KR101585775B1 - Context-aware content providing system for qos in mobile cloud system - Google Patents

Abstract

A context-aware content delivery system for service quality in a mobile cloud system is disclosed. The context awareness-based mobile cloud service method of a mobile cloud service server includes a step of collecting information on a service provided to a user from a software defined network agent of each of network entities, A step of performing reinforcement learning based on information about the service, and a step of providing a content transfer service in a mobile cloud environment by reflecting the reinforcement learning result.

Description

TECHNICAL FIELD The present invention relates to a context-based content providing system for a service quality in a mobile cloud system,

The present invention relates to a mobile cloud, and more particularly, to a context-aware content providing system for maintaining service quality in a mobile cloud system.

Recently, content services such as downloading movies, music, and applications through cloud services are increasing.

In particular, the use of cloud services through mobile terminals is increasing explosively.

Therefore, in a context aware mobile cloud, there is a need for a content transmission framework that has high reliability or QoS (Quality of Service).

At this time, mobile cloud provides mobile cloud computing environment with N screen directivity irrespective of device type and OS type, and provides service as a device for reusing terminal resources while providing existing cloud service (DaaS: Device as a service for cloud computing.

Korean Laid-Open Patent Application No. 2013-0070489 (March 31, 2013), "Collaborative Middleware Device and Service Method for Open Mobile Cloud" Korean Patent Laid-Open Publication No. 2014-0022078 (Feb. 21, 2014), "Method and Apparatus for Seamless Mobility Mechanisms in Content-Oriented Network" Korean Published Patent Application No. 2012-0060565 (2012.06.12), "Method for gradual transition to content-centric network and network system based on it" Korean Patent Publication No. 2010-0112742 (Oct. 20, 2010), "Behavior for Enhanced Learning - Infrastructure"

It is an object of embodiments of the present invention to provide a method and system for recognizing a situation for efficient content transmission in a mobile cloud environment.

It is also an object of embodiments of the present invention to provide a content delivery framework that ensures QoS for content provision in a mobile cloud environment.

 The context awareness-based mobile cloud service method of a mobile cloud service server according to an embodiment includes a process of collecting information on services provided to a user from a software defined network agent of each of network entities ; Performing reinforcement learning based on information about the collected services; And providing a content transmission service in a mobile cloud environment by reflecting the reinforcement learning result.

In addition, the mobile cloud service method of the mobile cloud service server according to an exemplary embodiment of the present invention includes receiving an event message for a service provided to a user from a software defined network agent of a network entity ; Generating a new flow table corresponding to the event message and transmitting the new flow table to the network entity; And providing a service corresponding to the event message through the network entity.

In addition, the mobile cloud service method based on the context awareness of the mobile cloud service server according to an exemplary embodiment of the present invention is a method for providing a context aware environment based on environment information including information on performance of a network for mobile cloud service, information on content delivery delay, ; Determining a compensation value and a penalty value by comparing the obtained reference values with the obtained environment information; And generating the context awareness information by reflecting the compensation value and the penalty value.

In this case, the compensation value may include at least one of a network throughput, a content delivery time, an overhead in the network, a data throughput of the user access network of the mobile cloud service, Access network throughput).

According to an embodiment of the present invention, a method and system for recognizing a situation for efficient content transmission in a mobile cloud environment are provided.

According to an embodiment of the present invention, there is also provided a content transmission framework for assuring QoS for content provision in a mobile cloud environment.

Also, according to an embodiment of the present invention, it is possible to provide a new path reducing the control plane message in the node movement using the SDN technique even in a cloud environment not using the CCN, .

1 is an exemplary diagram for explaining a content transfer operation of a content-oriented network.
2 is an exemplary diagram for explaining the concept of the SDN.
3 is a diagram for explaining a context-aware content delivery framework according to an embodiment.
4 is a diagram for explaining a configuration of a context aware mobile cloud service server according to an embodiment.
FIG. 5 is a view for explaining an example of a content transmission scenario in the context-awareness-based content providing system according to an embodiment.
6 is a flowchart illustrating a context aware mobile cloud service method according to an exemplary embodiment of the present invention.
7 is a flowchart illustrating a context aware mobile cloud service method according to another embodiment of the present invention.
8 is a flowchart illustrating a context aware mobile cloud service method according to another embodiment of the present invention.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

In a conventional cloud computing environment, when a mobile node moves, a control plane message for processing content retransmissions and mobile nodes increases in the cloud, and a content request message is delivered to the server again. Accordingly, an embodiment of the present invention proposes a software-defined network (SDN) -based content delivery framework in a mobile cloud to solve such a problem, .

