KR20170089348A - Method, apparatus, and computer program for managing instant network congestion - Google Patents

Abstract

The present invention relates to a method for operating network which processes an instant congestion connecting request which comprises: step A of confirming whether a received traffic is a predetermined management target traffic by referring to a header value of traffic in a congestion control server configured at the beginning of a network architecture for providing a random service; step B of confirming whether a predetermined simultaneous connection number is exceeded when the traffic is the management target traffic and the traffic is a TCP connection request signal, and holding the traffic when the simultaneous connection number is exceeded; and step C of transmitting the held traffic to the network architecture according to a first-in first-out method at a point of time when the simultaneous connection number is not exceeded, and dropping when a predetermined time is passed from a holding point of time.

Description

[0001] METHOD, APPARATUS, AND COMPUTER PROGRAM FOR MANAGING INSTANT NETWORK CONGESTION [0002]

The present invention relates to a method for controlling instantaneous congested network traffic. More specifically, the present invention can design a congestion control function in front of a certain network architecture so that congestion can be controlled without load of the entire network even when traffic is momentarily transmitted, and traffic without worry of congestion can be processed without delay ≪ / RTI >

Recently, the types and sizes of services provided online have increased exponentially, and traffic congestion has become a frequent problem.

For example, in case of the enrollment application service, when the enrollment application service is opened, traffic may be concentrated on the enrollment application server, and the entire campus network may become inoperable.

Furthermore, when a hot-sale service is opened in an online mall, a holiday ticket sale service is opened, and a ticket sale service of a famous performance is opened, traffic to the web server may be congested and other services may be down .

Conventionally, such instantaneous congestion is generally controlled by a web service level. That is, the conventional instant congestion control method is a method of queuing a web client that allows a user's web client to pass through an access server before accessing the web server, restricts the number of concurrent users on the access server, .

However, there are some inefficient problems with the above method.

The first problem is that the load on network security devices can not be resolved. In general, because network security devices are designed on the edge of the network architecture, the access server is installed behind the security device, and even though the access server controls the congestion, the security devices still have to handle all traffic.

Furthermore, according to the conventional method, even if the traffic is free from the risk of congestion, when the security device in the front side of the network processes all the congestion traffic, a delay time occurs in passing through the security device. Therefore, the quality of the entire service may deteriorate.

The present invention is intended to solve the above problems. More particularly, it is an object of the present invention to provide a method for controlling traffic congestion at a network level. It is another object of the present invention to provide a method and apparatus for efficiently controlling traffic congestion by designing a congestion control function at the front end of a network architecture.

In the network operation method according to the embodiment of the present invention, in the congestion control server configured at the beginning of the network architecture for providing an arbitrary service, whether or not the received traffic is preset management target traffic is referred to the header value of the traffic Step A to confirm; Confirming whether the number of concurrent connections is exceeded when the traffic is the traffic to be managed and the traffic is a TCP connection request signal, and holding the traffic when the number of simultaneous connections is exceeded; And delivering the held traffic to the network architecture according to a first-in first-out method at a time when the number of simultaneous connections is not exceeded, and dropping a preset time from a held time.

Meanwhile, a network system for processing connection requests instantaneously according to an embodiment of the present invention includes a network architecture for providing arbitrary services; And a congestion control server configured at the beginning of the network architecture, wherein the congestion control server checks whether the received traffic is a preset managed traffic with reference to a header value of the traffic, If the traffic is a TCP connection request signal, it is checked whether a preset number of simultaneous connections is exceeded, if the number of simultaneous connections is exceeded, the traffic is held, and if the number of simultaneous connections is exceeded, And delivers the data to the network architecture according to the first-in first-out method, and drops when a preset time elapses from the time of holding.

According to the present invention, the congestion control function is provided at the network level, so that the network resources can be efficiently used, and the quality of other services in which congestion is not a problem can be maintained.

1 is a diagram for explaining a problem of a conventional web service level congestion control method
2 is a diagram for explaining a congestion control method and a network configuration diagram according to an embodiment of the present invention;
3 is a flowchart for explaining a congestion control method according to an embodiment of the present invention

It is to be understood that the present invention is not limited to the description of the embodiments described below, and that various modifications may be made without departing from the technical scope of the present invention. In the following description, well-known functions or constructions are not described in detail since they would obscure the invention in unnecessary detail.

In the drawings, the same components are denoted by the same reference numerals. And in the accompanying drawings, some of the elements may be exaggerated, omitted or schematically illustrated. It is intended to clearly illustrate the gist of the present invention by omitting unnecessary explanations not related to the gist of the present invention. Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a diagram for explaining a problem of a conventional web service level congestion control method. FIG. 1 illustrates a case where traffic of course enrollment in the campus network is busy.

