KR101146742B1 - METHOD OF DISTRIBUTED SESSION MANAGEMENT IN SaaS AND SESSION MANAGEMENT SYSTEM THEROF - Google Patents

Abstract

The present invention relates to a distributed session management method and a management system that are not tied to a specific web server in the cloud. Selecting a first node in a cloud to which a plurality of nodes are connected using service level agreement information, routing to a first node, and a session identifier included in session request information of a client at the first node Checking the session information to retrieve the session information, and transmitting the retrieved session information to the client can reduce the cost of building the service in the cloud computing environment.

Description

Decentralized session management method and management system of SARS {METHOD OF DISTRIBUTED SESSION MANAGEMENT IN SaaS AND SESSION MANAGEMENT SYSTEM THEROF}

The present invention relates to a session management method and a management system thereof in cloud computing, and more particularly, to a session management method and a management system distributed in software as a service (SaaS), a cloud-based software online rental business. Initiate.

Cloud computing is a form of providing everything in IT as a service. Until now, in order to use the IT environment, each component such as server, storage, SW, and solution was integrated to make a function, and then each solution such as website, mail, ERP, and CRM was developed. Therefore, the development process had to go through the process of introducing SW, integrating and customizing between systems. However, in a cloud computing environment, the IT environment can be constructed and used in the form of using services without the expertise of these individual technologies.

Cloud computing can be divided into three areas: Software as a Service (SaaS), Platform as a Service (PaaS), and Infrastructure as a Service (IaaS). Software as a service (SaaS) refers to providing a software application service as a service. Platform as a Service (PaaS) refers to a form that provides a standardized platform as a service, not SW. Infrastructure as a Service (IaaS) is a form of providing infrastructure, such as servers, storage, and networks as a service.

Recently, session management methods that are efficiently used by host providers are well used in web architectures. However, there is no concept of "pay as much as used" in the web architecture, and the use of resources is not a top priority. Instead, service providers focus on serviceability, reliability and robustness. The session management algorithms currently used in most web servers have been developed based on the hypothesis that resources are available during the life cycle of a session. Although web servers provide a recovery mechanism by copying session information from multiple servers, they do not fully reflect the dynamic nature of the cloud where resources can be added or removed with a single command in the cloud management console.

It enables distributed session management that is not tied to a specific web server in the cloud. In addition, it is possible to immediately supply a service whenever there is a service request, and to return resources used when there is no service request.

In a distributed session management method of Software as a Service (SaaS) according to an embodiment of the present invention, a first node in a cloud to which a plurality of nodes are connected is selected using service level agreement (SLA) information of a client requesting a service. Routing to the first node, retrieving session information by identifying a session identifier included in session request information of the client at the first node, and transmitting the retrieved session information to the client; do.

According to an embodiment of the present invention, a distributed session management system of Software as a Service (SaaS) includes a cloud gateway connecting a client requesting a service to a cloud to which a plurality of nodes are connected, and a session request from a client connected through the gateway. It includes a resource management server that receives information and selects and routes one node in the cloud, and a session management server that manages sessions for each node in the cloud. The session management server periodically updates node usage status information to manage resources. It includes a node management unit for transmitting to the server, and a job processing unit for analyzing the session request information to retrieve the session information, and processing the work required for session processing.

Accordingly, it is possible to reduce service construction costs in a cloud computing environment. You can also effectively use the available resources.

1 is a flowchart of a distributed session management method according to an embodiment of the present invention;
2A illustrates an example of transmitting session information through a first node according to an embodiment of the present invention;
2B is a diagram illustrating session information transmission using a second node and a third node according to an embodiment of the present invention;
3 is a block diagram of a session management server of a distributed session management system according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. The terms used are terms selected in consideration of functions in the embodiments, and the meaning of the terms may vary depending on the intention or precedent of the user or operator. Therefore, the meaning of the terms used in the embodiments to be described later, according to the definition if specifically defined herein, and if there is no specific definition should be interpreted to mean generally recognized by those skilled in the art.

1 is a flowchart of a distributed session management method according to an embodiment of the present invention.

Referring to FIG. 1, first, a first node in a cloud to which a plurality of nodes are connected using service level agreement (SLA) information of a client requesting a service is selected and routed to the first node (110). That is, when the client accesses the SaaS cloud to which a plurality of nodes are connected through the gateway 200, the client analyzes the service request information of the client, selects an optimal node, and routes the corresponding node. Here, the optimal node selects resources such as compute nodes and service instances according to the client's SLA.