Reinforcement learning learns a given state through trial and error, which agent gets from selecting an action, and determines and executes action according to a set policy. It is a type of machine learning that learns by modifying the behavioral policy based on the reward that is gained by action.

1 is an exemplary diagram for explaining a content transfer operation of a content-oriented network.

Contents Centric Network (CCN) is a part of future Internet research. It aims at solving the message duplication problem that is currently occurring on the Internet through data communication using contents instead of IP (Internet Protocol).

In the CCN, each node communicates through a request packet, Interesting Packet, and a data packet, which is a response message to the request, for data communication.

The content delivered through the data packet is stored in the content store of the CCN router. Accordingly, when an Interesting Packet for content previously stored in the CCN router is delivered, the router can search the content list and deliver the data packet to the requested location when the requested content is retrieved from the content list.

Referring to FIG. 1, (1) represents a case where Client 1 requests content. Referring to (1) of FIG. 1, Client 1 requests content to the network through an Interesting packet (Data Request), and the Interesting packet is delivered to the node A having the requested content.

FIG. 1 (2) shows a process in which a content is transferred from the source node A to the client 1. Referring to (2) of FIG. 1, the router A of the CCN receives the requested content from the server 110 and delivers the content to the client 1 in the order of interest packets.

In FIG. 1 (2), each of the CCN routers A, B, D, and H stores content in its cache every time the content is transmitted.

At this time, when Client 2 requests the same content as Client 1 as shown in (3) of FIG. 1, an Interesting Packet is delivered to the CCN router D. CCN router D

Since the already requested content is stored in its own cache, the data packet is transmitted to Client2 as shown in (1) of FIG. Through this delivery method, CCN can solve the problem of resource duplication that can occur in existing IP network.

2 is an exemplary diagram for explaining the concept of the SDN.

In order to solve the problem that the conventional network is closed, the SDN developed the first Openflow at Stanford University in USA and then developed the SDN which incorporated the software concept into the network based on this. With SDN, networking technology is modeled as computing technology, and network control and management is enabled by software program.

In the SDN, the SDN controller 230, which is the central controller, can manage the entire network and information about all network paths and equipment functions can be located in one application. Accordingly, the SDN controller can calculate a route based on the source 210 and destination 220 addresses, use various network paths for various traffic types, and respond quickly to changes in the network environment.

3 is a diagram for explaining a context-aware content delivery framework according to an embodiment.

The cloud service server providing the cloud service includes a plurality of virtual machines and can include the hypervisor of SDN.

At this time, each of the plurality of virtual machines may include an SDN VM (Virtual Machine), a Content Management Package, a Learning Engine (LE), and a Service Analyzer.

The LE is a content transmission method optimized in consideration of not only the network overhead, the network throughput, the content transmission time, but also the type of service, the performance of the network being used, and the like for content transmission guaranteeing QoS according to the embodiment of the present invention Can be selected.

Information about the services that the user uses is collected in the Service Analyzer through the SDN Agent, and the collected service information can be transferred to the LE.

At this time, information on the service may include, for example, whether the service is provided through CCN or IP-based transmission, the kind of the content service being provided, and the like.

Also, when using TCP / IP transmission, the content service uses the non-content service (SMS, e-mail etc) and CCN contents transmission technique which have better efficiency, , Large-capacity data transmission, etc.).

Referring to FIG. 3, a type of an access network used by a user terminal (UE) for content transmission may include an LTE network, a femtocell (small base station), and a wireless LAN (WiFi).

In FIG. 3, each of the network entities (eNB, Femto Cell AP, Wifi AP, Router, etc.) includes an SDN Agent for communicating with the SDN Controller of the cloud service server.

Each network entity sends a new event message (such as node movement and content request) that is not defined in the SDN Agent's Flow Table to the cloud when it arrives.

This message is delivered to the SDN VM in the cloud. The SDN Controller in the SDN VM defines a new flow table through the Agent Handling function and the Flow Management function, and notifies each SDN Agent to operate it. When a message such as a content delivery arrives, the corresponding content can be searched through the Forwarding Engine in the Content Management Package, and the searched content can be provided.

4 is a diagram for explaining a configuration of a context aware mobile cloud service server according to an embodiment.

