量的算力代币。
PCT/CN2020/115154 2019-09-12 2020-09-14 算力路由方法及装置 WO2021047678A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
EP20862333.0A EP4024785A4 (en) 2019-09-12 2020-09-14 ROUTING METHOD AND APPARATUS FOR COMPUTING POWER
US17/692,868 US20220200898A1 (en) 2019-09-12 2022-03-11 Compute-aware routing method and apparatus

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201910866871.4 2019-09-12
CN201910866871.4A CN112491716B (zh) 2019-09-12 2019-09-12 算力路由方法及装置

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US17/692,868 Continuation US20220200898A1 (en) 2019-09-12 2022-03-11 Compute-aware routing method and apparatus

Publications (1)

Publication Number Publication Date
WO2021047678A1 true WO2021047678A1 (zh) 2021-03-18

Family

ID=74866135

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/115154 WO2021047678A1 (zh) 2019-09-12 2020-09-14 算力路由方法及装置

Country Status (4)

Country Link
US (1) US20220200898A1 (zh)
EP (1) EP4024785A4 (zh)
CN (2) CN116886601A (zh)
WO (1) WO2021047678A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114710437A (zh) * 2022-06-07 2022-07-05 南京邮电大学 一种结合区块链的物联网边缘网络路由架构
CN116405874A (zh) * 2023-06-06 2023-07-07 湖南省水务规划设计院有限公司 一种水库生态环境实时监控方法

Families Citing this family (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US12008557B2 (en) * 2019-10-03 2024-06-11 Titan IO, Inc. Method, apparatus, and computer-readable medium for routing data services over a decentralized network
CN115225722A (zh) * 2021-04-20 2022-10-21 中兴通讯股份有限公司 算力资源的通告方法及装置、存储介质、电子装置
CN115550234B (zh) * 2021-06-30 2024-04-02 中国移动通信有限公司研究院 一种信息通告方法、控制器及存储介质
CN114978978A (zh) * 2022-06-07 2022-08-30 中国电信股份有限公司 一种算力资源调度方法、装置、电子设备及介质
CN115086230B (zh) * 2022-06-15 2023-06-30 中国联合网络通信集团有限公司 算网路由的确定方法、装置、设备及存储介质
CN117354220A (zh) * 2022-06-28 2024-01-05 中兴通讯股份有限公司 一种算力路由查找方法、装置、算力网络节点及存储介质
CN115514696B (zh) * 2022-09-21 2023-12-08 北京世纪互联宽带数据中心有限公司 一种传递算力资源信息的方法、装置及设备
CN115695523A (zh) * 2022-10-25 2023-02-03 北京星网锐捷网络技术有限公司 一种数据传输控制方法、装置、电子设备及存储介质
CN115412482B (zh) * 2022-11-01 2023-01-10 北京邮电大学 算力路由方法、装置、电子设备和存储介质
CN116074390A (zh) * 2022-12-09 2023-05-05 重庆大学 一种新能源算力网络感知与路由的系统和方法

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103765832A (zh) * 2011-09-01 2014-04-30 华为技术有限公司 用于信息中心网络的通用双模式数据转发平面
WO2016096005A1 (en) * 2014-12-18 2016-06-23 Nokia Solutions And Networks Oy Trusted routing between communication network systems
CN108009811A (zh) * 2017-11-30 2018-05-08 中国人民解放军国防科技大学 一种面向云际计算环境价值交换的跨链通信方法
CN109462652A (zh) * 2018-11-21 2019-03-12 杭州电子科技大学 智能家居系统中基于哈希算法的终端网关负载分配方法
CN110191116A (zh) * 2019-05-24 2019-08-30 北京清红微谷技术开发有限责任公司 恶意节点隔离方法和系统、算力验证终端及p2p网络

