KR100276684B1 - Network resource management system and network resource reservation method to guarantee service level between end users - Google Patents

Abstract

The present invention relates to a network resource management system and a network resource reservation method for guaranteeing a service level between end users.

The present invention maintains network topology and link state information, and stores an agent for monitoring network status periodically and notifying each node of network load information, and load information, container information, transmission path, etc. by the agent. A plurality of packets to be sent to the same destination node, a packet to display the bulletin board, a message to check whether the bulletin board information has a container destined for a destination node, or a space for shipment in a container destined for a destination node; It includes a container for shipping a packet to be collectively transmitted, and a node manager for modifying the bulletin board by determining the transmission path of the packet based on the information informed by the agent.

According to the present invention, it is possible to mass-transmit packets by collectively allocating bandwidth, and the resource reservation procedure is simple and clear, and an appropriate data transmission mechanism can be selected and provided according to an application service.

Description

Network resource management system and network resource reservation method to guarantee service level between end users

The present invention relates to a network resource management system and a network resource reservation method for guaranteeing a service level between end users. In particular, the present invention relates to a network resource reservation method for collectively reserving network resources, and an incoming packet autonomously determines a next node. The present invention relates to a network resource management system and a network resource reservation method capable of guaranteeing a service level between end users in a low overhead and efficient manner.

In a traditional network, transmitted packets are managed identically without any special requirements for application services. However, with the proliferation of Internet users and the application of on-demand video, video conferencing, etc., in which large-scale data has to be transmitted in real time, existing data transmission protocols cannot meet the service requirements of end users. In other words, packets flowing into the intermediate nodes connecting the end user are retransmitted after a certain time elapses or abandon transmission when the data flow on the network increases. In addition, although some operating systems provide services to satisfy the requirements of multimedia and real-time applications, it is difficult to provide a service that satisfies users because it cannot control data flow on a network.

According to such a request, a technique of controlling the flow of data by referring to the size and transmission rate of data to be transmitted between end users in consideration of application requirements has been used. In the network, the data flow on the network can be controlled according to the requirements such as packet transmission rate and packet arrival time required by the application. In this way, guaranteeing the service level between end users is called network quality of service (QoS). .

Protocols that guarantee the level of application services through network resource reservation are RSVP and DiffServ. RSVP (ReSerVation Protocol) is a protocol that reserves network resources based on end user requirements at the time an application starts. When the receiving side generates a request to reserve the necessary resources to receive the desired level of service, the intermediate nodes check the network resources they have and allow the inflow of packets and provide services if there are sufficient resources. Set. This process is possible by expressing values for resources such as bandwidth, maximum data rate, and packet size as parameters. However, while RSVP provides a sophisticated packet admission control mechanism, it is pointed out that the overhead at the intermediate node is too large and too complex to implement and is not scalable when the flow of data increases.

DiffServ is a protocol focused on scalability, proper performance maintenance, and ease of implementation. DiffServ provides a way to bulk transfer packets destined for the same destination at once using aggregated bandwidth, rather than considering each request individually. In other words, when a packet is reserved in advance and a packet must be transmitted to the same destination, the packet is transmitted on this bandwidth which is reserved in advance. By using aggregated bandwidth, there is no need to handle the need for resource reservation of incoming packets separately, thereby reducing the overhead at the intermediate node. However, although DiffServ is more efficient than RSVP, the overhead at the intermediate node still remains and cannot operate autonomously.

Accordingly, the present invention provides a large amount of packets by using the bandwidth that is collectively reserved, and the incoming packets have a computational processing capability to enable autonomous resource reservation, thereby reducing overhead and efficient service between end users. The purpose of the present invention is to provide a network resource management system and a network resource reservation method that can be provided.

The network resource management system for guaranteeing the service level between end users according to the present invention for achieving the above object maintains the network topology and link state information, and periodically monitors the network state to periodically monitor the network load information. Whether there is an agent for notifying a node, a load information, container information, transmission path, etc. by the agent, a bulletin board for displaying it, and a container directed to the destination node among the containers by inquiring the bulletin board information. A packet for checking whether there is space to be shipped in a container destined for the node, a container for loading and collectively transmitting a plurality of packets transmitted to the same destination node, and the information notified by the agent; Determine the transmission path of packets And a node manager for modifying the bulletin board.

In addition, the network resource reservation method for ensuring the service level between end users to achieve the above object is to determine the destination node by the packet entered to the node to search the bulletin board and to determine the container identifier to be shipped packet using the node identifier A first step of inquiring, a second step of determining whether a container of a corresponding class is allocated using the container identifier and a class, and a container assignment to a node manager if no container is allocated as a result of the determination of the second step A third step of requesting a request, a fourth step of determining whether a packet class and a container class are an express class or an ordinary class when there is an assigned container as a result of the determination of the second step, and the determination result packet of the fourth step The class and the container class are determined to be express classes, and the A fifth step of updating the status information of the container and requesting shipment from a node manager when a departure time is earlier than the packet request departure time and the container capacity is larger than the packet request capacity; A sixth step of updating the status information of the container and requesting shipment from the node manager when the container class is determined to be a normal class, the container capacity is larger than the packet request capacity, and the container departure time is earlier than the packet request departure time. And a seventh step of transmitting a packet according to a general packet transmission method when the packet class is a general class as a result of the fourth step determination; and the packet class is a normal class and the container class as a result of the determination in the fourth step. Assigns a container to the node manager if it is a generic class Including an eighth step of requesting is characterized in that is made.