The campus network shown in FIG. 1 may include a plurality of enrollment request servers 150, an email server 160, and a bulletin board server 170. In this case, the connection server for controlling the traffic congestion is generally located in front of the web service server, that is, the course entry server 150, the e-mail server 160 and the bulletin board server 170.

Meanwhile, campus networks are designed to include security devices such as switches, routers and firewalls, IPS, and web firewalls. These security devices are typically deployed at the beginning of the campus network. Thus, as shown in FIG. 1, the connection server 140 is disposed behind the security device such as the firewall 120 and / or the IPS 130.

Therefore, in the network configuration as shown in FIG. 1, the load of the security devices can not be solved. This is because the security devices 120 and 130 must process the traffic 110 instantaneously pushed even if the connection server controls the instantaneous congestion.

Furthermore, in the network configuration shown in FIG. 1, since the security devices 120 and 130 have to process all the traffic 110 entering the campus network, traffic communicating with the email server 160 irrelevant to the enrollment application, Traffic communicating with the server 170 may also be delayed and / or dropped due to the limitations of the security device capacity. That is, if the traffic of the enrollment application is crowded, the service which is not related to the enrollment application may be down, and the quality of all the services of the campus network may be degraded.

Accordingly, the present invention provides a method for controlling traffic congestion at a network level in order to solve the above problems.

2 is a diagram for explaining a congestion control method and a network configuration diagram according to an embodiment of the present invention.

As shown in FIG. 2, the congestion control server 220 according to an embodiment of the present invention can perform a role of buffering a connection request signal located at the front end of the campus network and instantly driven.

On the other hand, all or a part of the network configuration shown in FIG. 2, that is, the congestion control function and the campus network configuration equipment, may be implemented by network function virtualization technology and provided as a series of service chains. In particular, the congestion control function can be implemented with network virtualization technology and designed to pass traffic ahead of other network functions.

In this specification, the congestion control server is assumed to be implemented by software, but the present invention is not limited thereto. That is, the congestion control server according to the embodiment of the present invention can be designed in such a manner that a separate network hardware device that provides a congestion control function is installed in front of the existing campus network.

In particular, the congestion control server 220 according to the embodiment of the present invention is characterized in that it controls only traffic preset by the administrator. In the example of FIG. 2, when the congestion traffic 110 is traffic registration and the general traffic 210 is e-mail or bulletin board traffic, when the traffic is received, the congestion control server refers to a pre-stored table value And passes the normal traffic 210 as it is to the campus network.

The general traffic 210 will then enter the campus network without any additional sanctions.

The method for checking whether the received traffic is the management target traffic can follow the following example. For example, the congestion control server 220 can confirm a predetermined field value among the headers of the traffic to confirm the management target traffic. If the management target traffic is designated by using a field such as a destination IP address or a destination port number among the headers of the traffic, the destination of the received traffic can be identified and it can be confirmed whether there is a possibility of congestion. Therefore, the congestion control server according to the embodiment of the present invention can store a table describing the range of the field value of the header for confirming the congestion target traffic.

On the other hand, the administrator can preset the simultaneous connectable range to the congestion control server 220 in advance. Thereafter, the congestion control server 220 can confirm whether the number of simultaneous connections preset at that time is exceeded when the received traffic is the traffic to be managed and the TCP connection request signal.

How to check the number of concurrent connections is as follows. For example, the congestion control server counts +1 when the TCP connection request signal is passed, and -1 when the TCP connection confirmation signal is passed to check whether the count result at that point exceeds the pre-stored concurrent connectable range have.

If the number of concurrent connections is not exceeded at that point in time, the congestion control server will enter the corresponding TCP connection request packet into the campus network according to first-in-first-out, since there is room for network connection.

However, if the number of concurrent connections is exceeded, the congestion control server can buffer connection requests that are pushed in a moment by holding the corresponding traffic in the queue. If the number of simultaneous connections is not exceeded, the connection request signal held in the queue according to the first-in first-out method can be transmitted to the campus network. However, if the waiting time elapses, the corresponding traffic is dropped.

Meanwhile, according to another embodiment of the present invention, the congestion control server 220 may process a connection request which is instantaneously transmitted in a manner of controlling an HTTP connection signal.

For example, the congestion control server tracks the number of session connections in a manner similar to a load balancer and can hold the signals in a queue when the number of preset sessions is exceeded at the time the HTTP connection request signal is received. And may further transmit a web document notifying the standby state to the source of the traffic.

If the number of session connections does not exceed the predetermined value, the HTTP connection request signal held in the queue may be transmitted to the campus network according to the first-in first-out method. However, when the waiting time elapses, the corresponding traffic may be dropped.

3 is a flowchart illustrating a congestion control method according to an embodiment of the present invention.