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP5410998B2 (ja) * 2007-03-01 2014-02-05 イクストリーム・ネットワークス・インコーポレーテッド スイッチ及びルータのためのソフトウェア制御プレーン
CN107483345B (zh) * 2012-07-11 2020-09-11 华为技术有限公司 业务处理方法、设备及系统
US10341188B2 (en) * 2015-01-27 2019-07-02 Huawei Technologies Co., Ltd. Network virtualization for network infrastructure
CN108768856A (zh) * 2018-05-31 2018-11-06 新华三技术有限公司 一种路由处理方法和装置

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103765832A (zh) * 2011-09-01 2014-04-30 华为技术有限公司 用于信息中心网络的通用双模式数据转发平面
WO2016096005A1 (en) * 2014-12-18 2016-06-23 Nokia Solutions And Networks Oy Trusted routing between communication network systems
CN108009811A (zh) * 2017-11-30 2018-05-08 中国人民解放军国防科技大学 一种面向云际计算环境价值交换的跨链通信方法
CN109462652A (zh) * 2018-11-21 2019-03-12 杭州电子科技大学 智能家居系统中基于哈希算法的终端网关负载分配方法
CN110191116A (zh) * 2019-05-24 2019-08-30 北京清红微谷技术开发有限责任公司 恶意节点隔离方法和系统、算力验证终端及p2p网络

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114710437A (zh) * 2022-06-07 2022-07-05 南京邮电大学 一种结合区块链的物联网边缘网络路由架构
CN116405874A (zh) * 2023-06-06 2023-07-07 湖南省水务规划设计院有限公司 一种水库生态环境实时监控方法
CN116405874B (zh) * 2023-06-06 2023-08-15 湖南省水务规划设计院有限公司 一种水库生态环境实时监控方法

Also Published As

Publication number Publication date
CN112491716A (zh) 2021-03-12
CN116886601A (zh) 2023-10-13
US20220200898A1 (en) 2022-06-23
EP4024785A4 (en) 2022-11-02
CN112491716B (zh) 2023-07-18
EP4024785A1 (en) 2022-07-06

Similar Documents

Publication Publication Date Title
WO2021047678A1 (zh) 算力路由方法及装置
KR102113749B1 (ko) 소스 라우팅에 근거한 별개의 인피니밴드 서브넷들 간에 트래픽을 라우팅하기 위한 시스템 및 방법
US10148560B2 (en) Enhanced error signaling and error handling in a network environment with segment routing
CN112470436A (zh) 使用srv6和bgp的多云连通性
US7447166B1 (en) Method to distribute IEEE 802.1X authenticated users among multiple broadcast domains
KR101097548B1 (ko) 디지털 오브젝트 타이틀 인증
CN107046506B (zh) 一种报文处理方法、流分类器和业务功能实例
US9647876B2 (en) Linked identifiers for multiple domains
EP2869510B1 (en) Express header for packets with hierarchically structured variable-length identifiers
CN103716213A (zh) 在固定接入网中和在用户设备中运行的方法
WO2016029345A1 (zh) 网络流的信息统计方法和装置
US8055897B2 (en) Digital object title and transmission information
CN113395212B (zh) 网络装置及其操作方法和非暂时性计算机可读介质
US20220182243A1 (en) Method and Apparatus for Distributed Ledger
US20180316644A2 (en) System and method for identifying a communication for routing purposes using Internet Protocol addresses that are allocated by and shared amongst Internet service provider networks
US11218918B2 (en) Fast roaming and uniform policy for wireless clients with distributed hashing
US11196666B2 (en) Receiver directed anonymization of identifier flows in identity enabled networks
CN114513450A (zh) 计算路由信息发送方法、装置、设备及存储介质
US8036218B2 (en) Technique for achieving connectivity between telecommunication stations
US10841283B2 (en) Smart sender anonymization in identity enabled networks
US20240214292A1 (en) Multicast tracing in hybrid networks

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20862333

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

ENP Entry into the national phase

Ref document number: 2020862333

Country of ref document: EP

Effective date: 20220331