WO2022027224A1

WO2022027224A1 - In-network computing power or resource service-oriented communication method

Info

Publication number: WO2022027224A1
Application number: PCT/CN2020/106797
Authority: WO
Inventors: 雷凯; 余锡权; 徐婷; 张梅梅
Original assignee: 北京大学深圳研究生院; 深圳市智慧城市科技发展集团有限公司
Priority date: 2020-08-04
Filing date: 2020-08-04
Publication date: 2022-02-10
Also published as: CN116250219A

Abstract

An in-network resource service-oriented communication method and an in-interwork computing power service-oriented communication method. Taking an in-network resource service as an example, a resource task request packet, a resource task result packet, and a corresponding forwarding strategy are constructed, so as to design a reasonable and efficient communication method and communication device used for resource service request and resource service response in an in-network resource service scenario.

Description

A communication method for in-network computing power or resource services

technical field

The invention relates to a communication method oriented to resource services in a network, in particular to a communication method oriented to computing power services in a network.

Background technique

The core value of the network is to improve efficiency. The telephone network improves the efficiency of human communication, and the Internet improves the efficiency of human collaboration. The basic topology abstraction of the traditional Internet architecture is the end-to-end model: the network is in the middle, the computing is in the periphery, and the host is fully connected to the logical virtual through the network. In the scenario of edge computing or ubiquitous computing, the topology becomes computing embedded in the middle of the network. From the perspective of completing the user's computing tasks, the embedded resources are no longer peer-to-peer, but need to consider differences in distance and network good or bad condition. The widespread deployment of edge computing brings new opportunities and challenges to operator networks. The development of edge computing promotes the integration of network domain computing at the basic resource level.

technical problem

The present invention mainly provides a communication method oriented to in-network computing power service and a communication method oriented to in-network resource service, which will be described in detail below.

technical solutions

According to a first aspect, an embodiment provides a communication method for in-network computing power services, including:

Build and issue a computing power task requirement package, wherein the computing power task requirement package at least includes a name field, the type of computing task, the required computing power and data input fields; the name field of the computing power task requirement package is used to uniquely identify the computing power The task requirement package, the data input field is used to store the data to be calculated;

Receive a computing power task result package, wherein the computing power task result package includes a name field and a calculation result field, the name field of the computing power task result package is used to uniquely identify the computing power task result package, and the calculation result field is used to store the calculation result. Result; the name field of the computing power task result package corresponds to the name field of the computing power task requirement package;

in:

After the computing task demand package is forwarded to any node in the network, when the node receives the computing task demand package, it determines whether it is by extracting the type of computing task and the computing power required in the computing task demand package. Able to complete the computing task requested by the computing power task requirement package; specifically, when the node judges that it has deployed the function corresponding to the type of computing task in the computing power task requirement package, and the function is greater than or equal to the computing task required in the computing power task requirement package. When the computing power is high, it is judged that it can complete the computing task requested by the computing power task demand package; otherwise, it is judged that it cannot be completed;

When it is judged that it can be completed, the node executes the corresponding computing task, and returns the calculation result through the computing power task result packet along the original path where the computing power task demand packet was forwarded; The type of computing task in the package and the required computing power are used as the basis for routing and forwarding, and the computing power task demand package is forwarded to another node.

In one embodiment, the computing power task requirement package further includes signature authentication information, which is used to indicate the authorization status of the sender of the computing power task requirement package; the computing power task result package further includes signature authentication information, which is used to indicate the calculation result. Trusted situation.

According to a second aspect, an embodiment provides a communication method for in-network computing power services, the communication method is based on the communication of a computing power task demand package and a computing power task result package, and the computing power task demand package is used for Request calculation from the node in the network, and the computing power task result package stores the calculation result returned by the node in the network for the computing power task demand package; the communication method includes:

When any node in the network receives the computing power task demand package, the node judges whether it can complete the computing task requested by the computing power task demand package;

If the judgment can be completed, the node executes the corresponding computing task, and returns the computing result through the computing power task result packet along the original path where the computing power task demand packet was forwarded;

If it is judged that it cannot be completed, the node continues to forward the computing power task demand package to other nodes in the network until a node in the network can complete the computing task requested by the computing power task demand package.

In one embodiment, if it is judged that it cannot be completed, the node also marks the computing task requirement package as unsatisfied; when the node receives the corresponding computing task result package returned along the original path, the previous corresponding Records marked as unsatisfied are deleted.

In one embodiment, when the node receives the corresponding computing power task result packet returned along the original path, it judges whether it has enough cache space, and if so, caches the computing power task result packet, otherwise, does not cache force task structure package; or,

The node that executes the computing task requested by the computing power task demand package, after giving the computing power task result package, also judges whether it has enough cache space. If so, it caches the computing power task result package. Adjacent nodes cache computing power task result packets.

In one embodiment, the computing power task requirement package at least includes a name field, the type of computing task and the required computing power, and the name field of the computing power task requirement package is used to uniquely identify the computing power task requirement package; The power task result bag includes a name field and a calculation result field, and the name field of the calculation power task result bag is used to uniquely identify the calculation power task result bag, and the calculation result field is used to store the calculation result; the name of the calculation power task result bag The field corresponds to the name field of the computing power task requirement package.