Referring to FIG. 4, the context awareness-based mobile cloud service server 400 includes a service analysis unit 410, an enhanced learning performing unit 420, and a control unit 430.

The service analyzer 410 can perform the function of the service analyzer of FIG. For example, the service analysis unit 410 collects information on a service provided to a user from a software defined network agent of each of network entities.

At this time, the service provided to the user may include a content transmission service through a content center network (CCN) and an internet protocol (IP) based network transmission service.

The reinforcement learning performing unit 420 may perform the functions of the learning engine of FIG. For example, the reinforcement learning performing unit 420 performs reinforcement learning based on information about the collected services.

The reinforcement learning performing unit 420 may infer the environment information related to the quality of service (QoS) of the entire network connected to the mobile cloud service based on the collected service information, And then perform reinforcement learning to generate context awareness information.

In this case, the environment information related to the service quality includes at least one of network performance, content transmission time, and service type. The reinforcement learning determines a compensation value and a penalty value by comparing the environment information with a predetermined reference value, The state information may be generated by reflecting the compensation value and the penalty value.

The controller 430 controls the overall operation of the cloud service, such as the Virtual Machine of FIG. 3, and can perform the functions of the SDM VM and Content Management Package.

Accordingly, the control unit 430 may include a SDN VM 431 and a content management unit 433 that performs the functions of the Content Management Package.

For example, the control unit 430 may provide a context-aware content delivery service in a mobile cloud environment by reflecting reinforcement learning results.

For example, the control unit 430 may select a route or a network for content transmission by reflecting the reinforcement learning result. In addition, the controller 430 can determine a content transmission path or a service type for the mobile cloud service based on the context recognition information.

FIG. 5 is a view for explaining an example of a content transmission scenario in the context-awareness-based content providing system according to an embodiment.

The example shown in FIG. 5 shows a situation in which a node UE that has been receiving video streaming service through an eNB 1 that is an LTE base station moves to another base station eNB 2.

At this time, the network entities generate handover events defined in the network environment. For example, eNB 2 can forward an event confirmation message to the SDN VM in the cloud via Router 2 of the CCN.

Due to the transmission characteristics of the CNN, the UE receives the content stored in the content store of the Router 1 that has received the content, not the Router 2.

That is, after the handover is completed using the SDN VM, the content request message is not transmitted to the server in order to retransmit the video streaming service to the UE, but the video streaming service is provided to the UE through the Router 1 having the corresponding content .

Accordingly, according to an embodiment, it is possible to reduce the number of messages related to the control plane occurring at the time of handover, to retransmit the service request message to the server, and to reduce the network load due to the content transmission.

Meanwhile, Reinforcement learning according to an embodiment of the present invention can be applied to select a content transmission method at the time of content transmission.

[Equation 1] represents a state (S), a motion set (A), and a compensation (R) for the behavior defined in the Learning Engine to solve the reinforcement learning problem.

The compensation for the behavior can be determined by the compensation value and the penalty value by each factor.

[Equation 1]

In Equation (1), the meaning of each parameter is as defined in [Table 1].

[Table 1]

In this case, U _{TCP / IP} act, which means the content transmission status of TCP / IP-based and, U _CCN mean, change D _{TCP / IP} is a TCP / IP-based service content delivery status of CCN-based, D _CCN Refers to an action that changes from an IP-based transport to a CCN-based transport, and NOP refers to an action that does not perform any particular action.

And, H _SLA is the current standard value of the network throughput (throughput) of the service, T _SLA is the reference value of time or delay applied to the content transmission, Oth mean threshold of the overhead in the network, and AN _SLA is It means the reference value of the throughput of the access network.

Thus, the parameters with subscripts may be determined experimentally, or may be reference values required in the network. For example, as the network throughput H of the current service is larger than the reference value H _SLA and the delay T is smaller than the reference value T _SLA , the compensation value for the current behavior becomes larger.

Looking at the components that make up the penalty value (P), P _rqst is a penalty for the network throughput, P _dl is the penalty for a content transmission time, P _ovhd is the penalty for a network overhead, P _AN are access network Quot; is the penalty value for the throughput of the < / RTI >

At this time, each element constituting R _srv , C _srv , and penalty value can be calculated in various ways used in the reinforcement learning algorithm.

For example, R _srv and C _srv can be defined as: " (2) "

&Quot; (2) "

In this case, NCS means Non-Content Service, CS means Content Service, and the probability that each content service is served through TCP / IP and CCN, respectively.