1 is a block diagram of a network resource management system for reserving network resources according to the present invention.

2 is a flowchart illustrating a network resource management method according to the present invention.

3 is a flowchart illustrating a container allocation process of a node manager according to the present invention.

<Explanation of symbols for the main parts of the drawings>

100: agent 110: node manager

120: bulletin board 130a to 130n: first to n-th container

140: packet

The present invention will be described in detail with reference to the accompanying drawings.

1 is a configuration diagram of a network resource management system for guaranteeing a service level between end users according to the present invention. The functions and roles of each member constituting the system are as follows.

The agent 100 maintains network topology and link state information. The agent 100 periodically monitors the network state and periodically informs each node of network load information necessary to determine an efficient data transmission path.

The bulletin board 120 stores load information, container information, a packet transmission path, and the like.

The node manager 110 determines the transmission path of the packet 140 based on the information notified by the agent 100 to modify the bulletin board 120. In addition, the packet 140 flowing into the node queries the bulletin board information in order to check whether there is a container 130 destined for the destination, and if there is a container 130, the space to ship the packet remains.

The container 130 may load a plurality of packets transmitted to the same destination node and transmit them collectively. The container 130 may be classified into several classes according to the capacity to be loaded and the scheduled time interval. Containers that have a small packet loading capacity and are scheduled within a short time are called express class containers. Containers that have a large packet loading capacity and are scheduled at regular time intervals are called regular class containers.

2 is a flowchart illustrating a network resource reservation method for maintaining a service level between end users according to an embodiment of the present invention, and illustrates a process of autonomously operating a packet introduced in FIG. 1. In the network resource reservation method according to the present invention, a packet having an operation processing capability enables resource reservation autonomously, and a plurality of packets transmitted to the same destination node are shipped in one container to enable mass transfer collectively.

Now, the resource reservation method of the network according to the present invention will be described with reference to FIG.

If the packet flows into the node, the bulletin board is searched to determine the next node to which the packet should be transmitted, that is, the destination node (201). When the destination node is determined, the node identifier is used to inquire information about the container in which the packet is to be shipped, that is, the container identifier. Containers are classified into regular containers and express containers according to the packet's shipping capacity and transmission time. The ordinary container is used to transmit a packet in which the size of the data to be transmitted is prior to the transmission time of the data, and the advanced container is used to transmit a packet in which the transmission time of the data is more important. It is determined whether a container of the corresponding class is allocated using the container identifier and the class (202). As a result of the determination, if there is an assigned container, the packet class is checked. First, it is determined whether the class is a priority class having priority. As a result of the determination in step 202, if there is no container assigned, the data transmission path cannot be established and thus the service requested by the user cannot be provided. Therefore, the container manager is requested to the node manager (213).

If it is determined in step 203 that the class is an express class, it is determined whether the container is already assigned to the class class (204). As a result of the determination in step 204, if a container of the express class is allocated, the departure time of the container is checked, and it is determined whether the packet is earlier than or equal to the required departure time (205). The reason for comparing the departure time first is that the express class packet has priority over the capacity of the container.

As a result of the determination in step 204, if the container of the express class is not allocated, the service is not immediately abandoned. Since a packet of an express class has a significant transmission time, even if it is a container of an ordinary class, a judgment is made because the packet can be shipped if all of the user's requirements are satisfied (209).

As a result of comparing the departure time in step 205, if the container is found to depart within the time required by the packet, it is determined whether there is space for shipping the packet. That is, to determine whether the packet can be shipped to the container, it is determined whether the remaining capacity of the container is equal to or larger than the capacity required by the packet (206). If it is determined in step 205 that the container departs later than the time required by the packet, a process of checking the capacity and time of another class is performed (209).

As a result of the determination of step 206, if there is space to ship the packet in the container, the container information, that is, the size of the remaining space is updated (211), and the node manager is requested to ship the packet (212).

As a result of the determination in step 203, it is determined whether the packet is a normal class packet if it is not an express class packet (207). If the packet is determined to be a normal class, it is determined whether a container of a normal class is allocated (208).

As a result of the determination in step 208, if a normal container is allocated, it is determined whether the capacity of the container has a capacity equal to or greater than that required by the packet (209). Normal containers are bandwidth allocated collectively to transmit large amounts of data.