In one embodiment, when the node in the network receives the computing power task demand package, it determines whether it can complete the computing power task demand package by extracting the type of the computing task and the required computing power in the computing power task demand package. Computing tasks; specifically, when a node in the network judges that it has deployed a function corresponding to the type of computing task in the computing power task demand package, and has a computing power greater than or equal to the computing power required in the computing power task demand package, it judges itself The computing task requested by the computing power task requirement package can be completed, otherwise, it is judged that it cannot be completed.

In one embodiment, when the node judges that the computing task requested by the computing power task requirement package cannot be completed, the node uses the computing task type and required computing power in the computing power task requirement package as the basis for routing and forwarding, and sends the computing power task The demand packet is forwarded to another node.

In one embodiment, the computing power task requirement package further includes a data input field and/or signature authentication information, the data input field of the computing power task requirement package is used to store the data to be calculated, and the signature authentication information is used to indicate the calculation The authorization status of the sender of the power task demand package; the computing power task result package also includes signature authentication information, which is used to indicate that the calculation result is trustworthy.

According to a third aspect, an embodiment provides a communication method for in-network resource services. The communication method communicates based on a resource task request package and a resource task result package, and the resource task request package is used to request an in-network node Resource task, the resource task result package stores the task result returned by the node in the network for the resource task request package, and the communication method includes:

When any node in the network receives the resource task request package, the node determines whether it can complete the resource task requested by the resource task request package;

If the judgment can be completed, the node executes the corresponding resource task, and returns the task result through the resource task result package along the original path where the resource task request package was forwarded;

If it is judged that it cannot be completed, the node continues to forward the resource task request packet to other nodes in the network until a node in the network can complete the resource task requested by the resource task request packet.

In one embodiment, if it is judged that it cannot be completed, the node also marks the resource task request package as unsatisfied; when the node receives the corresponding resource task result package returned along the original path, it marks the previous corresponding package as unsatisfied. Unsatisfied records are deleted.

In one embodiment, when the node receives the corresponding resource task result package returned along the original path, it determines whether it has enough cache space, and if so, caches the resource task result package, otherwise, does not cache the resource task result package; or,

The node that executes the resource task requested by the resource task request package, after giving the resource task result package, also judges whether it has enough cache space. If so, it caches the resource task result package. Resource task result package.

In one embodiment, the resource task request package at least includes a name field, the type of resource task and the required amount of resources, and the name field of the resource task request package is used to uniquely identify the resource task request package; the resource task result package Including a name field and a calculation result field, the name field of the resource task result bag is used to uniquely identify the resource task result bag, and the task result field is used to store the task result; the name field of the resource task result bag and the resource task request bag are used. corresponding to the name field.

In one embodiment, when the node in the network receives the resource task requirement package, it determines whether it can complete the resource task requested by the resource task requirement package by extracting the resource task type and required resource amount in the resource task requirement package; Specifically, when a node in the network judges that it has deployed a function corresponding to the type of the resource task in the resource task demand package, and the amount of resources required in the resource task demand package is greater than or equal to that of the resource task demand package, it judges that it can complete the resource task demand. The resource task requested by the package, otherwise, it is judged that it cannot be completed.

In one embodiment, when the node judges that the resource task requested by the resource task demand package cannot be completed, the node forwards the resource task demand package according to the type of resource task and the required amount of resources in the resource task demand package as the basis for routing and forwarding. to another node.

According to a fourth aspect, an embodiment provides a computer-readable storage medium comprising a program executable by a processor to implement the method described in any of the embodiments herein.

Description of drawings

1 is a schematic diagram of a resource task request package and a resource task result package according to an embodiment;

2 is a flow chart of a communication method oriented to resource services in a network according to an embodiment;

3 is a schematic diagram of forwarding of a computing power task request packet and a computing power task result packet according to an embodiment;

4 is a schematic diagram of a computing power task request package and a computing power task result package according to an embodiment;

5 is a flow chart of a communication method oriented to an in-network computing power service according to an embodiment;

FIG. 6 is a flow chart of a communication method for in-network computing power services according to another embodiment;

7 is a schematic structural diagram of a communication device for routing and computing according to an embodiment;

FIG. 8 is a schematic structural diagram of a communication device for routing and computing according to another embodiment.

Embodiments of the present invention

The present invention will be further described in detail below through specific embodiments in conjunction with the accompanying drawings. Wherein similar elements in different embodiments have used associated similar element numbers. In the following embodiments, many details are described so that the present application can be better understood. However, those skilled in the art will readily recognize that some of the features may be omitted under different circumstances, or may be replaced by other elements, materials, and methods. In some cases, some operations related to the present application are not shown or described in the specification, in order to avoid the core part of the present application from being overwhelmed by excessive description, and for those skilled in the art, these are described in detail. The relevant operations are not necessary, and they can fully understand the relevant operations according to the descriptions in the specification and general technical knowledge in the field.

Additionally, the features, acts, or characteristics described in the specification may be combined in any suitable manner to form various embodiments. At the same time, the steps or actions in the method description can also be exchanged or adjusted in order in a manner obvious to those skilled in the art. Therefore, the various sequences in the specification and drawings are only for the purpose of clearly describing a certain embodiment and are not meant to be a necessary order unless otherwise stated, a certain order must be followed.

The serial numbers themselves, such as "first", "second", etc., for the components herein are only used to distinguish the described objects, and do not have any order or technical meaning. The "connection" and "connection" mentioned in this application, unless otherwise specified, include both direct and indirect connections (connections).