는 기준값 또는 임계값을 위반을 제어하기 위한 가중치를 의미한다. Also, for example, each element constituting the penalty value can be defined as in Equation (3)

Means a weight for controlling violation of a reference value or a threshold value.

&Quot; (3) "

와 가장 가까운 Q값을 찾기 위해 Q(s, a)를 갱신할 수 있다. The Learning Engine uses the best policy during reinforcement learning,

Q (s, a) can be updated to find the closest Q value.

A policy for updating Q (s, a) can be defined as: " (4) "

&Quot; (4) "

는 learning rate(예를 들어, 0.3)이고,

는 discount rate(예를 들어, 0.8)을 나타낸다. 그리고, r은 보상값을 나타낸다. At this time,

Is a learning rate (for example, 0.3)

Represents a discount rate (for example, 0.8). And r represents the compensation value.

Table 2 below shows an example of a pseudo code indicating a procedure for updating Q (s, a).

[Table 2]

6 is a flowchart illustrating a context aware mobile cloud service method according to an exemplary embodiment of the present invention.

The method shown in FIG. 6 may be performed by the mobile cloud service server 400 shown in FIG.

In operation 610, the server 400 collects information on a service provided to a user from a software defined network agent of each of the network entities.

In step 620, the server 400 performs reinforcement learning based on the collected service information.

The step of performing reinforcement learning may include estimating environment information related to a quality of service (QoS) of the entire network connected to the mobile cloud service based on the collected service information.

At this time, the environment information may be, for example, each element constituting the penalty value P, and each element constituting the compensation value.

The step of performing the reinforcement learning may include the step of performing reinforcement learning based on the inferred environmental information to generate the context information.

At this time, the context information may be, for example, updated Q (s, a).

In step 630, the server 400 provides a content transfer service in a mobile cloud environment by reflecting the reinforcement learning result.

7 is a flowchart illustrating a context aware mobile cloud service method according to another embodiment of the present invention.

The method shown in Fig. 7 can be performed by the mobile cloud service server 400 shown in Fig.

In operation 710, the server 400 receives an event message for a service provided to the user from a software defined network agent of a network entity.

For example, the server 400 may receive a handover event message as shown in FIG.

In step 720, the server 400 generates a new flow table corresponding to the event message, and transmits the new flow table to the network entity.

In operation 730, the server 400 provides a service corresponding to the event message through the network entity.

At this time, the service corresponding to the event message may be, for example, providing the contents stored in Router 1 to the UE as shown in FIG.

8 is a flowchart illustrating a context aware mobile cloud service method according to another embodiment of the present invention.

The method shown in Fig. 8 can be performed by the mobile cloud service server 400 shown in Fig.

The method shown in FIG. 8 may be a process of performing reinforcement learning.

In step 810, the server 400 obtains environment information including information on performance of the network for mobile cloud service, information on content delivery delay, and information on network throughput.

In step 820, the server 400 compares the obtained reference values with the obtained environment information to determine a compensation value and a penalty value.

At this time, the reference values may be, for example, the parameters described in Equation (1).

In operation 830, the server 400 generates the context aware information by reflecting the compensation value and the penalty value.

In this case, as described in Equation (1), the compensation value includes a network throughput, a content delivery time, an overhead in the network, May be determined in consideration of at least one of the access network throughput of the user access network.

The apparatus described above may be implemented as a hardware component, a software component, and / or a combination of hardware components and software components. For example, the apparatus and components described in the embodiments may be implemented within a computer system, such as, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable array (FPA) A programmable logic unit (PLU), a microprocessor, or any other device capable of executing and responding to instructions. The processing device may execute an operating system (OS) and one or more software applications running on the operating system. The processing device may also access, store, manipulate, process, and generate data in response to execution of the software. For ease of understanding, the processing apparatus may be described as being used singly, but those skilled in the art will recognize that the processing apparatus may have a plurality of processing elements and / As shown in FIG. For example, the processing unit may comprise a plurality of processors or one processor and one controller. Other processing configurations are also possible, such as a parallel processor.

The software may include a computer program, code, instructions, or a combination of one or more of the foregoing, and may be configured to configure the processing device to operate as desired or to process it collectively or collectively Device can be commanded. The software and / or data may be in the form of any type of machine, component, physical device, virtual equipment, computer storage media, or device , Or may be permanently or temporarily embodied in a transmitted signal wave. The software may be distributed over a networked computer system and stored or executed in a distributed manner. The software and data may be stored on one or more computer readable recording media.