According to an embodiment of the present invention, in step 310, a congestion control server may be installed at the beginning of a network for providing a service.

The congestion control server may be located at a front end of a specific network and buffer the connection request signal instantaneously. Such a network configuration may be achieved by installing separate network hardware devices providing congestion control functions in front of the existing network, or applying virtualized congestion control functions implemented in software to general-purpose servers by applying network function virtualization technology to the campus network As shown in FIG.

On the other hand, the network configuration of step 310, i.e., the configuration in which the congestion control function is located in front of a specific network architecture, may be implemented by network function virtualization technology and may be provided as a series of service chains.

In step 320, the congestion control server can receive information on the managed traffic. The information may be provided in a flow rule conforming to the open flow protocol, for example.

More specifically, the information on the traffic to be managed may be provided using a specific field value among the headers of the traffic. For example, the congestion control server can store a table that specifies that a packet value should be processed by checking whether a field value such as a destination IP address or a destination port number of the packet matches a predetermined value.

On the other hand, in step 330, the congestion control server can receive information on the concurrent connectable range. This is for buffering the connection request signal at that point of time when the number of traffic connections at a specific point of time exceeds a predetermined range.

In this situation, when the congestion control server receives the traffic, it first checks whether it is the traffic to be managed. (Step 340)

For example, the congestion control server can check a predetermined field value among the headers of the traffic to check whether the received traffic is the traffic to be managed. For example, the congestion control server can confirm whether the traffic is the management target traffic by checking predetermined field values such as the destination IP address or the destination port number among the headers of the traffic.

On the other hand, in step 340, packets that are classified as general traffic other than managed traffic can be transmitted to the campus network through the congestion control server. (Step 380)

However, in case of the traffic to be managed, the congestion control server can confirm whether or not the number of simultaneous connections at that point in time is exceeded. (Step 345)

More specifically, the congestion control server counts +1 when a TCP connection request signal is passed and -1 when a TCP termination confirmation signal is passed, and it can be checked whether a count result of the time point exceeds a previously stored concurrent connectable range .

If the number of concurrent connections is not exceeded at that point in time, the congestion control server will enter the corresponding TCP connection request packet into the network architecture according to first-in-first-out because there is room for network connection. (Step 380)

However, if the number of concurrent connections is exceeded, the congestion control server can buffer connection requests that are pushed in a moment by holding the corresponding traffic in the queue. If the number of simultaneous connections is not exceeded (step 350), the connection request signal held in the queue according to the first-in first-out method may be transmitted to the campus network. However, if the waiting time elapses (step 360), it is appropriate to drop the corresponding traffic. (Step 370)

The embodiments of the present invention disclosed in the present specification and drawings are intended to be illustrative only and not intended to limit the scope of the present invention. It will be apparent to those skilled in the art that other modifications based on the technical idea of the present invention are possible in addition to the embodiments disclosed herein.

Claims (5)

1. A network operating method for processing a connection request instantaneously,
A step of checking whether a received traffic is a preset managed traffic by referring to a header value of the traffic in a congestion control server configured at the beginning of a network architecture for providing an arbitrary service;
Confirming whether the number of concurrent connections is exceeded when the traffic is the traffic to be managed and the traffic is a TCP connection request signal, and holding the traffic when the number of simultaneous connections is exceeded; And
And delivering the held traffic to the network architecture according to a first-in first-out method at a time when the number of concurrent connections is not exceeded, and dropping a preset time from a holding time. The method according to claim 1,
In the step A,
And forwarding the traffic to the network architecture when the traffic is not the management subject traffic. 3. The method of claim 2,
In the step B,
Counting -1 when a TCP connection request signal is passed, and counting -1 when a TCP termination confirmation signal is passed, thereby checking whether the number of simultaneous connections is exceeded. 3. The method of claim 2,
In the step B,
Confirming whether the number of concurrently connected sessions is exceeded when the traffic is the management target traffic and the traffic is an HTTP request message; And
And when the number of simultaneous connection sessions is exceeded, transmitting the web document holding the traffic and notifying the pre-stored waiting state to the source of the traffic. 1. A network system for processing a connection request instantaneously, comprising:
A network architecture for providing arbitrary services; And
And a congestion control server configured at the beginning of the network architecture,
Wherein the congestion control server checks whether the received traffic is a preset managed traffic with reference to a header value of the traffic and if the traffic is the managed traffic and the traffic is a TCP connection request signal, Wherein the number of simultaneous connections is exceeded, holding the traffic when the number of simultaneous connections is exceeded, delivering the held traffic to the network architecture according to a first-in first-out method at a time when the number of concurrent connections is not exceeded, And drops when a preset time elapses from the time point.

Images