Under the general development trend of cloud computing and edge computing, resources of different scales, such as computing resources, will be distributed at different distances close to users in the future society, and various personalized services will be provided to users through the network. Therefore, the computing power service network needs to coordinately consider the needs of the evolution of network and computing integration, and realize the network-wide optimization of ubiquitous links and computing power resources, flexible scheduling of computing power requirements, and reasonable distribution of computing power tasks. The computing power service network realizes the global optimization of connection and computing power in the network through unified management and coordinated scheduling of multi-dimensional resources such as network, storage and computing power. There can be one or more targets for the computing power service network. For example, to achieve user experience consistency: the network can perceive ubiquitous computing and services, and users do not need to care about the location and deployment status of computing resources in the network. Another example is to realize flexible and dynamic service deployment: according to real-time resource status and user SLA (Service-Level Agreement, service-level agreement) requirements, flexibly schedule services to the optimal edge node.

The above is an example and description using computing power as an example, and other resources such as storage resources and bandwidth resources are similar. As mentioned above, the core value of the network is to improve efficiency. For in-network resource service scenarios, especially in computing power service network scenarios, it is critical to design an efficient communication logic for resource service requests and responses. Therefore, the technical problem to be solved by the present invention is to design a reasonable and efficient communication method and communication device, which are used for resource service request and resource service response in the scenario of in-network resource service.

Before introducing the present invention, some prior art is introduced. Information Center Network (ICN, Information-Center Network) architecture, which uses information-centric communication to replace the existing end-to-end communication, which defines the protocol of the network layer. In this mode of ICN, the importance of data location is downplayed. Compared with physical logic, the focus is on the data content itself; in ICN, users only need to care about any network node to quickly obtain data (or copies), Don't care about the location of the content. An important representative of ICN is the Named Data Network (NDN, Named Data Networking), also known as Content-Centric Networking (CCN). All data in the named data network is identified based on the name. The concept of IP address and communication connection no longer exists in the network, and its network communication is based on two types of data packets, Interest packets (interest packets, in some cases can also be called as request packets) and Data packets (data packets), which are matched by full or relative names. Interest packets are used to request data from the data requesting direction network, and data packets are the data returned by the data owner after receiving the request. NDN is a pull-based architecture, in which users request data by sending Interest packets, and the corresponding data packets return along the same path, achieving good content distribution capabilities through the data's built-in caching function and flexible forwarding strategy layer . The content-oriented network transmission characteristics of NDN enable it to be used for content distribution of massive data.

As mentioned above, there are two types of data packets in the NDN network, one is the Interest packet, the other is the Data packet, and the two are matched according to the complete or relative name. The network communication process is completely driven by the content consumer. The requester sends an Interest packet to the network through one or more locally owned interfaces, names the data as a link, and uses a pull method to obtain the corresponding Data packet. Specifically, the Interest package mainly contains fields such as "Content Name", "Selector", and "Nonce"; the "Content Name" field is used to uniquely identify a content block; the "Selector" field is provided by the content consumer who generates the Interest package , which is mainly used to attach some additional information for forwarding decisions; the "Nonce" field is generated by the content consumer that generates the Interest packet, and is mainly used to prevent loops in the forwarding process of the Interest packet. The Data package mainly includes "Content Name", "Signature", "Signed Info" and "Content"; the "Content Name" field corresponds to the "Content Name" in the Interest package one-to-one; the "Content" field carries a content block specific data. Since each Data packet needs to be signed by the content producer that generates this block to achieve the integrity and reliability of the content, the "Signature" field and the "Signed Info" field are respectively used to carry specific signature content and related Signer information.

When a routing node in an NDN network receives an Interest packet, it will perform the following actions:

(1) Use the "Content Name" field contained in the received Interest packet to retrieve the CS (Content Store, content storage), to find out whether the block corresponding to the Interest package exists in the local CS. If the requested block exists, generate a corresponding Data packet and send the Data packet through the interface to which the received Interest packet arrives, and then discard the Interest packet;

(2) If the requested block does not exist in the local CS, search the PIT table to confirm whether the exact same Interest package as this time has been received before. If the received Interest package is received for the first time, a new PIT (Pending Interest Table) entry containing the content name of the Interest package and the interface that received the Interest package will be created;

(3) Once a new entry is created in the PIT table, it indicates that the received Interest packet will be further forwarded to the location where the requested content is located. In order to determine which interface or interfaces the Interest packet will be forwarded from, the routing node uses the Longest Prefix Matching (LPM) method to retrieve the local FIB (Forwarding Information Base) table. The forwarding strategy selects the forwarding interface and forwards the Interest packet, and then the Interest packet is discarded, thereby ending its processing flow in the routing node;

(4) If a corresponding entry already exists in the PIT table, add the interface to which the Interest packet arrives to the PIT entry and discard the Interest packet;

(5) Each PIT entry has a life cycle. When this life cycle arrives, the corresponding PIT entry will be removed from the PIT table.

When a routing node in an NDN network receives a Data packet, it will perform the following operations:

(1) Use the "Content Name" field contained in the Data package to retrieve the local PIT table, and confirm whether the Data package has a corresponding Interest package stored in the PIT table;

(2) If there is a corresponding PIT entry, forward the Data packet from all interfaces included in the corresponding PIT entry, and delete the PIT entry at the same time;

(3) If there is no corresponding PIT entry, the received Data will be discarded directly;

(4) If the cache replacement policy of the local CS of the routing node decides to cache the Data packet, insert the Data packet into the local CS.

