KR100782920B1 - Method for packet loss prevention for optimization of wan - Google Patents

Abstract

A packet loss prevention method for WAN(Wide Area Network) optimization is provided to decide failure of a tunnel by checking the number of packets received through a tunnel when traffic is properly distributed by multiplexing the tunnel between two gateways, so that the number of lost packets during the failure is limited to a particular number of packets irrespective of a transmission amount of packets. The number of packets received through a tunnel is checked(ST11). Tunnel failure is sensed on the basis of the number of packets incoming to a virtual tunnel device connected to each tunnel(ST12). If a tunnel is sensed as a failed tunnel exists, packets routed to the corresponding failed tunnel are routed again to a normal tunnel, and copied packets are also transmitted to the failed tunnel(ST13). Queueing delay is prevented by a queue timer(ST14).

Description

Method for Packet Loss Prevention for optimization of WAN

1 is a conceptual diagram of tunneling and VPN in a general WAN section.

2 is a conceptual diagram illustrating a packet division tunnel multiplexing transmission method in a conventional WAN.

3 is a flowchart illustrating a packet split tunnel multiplexing transmission method in a conventional WAN.

4 is a flowchart illustrating a packet loss prevention method for WAN optimization according to an embodiment of the present invention.

5 is a conceptual diagram showing an example of the configuration of a TCP tunnel used in the present invention.

6 is a conceptual diagram illustrating an example of detecting a tunnel failure in FIG. 4.

7 is a conceptual diagram illustrating an example of preventing the tunnel failure in FIG. 4.

8 is a conceptual diagram illustrating an example of preventing a delay due to packet loss by a queue timer when a tunnel fails in FIG. 4.

Explanation of symbols on the main parts of the drawings

A: Gateway A

B: Gateway B

The present invention relates to a WAN (WAN, Wide Area Network, hereinafter referred to as "WAN"), in particular when the number of packets received through the tunnel when the traffic is properly distributed by multiplexing the tunnel between the two gateways in the WAN The present invention relates to a packet loss prevention method for WAN optimization in which the number of packets lost during the tunnel failure is determined by checking the failure of the tunnel so that the number of packets lost during the tunnel failure is limited to a specific number of packets.

In general, a wide area network (WAN) corresponds to a local area network (LAN), and refers to a variety of comprehensive computer networks that are constructed over a wide area. Such a WAN refers to a computer communication network that is formed in a wide area such as a city, a wide area such as a country or a continent, and is a general term for a conventional information communication network. All communication networks established by financial institutions, large corporations or public institutions can be called wide area networks.

LANs connect computers from just a few to a few teens, but in a wide area network (WAN) it connects hundreds and thousands of computers and various computer equipment, so it is necessary to use a host computer capable of high speed processing. Nowadays, wide area network has a form of connecting LAN and LAN to a network capable of high-speed transmission, and devices are installed in various places to reduce the burden on the main computer.

Most of the wide area networks consist of a transmission line and two exchanges. Transmission lines, also called circuits, channels, or relay lines, carry information between devices. The exchange device is a router used to connect two or more transmission lines. When designing a wide area network, an important design issue is how topologies of routers should be interconnected. These types include star, ring, tree, finished, crossed ring and irregular shapes.

Tunnel multiplexing is a technique for improving the quality and performance of multiple circuits for such a WAN. Tunnel multiplexing technology is a technology that forms virtual tunnels on one or more circuits and load balances these tunnels so that they can be used as a single circuit with extended bandwidth. Tunnel multiplexing is an EN-TO-END connection between two gateways. The tunnel multiplexing is applied to the gateway section at the beginning of the tunnel and the gateway section terminating the tunnel, and the VPN can be configured between the networks at both ends.

The most common technologies for implementing tunneling include IP Security Protocol (IPSec), Layer 2 Tunneling Protocol (L2TP), etc. In addition, Generic Routing Encapsulation (GRE) and Point to Point Tunneling Protocol (PPTP). , Layer 2 Forwarding Protocol (L2F), Virtual Tunneling Protocol (VTP), and the like are used. Common encryption technologies used to provide WAN services include authentication, encryption, and encryption.The encryption methods include the Data Encryption Standard (DES) used for IPSec and RC4 used for 3DES, AES, and PPTP. There is this.

1 is a conceptual diagram of two sites having a general WAN section and a VPN tunneled through the gateway.

Where A is gateway A at a particular point and B is gateway B at another particular point tunneled with gateway A. Internal A network and internal B network are connected by VPN.

2 is a conceptual diagram illustrating a packet split tunnel multiplexing transmission method in a conventional WAN, and FIG. 3 is a flowchart illustrating a packet split tunnel multiplexing transmission method in a conventional WAN.

Thus, the transmitter A transmits each packet by dividing each packet into fragments corresponding to the number of tunnels formed by the line (ST1); Receiving unit (Gateway B) recombines the packet divided into the received fragment (ST2);

That is, the conventional packet division tunnel multiplexing transmission technology uses a multi-socket method, which is a bandwidth aggregation technology used in a multilink point to point protocol (PPP), and a single transmission side is used. It is a technique of dividing a packet into pieces as small as the number of tunnels and recombining them at the receiving end.