If it is determined in step 209 that the capacity of the container is smaller than the required capacity of the packet, the remaining capacity of the container is updated (211), and the node manager is requested to allocate a new container (212). At this time, the node manager allocates a container of another class according to the class of the packet. The process performed by the node manager is described in detail in FIG. 3.

Next, the departure time of the container and the departure time of the packet are compared (210). As a result, if the departure time of the container is earlier than the departure time of the packet or the request departure time of the packet is the same, the remaining capacity of the container is updated (211), and the node manager is requested to ship the packet (212). As a result of the determination of step 210, if the departure time of the container is later than the departure time of the packet, the node manager requests the node manager to allocate the container (213).

As a result of the determination in step 207, if the packet does not belong to any class of express and normal class, it follows the traditional packet transmission scheme (214).

When the process shown in the above flowchart is finished, all the work to be done by the packet is finished.

FIG. 3 is a detailed flowchart illustrating a process of allocating a container when a node manager requests a resource from a packet. This process is described in detail for the operation of the node manager in FIG. 1.

When a request for resource allocation comes from the packet, information on the container required by the packet, that is, the container class and the size of the container is read (301). Next, the capacity of the remaining container is checked, and it is checked whether the container of the class requested by the packet can be allocated (302). As a result of determination, if there is space for allocating the container, the information on the board is updated to indicate that a new container is allocated to the corresponding destination node, and the container sets information such as the class and departure time of the container to be linked to the board. (303).

As a result of the determination of step 302, if there is no space to allocate the container, another way to handle the user request should be provided. To this end, it is first determined whether the packet adopts a delayed transmission policy (304). The delayed transmission policy is a policy that periodically checks whether the capacity of a container is allowed while keeping a packet at a node for a certain time and ships the packet.

As a result of the determination of step 304, if the delayed transmission policy is used, the timer is set for the next attempt (305), and when the timer setting is completed, the process is performed again from the step 204 of attempting shipment to the container in FIG. As a result of the determination of step 304, if the delayed transmission policy is not used, it is regarded as abandonment of transmission and a service rejection message is transmitted (307).

The request of the packet is a node administrator has reserved a resource for satisfying the user service request by allocating a container, thereby ending the present invention.

The present invention described above is capable of various substitutions, modifications, and changes within the scope without departing from the spirit of the invention for those skilled in the art to which the present invention pertains, the foregoing embodiments and the accompanying drawings. It is not limited to.

As described above, according to the present invention, by collectively reserving network resources to transmit a large amount of packets, and the incoming packets autonomously determine the next node, the service level between end users is guaranteed in a low overhead and efficient way There is. That is, the protocol proposed by the present invention can collectively transmit a packet by allocating bandwidth collectively, and the resource reservation procedure is simpler and clearer than the conventional method (DiffServ). In addition, by providing different classes of containers according to packet requirements, an appropriate data transmission mechanism can be selected and provided according to an application service.

Claims (3)

It maintains network topology and link status information, and periodically monitors network status and informs each node of network load information. A bulletin board for storing load information, container information, transmission path, etc. by the agent and displaying the same; A packet for autonomously checking whether there is a container destined for a destination node among the containers and whether there is a space for shipment in a container destined for a destination node by inquiring the bulletin board information; A container for loading a plurality of packets sent to the same destination node and transmitting them collectively; And a node manager for modifying the bulletin board by determining a transmission path of the packet based on the information notified by the agent. A first step of retrieving a bulletin board, determining a destination node by searching a bulletin board, and using the node identifier to inquire a container identifier to which a packet is shipped; A second step of determining whether a container of a corresponding class is allocated using the container identifier and the class; A third step of requesting a node manager to allocate a container when there is no container allocated as a result of the determination of the second step; A fourth step of determining whether the packet class and the container class are the express class or the normal class when there is an assigned container as a result of the determination of the second step; A fifth step of updating the status information of the container and requesting shipment from the node manager according to the departure time and container capacity of the container class, if it is determined that the packet class and the container class are express classes; A sixth step of updating the state information of the container according to the container capacity and the container departure time and requesting a node manager to ship if it is determined that the packet class and the container class are ordinary classes; A seventh step of transmitting a packet according to a general packet transmission method when the packet class is a general class as a result of the fourth step determination; And in the fourth step, if it is determined that there are no containers that can be shipped, the eighth step of requesting a node manager to allocate a container comprises the eighth step. . The method of claim 2, wherein the node manager, which has been requested to allocate the container, queries the container class and capacity information of the container to determine whether to allocate a container of a class required by the packet; A second step of updating the bulletin board information if there is space to be allocated to the container and setting information such as container class, departure time, etc. in the container and interworking with the bulletin board; A third step of determining whether the packet adopts a delayed transmission policy when there is no space to be allocated to the container as a result of the determination of the first step; A fourth step of setting a timer and retrying shipment to the container if the packet uses the delayed transmission policy as a result of the third step; And a fifth step of transmitting a service rejection message when the packet does not use the delayed transmission policy as a result of the determination of the third step.