The applicant refers to some ideas of the NDN network, draws lessons from and improves it, and proposes the present invention, and in some cases, some disadvantages that the applicant considers inherent in the study of the NDN network are also eliminated. Applicants believe that NDN networks have the following disadvantages:

(1) The interest package and data package of the NDN network both contain the name field, and the name field contains routing information. However, the data packet does not need to be routed, so there is additional redundant routing information in the data packet.

(2) The interest package and data package of the NDN network can only be used to request content data. It cannot be directly used in the computing power service network to perform computing requests, computing execution and return of computing results for computing tasks.

(3) When the routing node of NDN receives interest, it first checks whether there is corresponding data in the cache through name. This processing method is no longer suitable for in-network resource service scenarios, especially the processing logic of computing tasks and computing results in the computing power service network.

(4) The data package of NDN can only use the path cache. Path caching is a simple approach that requires less computation and communication overhead of placing replicas in the network. Copies of content are cached at intermediate nodes in the path from the content provider to the user. However, due to the continuous accumulation of copies on the path, the content is redundant, which will seriously waste the storage space of the node and reduce the overall hit rate of the cache.

The present invention proposes a communication method oriented to in-network resource services, particularly a communication method oriented to in-network computing power services, which will be described in detail below.

For the communication method of the resource service in the network, the present invention constructs two kinds of data packages, namely, the resource task request package and the resource task result package. The resource task request package is used to request resource tasks from the nodes in the network, and the resource task result package stores the task results returned by the nodes in the network for the resource task request package.

A resource task request package can correspond to one or more resource task result packages. The resource task request packet needs to find a routing path through routing. The resource task result packet does not need to be routed, and is directly returned along the routing path of the resource task request packet.

The communication method is based on the pull (pull) communication model, that is, only when the user or the client application actively sends a resource task request packet, the nodes in the network will execute the resource task and return one or more resource task result packets to the user. .

Referring to FIG. 1, in some embodiments, the resource task request package at least includes a name field, the type of resource task and the required amount of resources; in some embodiments, the resource task request package may also include a data input field and/or signature authentication information . Specifically, the name field of the resource task request package: used to uniquely identify the resource task request package; the name field may contain a timestamp and an identifier of the resource task requester. The type of resource task in the resource task request package: determines the tools required to implement the resource task. The amount of resources required in the resource task request package may be a definite value, indicating the amount of resources required to complete the resource task. The data input field of the resource task request package: used to store the data required to execute the resource task. The signature authentication information of the resource task request package is used to indicate the authorization status of the sender of the resource task request package.

In some embodiments, the resource task result package includes a name field and a calculation result field; in some embodiments, the resource task result package further includes signature authentication information. Specifically, the name field of the resource task result package: used to uniquely identify the resource task result package, and the word field may contain a timestamp and an identifier of the resource task requester. The task result field of the resource task result package: used to store the task result. The name field of the resource task result package corresponds to the name field of the computing power task requirement package.

It can be seen that the resource task request packet contains routing information. The combination of the two fields of the resource task type and the required resource amount is the basis for routing forwarding, while the resource task result packet does not need to contain additional routing information. , so there is no redundant routing information in the resource task result packet like the data packet in the NDN network. In the specific design, the name field of the package (used to identify the protocol package) and the field used for routing and forwarding (such as the above-mentioned resource task type and required resource amount) are separated, so that the resource task result package does not contain Include redundant routing information.

The communication method for in-network resource service of some embodiments, the communication method is based on resource task request package and resource task result package communication, the resource task request package is used to request resource tasks from nodes in the network, and the resource task result package is stored in the network. The task result returned by the inner node for the resource task request package. Please refer to Fig. 2, the communication method includes the following steps:

Step 100: Any node in the network (eg, node A) receives a resource task request packet.

Step 110: When any node in the network (eg, the above-mentioned node A) receives the resource task request package, the node (node A) determines whether it can complete the resource task requested by the resource task request package.

Specifically, when the node in the network receives the resource task requirement package, it judges whether it can complete the resource task requested by the resource task requirement package by extracting the resource task type and required resource amount in the resource task requirement package; specifically , when the node in the network judges that it has deployed the function corresponding to the type of the resource task in the resource task demand package, and the amount of resources required in the resource task demand package is greater than or equal to The requested resource task, otherwise, it is judged that it cannot be completed.

Step 120: If the judgment can be completed, the node (node A) executes the corresponding resource task, and returns the task result through the resource task result packet along the original path where the resource task request packet was forwarded.

Step 130: If it is judged that it cannot be completed, the node (node A) continues to forward the resource task request packet to other nodes in the network until a node in the network can complete the resource task requested by the resource task request packet.

Specifically, when the node judges that the resource task cannot be completed, the node forwards the resource task request package to another node according to the type of the resource task and the required resource amount in the resource task request package as the routing and forwarding basis.

In some examples, when the node judges that the resource task cannot be completed in step 130, the node also marks the resource task request package as unsatisfied; when the node receives the corresponding resource task result package returned along the original path, Records previously marked as unsatisfied will be deleted.