The method according to an embodiment may be implemented in the form of a program command that can be executed through various computer means and recorded in a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, and the like, alone or in combination. The program instructions to be recorded on the medium may be those specially designed and configured for the embodiments or may be available to those skilled in the art of computer software. Examples of computer-readable media include magnetic media such as hard disks, floppy disks and magnetic tape; optical media such as CD-ROMs and DVDs; magnetic media such as floppy disks; Magneto-optical media, and hardware devices specifically configured to store and execute program instructions such as ROM, RAM, flash memory, and the like. Examples of program instructions include machine language code such as those produced by a compiler, as well as high-level language code that can be executed by a computer using an interpreter or the like. The hardware devices described above may be configured to operate as one or more software modules to perform the operations of the embodiments, and vice versa.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed exemplary embodiments. For example, it is to be understood that the techniques described may be performed in a different order than the described methods, and / or that components of the described systems, structures, devices, circuits, Lt; / RTI > or equivalents, even if it is replaced or replaced.

Therefore, other implementations, other embodiments, and equivalents to the claims are also within the scope of the following claims.

Claims (12)

Collecting information on a service provided to a user from a software defined network agent of each of the network entities;
Performing reinforcement learning based on information about the collected services; And
And providing a content transmission service in a mobile cloud environment by reflecting the reinforcement learning result,
The process of performing the reinforcement learning
Estimating environment information related to a quality of service (QoS) of the entire network connected to the mobile cloud service based on the collected service information; And
And generating context aware information by performing reinforcement learning based on the inferred environmental information
A mobile cloud service method based on context awareness of a mobile cloud service server. The method according to claim 1,
The service provided to the user
A content transmission service through a content-centric network (CCN), and an internet protocol (IP) -based network transmission service.
A mobile cloud service based on context awareness of mobile cloud service server. delete The method according to claim 1,
Wherein the environment information related to the service quality includes at least one of a network performance, a content transmission time, and a service type,
The reinforcement learning determines a compensation value and a penalty value by comparing the predetermined reference value and the environment information, and generates the context awareness information by reflecting the compensation value and the penalty value
A mobile cloud service based on context awareness of mobile cloud service server. The method according to claim 1,
The process of performing the reinforcement learning
Determining a content transmission path or service type for the mobile cloud service based on the context awareness information
The mobile cloud service method comprising: Receiving an event message for a service provided to a user from a software defined network agent of a network entity;
Generating a new flow table corresponding to the event message and transmitting the new flow table to the network entity; And
And providing a service corresponding to the event message through the network entity
A mobile cloud service based on context awareness of mobile cloud service server. Obtaining environmental information including information on performance of a network for a mobile cloud service, information on content delivery delay, and information on network throughput;
Determining a compensation value and a penalty value by comparing the obtained reference values with the obtained environment information; And
And generating the context recognition information by reflecting the compensation value and the penalty value,
The compensation value includes a network throughput, a content delivery time, an overhead in the network, an access network of the mobile cloud service, throughput) is determined in consideration of at least any one of
A mobile cloud service based on context awareness of mobile cloud service server. A service analyzer for collecting information on a service provided to a user from a software defined network agent of each of network entities;
A reinforcement learning performing unit for performing reinforcement learning based on the collected service information; And
And a controller for providing a content transmission service in a mobile cloud environment by reflecting the reinforcement learning result,
The reinforcement learning performing unit
Based on information about the collected service, environment information related to the quality of service (QoS) of the entire network connected to the mobile cloud service is estimated, and reinforcement learning is performed based on the inferred environment information To generate context-aware information
Situational aware mobile cloud service server. 9. The method of claim 8,
The service provided to the user
A content transmission service through a content-centric network (CCN), and an internet protocol (IP) -based network transmission service.
Situational aware mobile cloud service server. delete 9. The method of claim 8,
Wherein the environment information related to the service quality includes at least one of a network performance, a content transmission time, and a service type,
The reinforcement learning determines a compensation value and a penalty value by comparing the predetermined reference value and the environment information, and generates the context awareness information by reflecting the compensation value and the penalty value
Situational aware mobile cloud service server. 9. The method of claim 8,
The control unit
And determines a content transmission path or a service type for the mobile cloud service based on the context awareness information
Situational aware mobile cloud service server.

Images