However, the prior art has a problem that the TCP (Transmission Control Protocol) session is disconnected when the packet is retransmitted because it does not detect the tunnel failure.

Accordingly, the present invention has been proposed to solve the above conventional problems, and an object of the present invention is to multiplex the tunnels between two gateways in order to optimize the WAN so that the number of packets received through the tunnels is properly distributed. The present invention provides a packet loss prevention method for WAN optimization, in which a failure of a tunnel is determined by checking the error, and the number of packets lost in the tunnel failure is limited to a specific number of packets irrespective of the amount of packet transmission.

Packet loss prevention method for WAN optimization according to an embodiment of the present invention to achieve the above object,

Checking a number of packets received through the tunnel; A second step of detecting a tunnel failure based on the number of packets flowing into the virtual tunnel device connected to each tunnel after the first step; If there is a tunnel detected as a failure tunnel in the second step, a third step of rerouting the packet routed to the failure tunnel toward the normal tunnel, and transmitting the copied packet toward the failure tunnel; It is characterized by the configuration.

Hereinafter, an embodiment according to the present invention as described above, the technical spirit of the packet loss prevention method for WAN optimization with reference to the drawings as follows.

4 is a flowchart illustrating a packet loss prevention method for WAN optimization according to an embodiment of the present invention.

As shown therein, a first step ST11 of checking the number of packets received through the tunnel; A second step (ST12) of detecting a tunnel failure based on the number of packets flowing into the virtual tunnel device connected to each tunnel after the first step; If there is a tunnel detected as a failure tunnel in the second step, a third step (ST13) of rerouting the packet routed to the failure tunnel to the normal tunnel and transmitting the copied packet to the failure tunnel; It is characterized by performing.

In the second step, the criterion for determining whether a failure occurs in one tunnel among the tunnels constituting the multiplexing group is set such that the packet inflow does not occur toward the failed tunnel while the packet inflow increases to a normal tunnel. It is done.

In the third step, the copied packet is characterized by indicating that the copied packet using a specific bit of the header portion.

The third step may be characterized in that the receiver automatically discards the copied packet.

The packet loss prevention method for WAN optimization may further include a fourth step (ST14) of preventing a queuing delay with a queue timer after the third step.

In the fourth step, the dequeue packet is forcibly released within a predetermined time from a queuing time point using the queuing timer.

A preferred embodiment of the packet loss prevention method for WAN optimization according to the present invention configured as described above will be described in detail with reference to the accompanying drawings. In the following description of the present invention, detailed descriptions of well-known functions or configurations will be omitted if it is determined that the detailed description of the present invention may unnecessarily obscure the subject matter of the present invention. In addition, terms to be described below are terms defined in consideration of functions in the present invention, which may vary according to intention or precedent of a user or an operator, and thus, the meaning of each term should be interpreted based on the contents throughout the present specification. will be.

First, the present invention is for optimization of the WAN, and can be applied to all related technologies such as multi-link (PP) as well as VPN. Hereinafter, the present invention will be described based on GRE tunneling.

According to the present invention, when the traffic is properly distributed by multiplexing the tunnel between two gateways to determine the WAN, the number of packets received through the tunnel is determined by checking the number of packets received through the tunnel. Is intended to be limited to a certain number of packets regardless.

If some tunnels among the tunnels forming the tunnel multiplexing group fail, the faulty tunnel of the multiplexing group is removed and packets are not routed to the failed tunnel. In this case, since the failure detection of the tunnel to be performed is inspected at a specific time period, a section in which the failure is not detected in the inspection period occurs, resulting in packet loss. In addition, the amount of lost packets may be small or large depending on the packet transmission rate. This loss of packets results in poorer performance than other protocols, especially for TCP applications. Packet loss in some tunnels not only results in retransmission of packets, but can also be a difficult factor in maintaining TCP sessions.

5 is a conceptual diagram showing an example of the configuration of a TCP tunnel used in the present invention.

It has a client and a server. In the TCP proxy, the ISEC G / W consists of a kernel and a TCP proxy daemon, and the kernel consists of a redirection module. So when a client makes a connection request to the gateway's kernel, the connection request is passed to the server through the TCP proxy daemon. If a tunnel failure occurs in such a configured TCP tunnel, packet loss may make it difficult to maintain a TCP session.

Therefore, the packet loss prevention technique proposed by the present invention is a technique for early detection of a tunnel failure, and does not detect the failure at a specific time, but determines the failure by checking the number of packets received through the tunnel. Therefore, the number of packets lost in a tunnel failure is limited to a certain number of packets regardless of the amount of packets transmitted. By doing so, it is possible to minimize the retransmission of packets even in the event of a tunnel failure and to suppress TCP session disconnection.

6 is a conceptual diagram illustrating an example of detecting a tunnel failure in FIG. 4.