For the returned resource task result package, in some embodiments, the node may adopt a strategy of not caching the resource task result package; in some embodiments, the node may also adopt a strategy of caching the resource task result package. Specifically, it may be further divided into paths (on -path) caching strategy and off-path caching strategy, as detailed below.

In some embodiments, when the node receives the corresponding resource task result package returned along the original path, it determines whether it has enough cache space, and if so, caches the resource task result package, otherwise, does not cache the resource task result. Bag.

In some implementations, the node that executes the resource task requested by the resource task request package, after giving the resource task result package, also judges whether it has enough cache space, and if so, caches the resource task result package, otherwise, then Notifies its neighbors to cache resource task result packets.

The above are some descriptions of the communication methods for in-network resource services. Such resources can be storage resources, bandwidth resources, and computing resources. The following may take computing resources as an example for description.

Aiming at the communication method of the computing power service in the network, the present invention constructs two kinds of data packets, namely the computing power task request package and the computing power task result package. The computing power task demand package is used to request calculation from the nodes in the network, and the computing power task result package stores the calculation results returned by the nodes in the network for the computing power task demand package.

A computing power task request package can correspond to one or more computing power task result packages. When one computing power task request package corresponds to one computing power task result package, the two have an atomic correspondence; when one computing task request package corresponds to multiple computing power task result packages, the two have a composite correspondence. Referring to Figure 3, the computing power task request packet needs to be routed to find a routing path. The computing power task result packet does not need to be routed, and returns directly along the routing path of the computing power task request packet.

The communication method is based on the pull (pull) communication model, that is, only when the user or the client application actively sends a computing task request packet, the nodes in the network will perform the computing task and return one or more computing task results. package to users. The computing power task can typically be, for example, tasks such as face recognition, speech recognition, and image recognition.

Referring to FIG. 4 , in some embodiments, the computing power task request package at least includes a name field, the type of computing power task and the required computing power; in some embodiments, the computing power task request package may also include a data input field and/or Signature authentication information.

Specifically, the name field of the computing power task request package: used to uniquely identify the computing power task request package; the name field may contain a timestamp and an identifier of the computing task requester.

The type of computing task in the computing power task request package: determine what the function to call is. For example, the type of computing task is a picture recognition, then the function that may need to be called is the function photoRecognition() of image recognition. If the computing task is a speech recognition, then the function that may need to be called is the function speechRecognition() of speech recognition.

The computing power required in the computing power task request package can be a definite value, indicating the computing power required to complete the computing task.

The data input field of the computing power task request package: used to store the data required to perform the computing power task. Some computing tasks must have input data, then put the input data into the data input field. For example, if the computing task is image recognition, then the image to be recognized needs to be input into the data input field, and the input data of the subsequent image recognition function photoRecognition() is called from the data input field of the computing power task request package. image to be recognized.

The signature authentication information of the computing power task request package is used to indicate the authorization status of the sender of the computing power task request package. Requesters of network computing must be authorized.

The above are some descriptions of the computing power task request package.

In some embodiments, the computing power task result package includes a name field and a calculation result field; in some embodiments, the computing power task result package further includes signature authentication information.

Specifically, the name field of the computing power task result package: used to uniquely identify the computing power task result package, and the word field may contain a timestamp and the identifier of the computing task requester. The name field of the computing power task result package corresponds to the name field of the computing power task requirement package.

The task result field of the computing power task result package: used to store the calculation results. In some examples, the task result field is variable in length; the calculation results expected by different computing tasks are different, so the length of the corresponding result field is also different, such as face recognition, image recognition and speech recognition. The calculation results expected by the three computing tasks are likely to be different.

The signature authentication information of the computing power task result package: used to indicate that the computing result is trustworthy. Signature authentication information This is to ensure that the calculation result must be trusted.

It can be seen that the computing power task request packet contains routing information. The combination of the two fields of the computing task type and the required computing power is the basis for routing and forwarding, while the computing power task result packet does not need to contain additional information. Routing information, so there is no additional redundant routing information in the computing power task result packet like the data packet in the NDN network. In the specific design, the name field of the packet (used to identify the protocol packet) and the field used for routing and forwarding (such as the type of the above-mentioned computing task and the required computing power) are separated, so that the computing power task result packet contains No longer contains redundant routing information.

A communication method for in-network computing power services according to some embodiments, the communication method is based on the communication of a computing power task demand package and a computing power task result package, the computing power task demand package is used to request calculation from a node in the network, and the computing power task result package is used to request calculation from a node in the network. The calculation results returned by the nodes in the network for the computing power task demand package are stored. Please refer to FIG. 5, the communication method includes the following steps:

Step 200: When any node (eg, node A) in the network receives the computing power task requirement package, the node (node A) determines whether it can complete the computing task requested by the computing power task requirement package.

Specifically, when the node in the network receives the computing power task demand package, it determines whether it can complete the computing task requested by the computing power task demand package by extracting the type of computing task and the required computing power in the computing power task demand package. Specifically, when the node in the network judges that it has deployed the function corresponding to the type of the computing task in the computing power task demand package, and has the computing power greater than or equal to the computing power required in the computing power task demand package, it judges that it can complete the task itself. The computing task requested by the computing power task requirement package, otherwise, it is judged that it cannot be completed.

Step 210 : If the judgment can be completed, the node (node A) executes the corresponding computing task, and returns the computing result through the computing power task result packet along the original path where the computing power task demand packet was forwarded.