This resource type requires resource utilization information that provides information about the processability of each node. For example, if the selected node is currently unable to utilize resources, support for client services will be delayed. Therefore, by periodically updating the resource utilization information of the nodes currently connected in the cloud, it is possible to determine which node can use the resource.

Next, the first node searches for session information by checking a session identifier included in session request information of the client (120). Here, the first node refers to a node having the most suitable resource in response to the service request of the client. In the first node, a session must be established to provide a service to a client. In this case, the first node checks a session identifier among session request information sent by the corresponding client. In operation 130, it is determined whether session information matching the corresponding session identifier exists in the first node.

In the case of a hypertext transfer protocol (HTTP) post request, once a service request is received, each service needs session information. However, the session management method of the present invention is not affected by a specific instance of a service. The session is affected by the compute node corresponding to the session identifier. To locate a session in the cloud, the session identifier serves to indicate the node that holds the session information.

If session information exists in the first node, the session information is extracted and transmitted to the client side (170). Accordingly, a session between the client and the first node is established.

If the first node cannot obtain session information of the corresponding client, the second node 500 having the session information is searched for (140). This is to find the second node 500 having session information necessary to provide a service required by a client among a plurality of connected nodes in the cloud. In this case, the first node searches its own finger table to identify the node having the corresponding session identifier, transmits the session request information of the client to the second node 500, and transmits the session information from the second node 500. Receive 150. In other words, the first node searches for the second node 500 having the session information to obtain session information.

If identification information cannot be found in the finger table of the second node 500 having session information, the third node 400 is searched for the closest third node 400 connected to the second node 500, and the third node 400 is located. The client transmits the session request information of the client (160). That is, when there is no node corresponding to the session identifier included in the session request information in the finger table of the first node, the third node 400 that can transmit the session request information to the second node 500 as a second best option is searched. The session request information is transmitted to the third node 400. In this case, the third node 400 receives the session request information from the first node, searches for its own finger table, and transmits the session request information to the second node 500 which is a node corresponding to the corresponding session identifier.

The second node 500 extracts the session request information and directly transmits the session request information to the first node instead of the third node 400. This is to reduce the response time by directly transmitting the session information instead of the third node 400. The session information transmitted to the first node is transmitted to the client side. Accordingly, a session between the client and the second node 500 is established.

2A illustrates an example of transmitting session information through a first node according to an embodiment of the present invention, and FIG. 2B illustrates a session using a second node 500 and a third node 400 according to an embodiment of the present invention. An example of information transmission.

Referring to FIG. 2A, the cloud includes up to 16 compute nodes. In such chord rings, black circles represent the absence of compute nodes, and white circles represent the compute nodes that can actually be routed. Each compute node maintains a chord ring and maintains a close relationship for resource management regarding processing capacity. This information automatically requests routing and aids in scaling the service. Each compute node is connected to the successor of chord and additionally obtains a finger table for routing session search queries to other nodes.

In FIG. 2A, the resource management server 250 searches for the most suitable node for the service request of the client connected through the cloud gateway 200. The cloud gateway 200 is an access point of SaaS. All services provided by the cloud are accessible through the cloud gateway 200. Because client applications only communicate with the cloud gateway 200, there is no need to change the service binding when multiple instances of a service are executed.

The resource management server has information on resource utilization of all nodes in the cloud. This is to avoid bottlenecks. The resource management server 250 continuously routes to optimal utilization resources for the service request of the client. For example, when node 3 is selected as the first node, the resource management server 250 transmits session request information of the client to the session management server 300 of node 3. The session management server 300 extracts the session information corresponding to the session identifier using its finger cable, and transmits the extracted session information to the resource management server 250. Accordingly, distributed session management in the SaaS cloud can be performed.

On the other hand, if the session management server 300 of the third node that is the first node can not find the session information, the second node 500 having the session information can be searched to transmit the session information. However, if the direct session identifier information about the second node 500 is unknown, the third node 400 that is closest to the second node 500 is searched. When the first node transmits the session request information to the third node 400, the third node 400 may receive the request, search for a node having the corresponding session identifier, and transmit the session request information.