Therefore, the tunnel failure detection technique used in the packet loss prevention technique is based on the number of packets flowing into the virtual tunnel device connected to each tunnel. In other words, when one of the tunnels in the multiplexing group fails, the packet inflow to the normal tunnel is increased while the packet inflow into the normal tunnel increases. Therefore, if the total amount of packets flowing into the remaining tunnels reaches a certain threshold while the amount of packets flowing into the tunnel to be detected reaches zero, the tunnel is considered to have failed.

That is, as shown in FIG. 6, while a packet having a specified threshold or more is received, a tunnel in which no packet is received is considered as a failure. As a result, when a multiplexed tunnel fails, it can be detected early at the kernel level, thereby minimizing packet loss, thereby maintaining a TCP session even in the event of a tunnel failure.

7 is a conceptual diagram illustrating an example of preventing the tunnel failure in FIG. 4.

Therefore, if it detects a fault tunnel, it routes the packet routed back to the normal tunnel. At this time, the copied packet is also sent to the failed tunnel, so that it can be determined as a normal tunnel when the failed tunnel is restored. That is, if the packet is not continuously transmitted to the fault tunnel, even if the tunnel state is restored to normal, no packet is introduced from the receiver side, and thus the packet is considered as the fault tunnel. Here, the copied packet indicates that the packet is copied by turning on a specific bit in the header part, and the receiver automatically discards it to use it only for inspection for tunnel recovery.

8 is a conceptual diagram illustrating an example of preventing a delay due to packet loss by a queue timer when a tunnel fails in FIG. 4.

Another technology included in packet loss prevention technology is the Queue Timer. This is to prevent queuing delay due to packet loss. For example, if some packets are lost due to the instability of the tunnel, the preceding packets queued for reordering the packets will not be able to exit the queue, waiting for the lost packets. As a result, the transmission rate of the tunnel becomes poor and TCP session maintenance is difficult. Therefore, it is the role of the queue timer to solve the queuing delay by forcibly releasing packets enqueued after a certain time.

So, as shown in Figure 8, for example, assume that the uppermost of the three tunnel device is a failure state, the tunnel in the middle and the lower tunnel operates normally. At this time, the sequence number 3 and 6 packets are transmitted through the tunnel above the fault, the sequence number 2 and 5 packets are transmitted through the tunnel in the normal center, and the sequence number 1 through the normal tunnel below. Let's say you are sending a packet of 4. If a failure occurs in the upper tunnel, the packets of sequence numbers 3 and 6 are lost, and in the receiver queue, the lost packets wait to receive the packets of sequence numbers 3 and 6, so they cannot dequeue. The delay phenomenon occurs. Therefore, the dequeuing delay problem is solved by forcibly releasing packets of sequence numbers 1, 2, 4, and 5 which are packets not dequeued within a predetermined time from the queuing time point.

As described above, the present invention is to determine the failure of the tunnel by checking the number of packets received through the tunnel so that the number of packets lost during the tunnel failure is limited to a specific number of packets regardless of the amount of packets transmitted.

As described above, in the packet loss prevention method for WAN optimization according to the present invention, when the traffic is properly distributed by multiplexing the tunnel between two gateways, the failure of the tunnel is determined by examining the number of packets received through the tunnel. There is an effect that the number of packets lost in case of failure can be limited to a certain number of packets irrespective of the amount of packets transmitted.

Although the above has been described as being limited to the preferred embodiment of the present invention, the present invention is not limited thereto and various changes, modifications, and equivalents may be used. Therefore, the present invention can be applied by appropriately modifying the above embodiments, it will be obvious that such application also belongs to the scope of the present invention based on the technical idea described in the claims below.

Claims (6)

Checking a number of packets received through the tunnel; A second step of detecting a tunnel failure based on the number of packets flowing into the virtual tunnel device connected to each tunnel after the first step; A third step of, if there is a tunnel detected as a failure tunnel in the second step, rerouting the packet routed to the failure tunnel toward the normal tunnel and transmitting the copied packet to the failure tunnel; Packet loss prevention method for WAN optimization, characterized in that performing including. The method of claim 1, wherein the second step, The criterion for determining whether one tunnel fails among the tunnels constituting the multiplexing group is set such that the packet inflow does not occur toward the failed tunnel while the packet inflow increases to a normal tunnel. How to avoid packet loss. The method of claim 1, wherein the third step, The copied packet is a packet loss prevention method for WAN optimization, characterized in that the packet is copied using a specific bit of the header portion. The method of claim 3, wherein the third step, The packet loss prevention method for WAN optimization, characterized in that the receiver automatically discards the copied packet. The method according to any one of claims 1 to 4, wherein the packet loss prevention method for WAN optimization, A fourth step of preventing a queuing delay with a queue timer after the third step; Packet loss prevention method for WAN optimization, characterized in that further comprising. The method of claim 5, wherein the fourth step, And dequeuing a packet that is not dequeued within a predetermined time from a queuing time point by using the queue timer.

Images