Step 220: If it is judged that it cannot be completed, the node (node A) continues to forward the computing power task demand package to other nodes in the network until a node in the network can complete the computing task requested by the computing power task demand package.

Specifically, when the node judges that the computing task requested by the computing power task requirement package cannot be completed, the node uses the computing task type and required computing power in the computing power task requirement package as the basis for routing and forwarding, and sends the computing power task request package forwarded to another node.

In some examples, when the node judges that the computing task cannot be completed in step 220, the node also marks the computing task requirement package as unsatisfied; when the node receives the corresponding computing power task result package returned along the original path , the records previously marked as unsatisfied will be deleted.

For the returned computing power task result package, in some embodiments, the node may adopt a strategy of not caching the resource task result package; in some embodiments, the node may also adopt a strategy of caching the computing power task result package. Specifically, it can be divided into paths The (on-path) caching strategy and the off-path (off-path) caching strategy are described in detail below.

In some embodiments, when the node receives the corresponding computing power task result package returned along the original path, it judges whether it has enough cache space, if so, caches the computing power task result package, otherwise, does not cache the computing power task result package. Force Task Structure Pack.

In some embodiments, the node that executes the computing task requested by the computing power task requirement package, after giving the computing power task result package, also judges whether it has enough cache space, and if so, caches the computing power task result package, On the contrary, it informs its adjacent nodes to cache the computing power task result packet.

Fig. 5 illustrates the communication method oriented to the computing power service in the network from the point of view of the nodes in the network. The following describes the communication method oriented to the computing power service in the network from the perspective of the user.

Referring to FIG. 6 , a communication method for in-network computing power services in some embodiments includes the following steps:

Step 300: Construct and issue a computing power task requirement package. The description of the computing power task requirement package can refer to the description of Figure 4 above. For example, the computing power task requirement package includes at least the name field, the type of computing task, the required computing power and the data input field. The computing power task requirement package can also include Signature authentication information.

Step 310: Receive a computing power task result package. The description of the computing power task result package can refer to the description of FIG. 4 above. For example, the computing power task result package includes a name field and a calculation result field, and the name field of the computing power task result package and the name field of the computing power task requirement package. Correspondingly; the computing power task requirement package may also include signature authentication information.

How the computing power task demand packet sent in step 300 is forwarded and executed in the network, and how step 310 can receive the computing power task result packet forwarded from the node in the network, is described in detail in the method steps of FIG. 5 , For example, after the computing task demand package is forwarded to any node in the network, when the node receives the computing power task demand package, it can determine whether it is able to obtain the computing power by extracting the type of computing task and the required computing power in the computing task demand package. Complete the computing task requested by the computing power task requirement package; specifically, when the node judges that it has deployed the function corresponding to the type of computing task in the computing power task requirement package, and has a function greater than or equal to that required in the computing power task requirement package When the computing power is used, it is judged that it can complete the computing task requested by the computing power task demand package. Otherwise, it is judged that it cannot be completed; when the judgment can be completed, the node executes the corresponding computing task and passes the calculation result through the computing power task result. The packet is returned along the original path where the computing power task demand packet was forwarded; when it is judged that it cannot be completed, the node uses the type of computing task and the required computing power in the computing power task demand packet as the basis for routing and forwarding, and sends the computing power task demand packet. forwarded to another node.

The nodes involved in the communication method for in-network computing power services can be routers or interactive machines, or pure computing nodes, which can be used for routing and computing. Please refer to FIG. 7 , it is advisable to define a node as a communication device for routing and computing. This communication device can be functionally divided into four components: computing resource component 10, functional component 20, storage resource component 30 and functional modules 40. The computing power resource component 10 represents the computing power available to the node. The function component 20 represents the computing function deployed by the node. A node may deploy one or more different computing functions, or none of the computing functions may be deployed. The storage resource component 30 is used for storing calculation results or other data that needs to be stored. The functional module 40 may include functional modules such as service orchestration, resource scheduling, forwarding of computing power requirements, and security authentication.

Please refer to FIG. 8 , which is a schematic structural diagram of a communication device for routing and computing. Routing module: perform computing power demand scheduling and service orchestration. Security authentication module: perform security authentication on computing task requesters and computing results, the requester must be authorized, and the computing results must be trusted. Forwarding module: forwarding computing power requirements and forwarding calculation results. Computing engine: The computing power engine that performs computing tasks, including computing power. Computing function library: Deployed computing functions, possibly deploying multiple functions. Routing/computation table: FIB/CIB table, used for forwarding and execution of computing power task requirements. Cache: Cache computational data, or other data. Physical high-speed switching fabric: perform high-speed packet switching. NIC network card: physical network card.

Through the above communication device for routing and calculation and the communication method for in-network computing service, a computing service network can be constructed. As an open infrastructure, the computing power service network realizes the global optimization of connection and computing power in the network, so that the operator network has the comprehensive service capability of ICT system integration and provides a consistent user experience.