The second node 500 receives service request information from the third node 400 and extracts session information correspondingly. The extracted session information may be transmitted to the third node 400, but is directly transmitted to the first node. This is to reduce the response time. Accordingly, the first node receives the session information from the second node 500 and transmits it to the resource management server 250. As such, a session management method using the second node 500 and the third node 400 is illustrated in FIG. 2B.

3 is a block diagram of a session management server 300 of a distributed session management system according to an embodiment of the present invention.

Referring to FIG. 3, the session management server 300 of the distributed session management system according to an embodiment of the present invention includes a communication unit 310, a node manager 330, a job processor 350, a session manager 370, The service manager 390 is included. The communication unit 310 performs network communication with other nodes in the cloud and also performs data communication with the resource management server 250.

The node manager 330 assists the resource management server 250 by selecting an appropriate service instance according to the SLA of the client. The node manager 330 periodically updates information on resource utilization and transmits the information to the resource management server 250. The node manager 330 monitors a worker thread of the web server to analyze the response time of the node. In order to avoid the situation that too many requests to the same node, the node manager 330 continuously checks the processing capacity of the node, and updates it to the resource management server 250.

The job processor 350 controls the overall functions of the resource management server 250. The task processor 350 controls functions such as key value (session identifier) search, finger table scan, session request information transfer, and the like. In the SaaS cloud, the task processor 350 stores a part of all activated sessions. Resource management server 250 assigns a unique job identifier to the compute node. Sessions are assigned to each compute node corresponding to the job identifier.

The session manager 370 creates a new session when the session does not exist or the validity period of the session expires. All sessions are valid only for a certain period of time, after which they are dropped. If the client application is an interactive application, then session management becomes more important because it requires faster response compared to traditional web applications. In addition, as web applications provide features with the desktop, business objects are demanding session variables that require more memory space and reduced retrieval time.

The service manager 390 keeps track of all of the host services on a single compute node. The service manager 390 determines a service that requires scaling for efficient resource utilization. The service manager 390 analyzes the job processing of the web server, which helps reduce the session retrieval time, and transmits the job processing to the node manager 330.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, Accordingly, the invention should be construed by the description of the claims, which are intended to cover obvious variations that can be derived from the described embodiments.

200: cloud gateway
250: resource management server
300: session management server
310; Communication
330: node management unit
350: work processing unit
370: session management unit
390: service management unit
400: third node
500: second node

Claims (8)

In the session management method of SaaS (Software as a Service),
Selecting a first node in a cloud to which a plurality of nodes are connected by using service level agreement (SLA) information of a client requesting a service, and routing to the first node;
Retrieving session information from the first node by checking a session identifier included in session request information of the client; And
The distributed session management method of Software as a Service (SaaS) comprising the step of transmitting the retrieved session information to the client. The method of claim 1, wherein the distributed session management method comprises:
If the session information is not found at the first node, searching for a second node having the session information using a finger table;
The method of claim 1, further comprising transmitting the session request information to the second node. The method of claim 2, wherein the distributed session management method comprises:
If a second node having the session information is not found, searching for a third node closest to the second node and transmitting the session request information to the third node; And
And searching for the second node in the third node. The distributed session management method of Software as a Service (SaaS). The method of claim 3,
The second node is a distributed session management method of Software as a Service (SaaS) for transmitting the session information to the first node. A cloud gateway connecting a client requesting a service to a cloud to which a plurality of nodes are connected; a resource management server configured to receive session request information from a client connected through the gateway and to select and route one node in the cloud; In the session management system of Software as a Service (SaaS) including a session management server for managing sessions for each node in the cloud,
The session management server,
A node manager for periodically updating the use state information of the node and transmitting the updated state information to the resource management server; And
And a task processing unit for analyzing the session request information, searching for session information, and managing a task required for processing a session. The method of claim 5, wherein the session management server,
A session manager for approving a session with the client or creating a new session; And
And a service management unit for calculating a resource necessary for executing a service requested by the client and transmitting the calculated resource to the resource management server and the node management unit. According to claim 5, The job processing unit,
When the session information is not retrieved from the node, another node in the cloud having the session information is searched, and the session request information is transmitted to the retrieved node. Session Management System. The method of claim 7, wherein the job processing unit,
If the node having the session information is not found, the node that is closest to the node having the session information is searched, and distribution of the Software as a Service (SaaS) is characterized by transmitting the session request information to the neighbor node. Session management system.

Images