It can be seen from the present invention that both the interest package and the data package of the NDN network contain the name field, and the name field contains routing information. However, the data packet does not need to be routed, so there is additional redundant routing information in the data packet. When the computing power serves the data packets of the network designed by the present invention, the computing power task demand packet needs to be routed and forwarded, so it needs to contain routing information, and the computing power task result packet does not need to be routed and forwarded, and directly follows the computing power task demand packet. The original routing path is returned, so the computing power task result packet no longer contains redundant routing information. The interest package and data package of the NDN network can only be used to request content data. However, the present invention can be implemented for requesting computations, performing computations and returning computation results. When an NDN routing node receives an interest packet, it can only look up whether there is corresponding data in the cache through the name field. In the present invention, it is no longer necessary to search for data in the cache, but the node first judges whether the corresponding function (computing function) can be provided, and then judges whether it can provide sufficient computing power. The data package of NDN can only be cached on-path path. Due to the continuous accumulation of copies on the path, the content is redundant, which will seriously waste the storage space of the node and reduce the overall hit rate of the cache. The calculation results are simultaneously cached on-path and off-path, and some cache optimizations are performed.

Descriptions are made herein with reference to various exemplary embodiments. However, those skilled in the art will recognize that changes and modifications may be made to the exemplary embodiments without departing from the scope of this document. For example, the various operational steps, and the components used to perform the operational steps, may be implemented in different ways depending on the particular application or considering any number of cost functions associated with the operation of the system (eg one or more steps may be deleted, modified or incorporated into other steps).

In the above-mentioned embodiments, it may be implemented in whole or in part by software, hardware, firmware or any combination thereof. Additionally, as understood by those skilled in the art, the principles herein may be reflected in a computer program product on a computer-readable storage medium preloaded with computer-readable program code. Any tangible, non-transitory computer-readable storage medium may be used, including magnetic storage devices (hard disks, floppy disks, etc.), optical storage devices (CD to ROM, DVD, Blu Ray disks, etc.), flash memory, and/or the like . These computer program instructions may be loaded on a general purpose computer, special purpose computer or other programmable data processing apparatus to form a machine such that the instructions executed on the computer or other programmable data processing apparatus may generate means for implementing the specified functions. These computer program instructions may also be stored in a computer-readable memory that instructs a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer-readable memory form a piece of Articles of manufacture, including implementing means for implementing specified functions. Computer program instructions may also be loaded on a computer or other programmable data processing device to perform a series of operational steps on the computer or other programmable device to produce a computer-implemented process such that a process executed on the computer or other programmable device Instructions may provide steps for implementing specified functions.

Although the principles herein have been shown in various embodiments, many modifications may be made in structure, arrangement, proportions, elements, materials and components as are particularly suited to particular environmental and operating requirements without departing from the principles and scope of the present disclosure use. The above modifications and other changes or corrections are intended to be included within the scope of this document.

The foregoing Detailed Description has been described with reference to various embodiments. However, those skilled in the art will recognize that various modifications and changes can be made without departing from the scope of the present disclosure. Accordingly, this disclosure is to be considered in an illustrative rather than a restrictive sense, and all such modifications are intended to be included within its scope. Likewise, the advantages, other advantages, and solutions to problems of the various embodiments have been described above. However, the benefits, advantages, solutions to the problems, and any elements that give rise to these, or make them more explicit, should not be construed as critical, necessary, or essential. As used herein, the term "comprising" and any other variations thereof are non-exclusive inclusion, such that a process, method, article or device including a list of elements includes not only those elements, but also not expressly listed or included in the process , method, system, article or other elements of a device. Furthermore, as used herein, the term "coupled" and any other variations thereof refer to physical connections, electrical connections, magnetic connections, optical connections, communication connections, functional connections, and/or any other connection.

Those skilled in the art will recognize that many changes may be made to the details of the above-described embodiments without departing from the underlying principles of the invention. Accordingly, the scope of the invention should be determined only by the claims.

Claims

A communication method for in-network computing power services, comprising:

Build and issue a computing power task requirement package, wherein the computing power task requirement package at least includes a name field, the type of computing task, the required computing power and data input fields; the name field of the computing power task requirement package is used to uniquely identify the computing power The task requirement package, the data input field is used to store the data to be calculated;

Receive a computing power task result package, wherein the computing power task result package includes a name field and a calculation result field, the name field of the computing power task result package is used to uniquely identify the computing power task result package, and the calculation result field is used to store the calculation result. Result; the name field of the computing power task result package corresponds to the name field of the computing power task requirement package;

in:

After the computing task demand package is forwarded to any node in the network, when the node receives the computing task demand package, it determines whether it is by extracting the type of computing task and the computing power required in the computing task demand package. Able to complete the computing task requested by the computing power task requirement package; specifically, when the node judges that it has deployed the function corresponding to the type of computing task in the computing power task requirement package, and the function is greater than or equal to the computing task required in the computing power task requirement package. When the computing power is high, it is judged that it can complete the computing task requested by the computing power task demand package; otherwise, it is judged that it cannot be completed;

When it is judged that it can be completed, the node executes the corresponding computing task, and returns the calculation result through the computing power task result packet along the original path where the computing power task demand packet was forwarded; The type of computing task in the package and the required computing power are used as the basis for routing and forwarding, and the computing power task demand package is forwarded to another node.
The communication method according to claim 1, wherein the computing power task requirement package further includes signature authentication information, which is used to indicate the authorization status of the sender of the computing power task requirement package; the computing power task result package further includes Signature authentication information, used to indicate that the calculation result is trustworthy.
A communication method for in-network computing power services, characterized in that the communication method is based on the communication of a computing power task demand package and a computing power task result package, and the computing power task demand package is used to request calculation from an in-network node, The computing power task result package stores the calculation results returned by the nodes in the network for the computing power task demand package; the communication method includes:

When any node in the network receives the computing power task demand package, the node judges whether it can complete the computing task requested by the computing power task demand package;

If the judgment can be completed, the node executes the corresponding computing task, and returns the computing result through the computing power task result packet along the original path where the computing power task demand packet was forwarded;

If it is judged that it cannot be completed, the node continues to forward the computing power task demand package to other nodes in the network until a node in the network can complete the computing task requested by the computing power task demand package.
The communication method according to claim 3, wherein if it is judged that it cannot be completed, the node also marks the computing task requirement package as unsatisfied; when the node receives the corresponding computing power returned along the original path When the task result package is completed, the records previously marked as unsatisfied will be deleted.
The communication method according to claim 3 or 4, characterized in that when the node receives the corresponding computing power task result packet returned along the original path, it judges whether it has enough cache space, and if so, caches the computing power. The power task result package, otherwise, the power task structure package is not cached; or,

The node that executes the computing task requested by the computing power task demand package, after giving the computing power task result package, also judges whether it has enough cache space. If so, it caches the computing power task result package. Adjacent nodes cache computing power task result packets.
The communication method according to claim 3, wherein the computing power task requirement package at least includes a name field, the type of computing task and the required computing power, and the name field of the computing power task requirement package is used to uniquely identify A computing power task requirement package; the computing power task result package includes a name field and a calculation result field, the name field of the computing power task result package is used to uniquely identify the computing power task result package, and the calculation result field is used to store the calculation result; The name field of the computing power task result package corresponds to the name field of the computing power task requirement package.
The device communication method according to claim 6, wherein when the node in the network receives the computing power task demand package, it determines whether it is itself by extracting the type of the computing task and the required computing power in the computing power task demand package. Able to complete the computing task requested by the computing power task demand package; specifically, when the node in the network determines that it has deployed a function corresponding to the type of computing task in the computing power task demand package, and has a value greater than or equal to the computing power task demand package When the computing power is required in the computing power, it is judged that it can complete the computing task requested by the computing power task demand package, and vice versa, it is judged that it cannot be completed.
The communication method according to claim 6 or 7, wherein when the node judges that the computing task requested by the computing power task demand package cannot be completed, the node is based on the type and needs of the computing task in the computing power task demand package. The computing power is used as the basis for routing and forwarding, and the computing power task demand packet is forwarded to another node.
The communication method according to claim 6, wherein the computing power task requirement package further includes a data input field and/or signature authentication information, and the data input field of the computing power task requirement package is used to store the data to be calculated. data, and the signature authentication information is used to indicate the authorization status of the sender of the computing power task demand package; the computing power task result package also includes signature authentication information, which is used to indicate that the calculation result is trustworthy.
A communication method oriented to resource services in a network, characterized in that the communication method is based on resource task request packet and resource task result packet communication, the resource task request packet is used to request resource tasks from nodes in the network, and the resource The task result package stores the task result returned by the in-network node for the resource task request package, and the communication method includes:

When any node in the network receives the resource task request package, the node determines whether it can complete the resource task requested by the resource task request package;

If the judgment can be completed, the node executes the corresponding resource task, and returns the task result through the resource task result package along the original path where the resource task request package was forwarded;

If it is judged that it cannot be completed, the node continues to forward the resource task request packet to other nodes in the network until a node in the network can complete the resource task requested by the resource task request packet.
The communication method according to claim 10, wherein if it is judged that it cannot be completed, the node also marks the resource task request package as unsatisfied; when the node receives the corresponding resource task result returned along the original path When the package is included, the records that were previously marked as unsatisfied will be deleted.
The communication method according to claim 10 or 11, wherein when the node receives the corresponding resource task result package returned along the original path, it judges whether it has enough cache space, and if so, caches the resource task result package, otherwise, the resource task result package is not cached; or,

The node that executes the resource task requested by the resource task request package, after giving the resource task result package, also judges whether it has enough cache space. If so, it caches the resource task result package. Resource task result package.
The communication method according to claim 10, wherein the resource task request package at least includes a name field, the type of resource task and the required amount of resources, and the name field of the resource task request package is used to uniquely identify the resource task request package; the resource task result package includes a name field and a calculation result field, the name field of the resource task result package is used to uniquely identify the resource task result package, and the task result field is used to store the task result; the resource task result package The name field of the resource task request packet corresponds to the name field.
The communication method according to claim 13, wherein when the node in the network receives the resource task requirement package, it judges whether it can complete the resource by extracting the resource task type and required resource amount in the resource task requirement package. The resource task requested by the task requirement package; specifically, when a node in the network judges that it has deployed a function corresponding to the type of resource task in the resource task requirement package, and there is a resource greater than or equal to the amount of resources required in the resource task requirement package , then it is judged that it can complete the resource task requested by the resource task demand package; otherwise, it is judged that it cannot be completed.
The communication method according to claim 13 or 14 is characterized in that, when the node judges that the resource task requested by the resource task requirement package cannot be completed, then the node takes the type of the resource task and the required resource amount in the resource task requirement package as the Based on routing and forwarding, the resource task request packet is forwarded to another node.
A computer-readable storage medium, characterized by comprising a program that can be executed by a processor to implement the method according to any one of claims 1 to 15.