CN101778048A - Data forwarding method, load balance scheduler and load balance system - Google Patents

Abstract

The invention provides a data forwarding method, a load balancing scheduler and a load balancing system. It involves the computer field; it solves the problem of low data forwarding efficiency in four-layer load balancing. The method includes: a load balancing scheduler receives a client request, and determines a four-layer mark corresponding to the client request; the load balancing scheduler selects any port from at least one port with the four-layer mark, and forwards the described port to the server. customer request. The technical solution provided by the invention is suitable for four-layer load balancing.

Description

Data forwarding method, load balancing scheduler and load balancing system

technical field

The invention relates to the computer field, in particular to a data forwarding method and a load balancing scheduler.

Background technique

Today's computer technology has entered a network-centric computing period. Due to the simplicity, ease of management and ease of maintenance of the client/server model, the client/server computing model has been widely adopted on the Internet. The Web is also changing from a content delivery mechanism to a service platform, and a large number of services and applications (such as news services, online banking, e-commerce, etc.) are carried out around the Web. From the perspective of the development of network technology, the growth of network bandwidth is much higher than the growth of processor speed and memory access speed. Many studies have shown that the reason why it is difficult for Gigabit Ethernet to achieve a throughput rate of 1Gb/s on the server is the protocol stack (TCP /IP) and operating system inefficiencies, as well as processor inefficiencies. In order to solve this dilemma, load balancing technology came into being. Due to the use of load balancing technology, automatic fault recovery can be realized, service time can be extended, and 24×7 reliable and continuous operation becomes possible. In addition, the load balancing scheduler generally supports the out-of-path return mode, that is, bypassing the traffic distributor to provide faster response time for those customers who are anxiously waiting for a large number of data file request responses

Four-layer load balancing software, the open source project 1vs provides us with a good load balancing function, which can meet the needs of four-layer load balancing. At present, the LVS project has provided a Linux Virtual Server framework for scalable network services. The LVS framework can be used to realize highly scalable and highly available web services such as Web, Cache, Mail and Media; on this basis, it is possible to develop highly scalable and highly available e-commerce applications that support a large number of users. The fourth layer of load balancing adopts the load balancing scheduling technology based on the IP layer, and intercepts and rewrites/forwards the IP message at the IP layer by modifying the tcp/ip protocol stack of the operating system kernel. In this way, when a client request arrives, the load balancing scheduler selects a server from the server according to the set scheduling method and the load conditions of the servers in the server pool, and forwards the request to this server. Since these operations are completed in the core space of the operating system, the load balancing scheduler has high efficiency and low scheduling overhead, and can schedule many servers, making the entire system have excellent scalability. The load balancing scheduler can automatically screen out server failures in the server pool.

In the process of implementing the above four-layer load balancing, there are the following problems:

For the traditional four-layer load balancing scheduler, a load balancing service can only use one TCP or UDP port. For applications that need to occupy multiple ports, the load balancing scheduler can only establish these ports in different load balancing Among services, when providing load balancing for a multi-port application, multiple load balancing services need to be established, which takes up a lot of resources, and the load balancing forwarding efficiency is low.

Contents of the invention

The invention provides a data forwarding method and a load balancing scheduler, which solves the problem of low data forwarding efficiency in four-layer load balancing.

A data forwarding method, comprising:

The load balancing scheduler receives the client request and determines the four-layer mark corresponding to the client request;

The load balancing scheduler selects any port from at least one port with the four-layer mark, and forwards the client request to the server.

Further, the load balancing scheduler receives a client request, and before the step of determining the four-layer label corresponding to the client request, it also includes:

Bind at least one port between the same server and the load balancing scheduler.

Further, the binding of at least one port between the same server and the load balancing scheduler includes:

Carry out the mark operation under iptables marking mechanism to described at least one port, add identical mark as the four-layer mark of port for described at least one port;

Locating the four-layer mark on the server.

Further, the above data forwarding method also includes:

The corresponding relationship between the server and the four-layer mark is stored.

Further, the port is a TCP protocol port or a UDP protocol port.

The present invention also provides a load balancing scheduler, including:

A tag determination module, configured to receive a client request and determine a four-layer tag corresponding to the client request;

A forwarding module, configured to select any port from at least one port with the four-layer mark, and forward the client request to the server.

Further, the above load balancing scheduler also includes:

The binding module is used to bind at least one port between the same server and the load balancing scheduler.

The present invention also provides a load balancing system, including a load balancing scheduler, there is at least one port between the load balancing scheduler and each server, and at least one port of the same server has the same four-layer identification;

The load balancing scheduler is configured to receive a client request, determine a Layer 4 mark corresponding to the client request, select any port from at least one port with the Layer 4 mark, and forward the client request to a server.

Embodiments of the present invention provide a data forwarding method, a load balancing scheduler, and a load balancing system. After the load balancing scheduler receives a client request, it determines the corresponding four-layer mark according to the client, and from at least one specific four-layer Select any port among the marked ports, forward the client request to the server through the port, realize the data forwarding to the same server through multiple ports, and solve the problem of low data forwarding efficiency in the four-layer load balancing.

Description of drawings

FIG. 1 is a flowchart of a data forwarding method provided by an embodiment of the present invention;

FIG. 2 is a schematic diagram of an application scenario of an embodiment of the present invention;

FIG. 3 is a schematic structural diagram of a load balancing scheduler provided by an embodiment of the present invention;

FIG. 4 is a schematic structural diagram of a load balancing scheduler provided by another embodiment of the present invention.

Detailed ways

In the process of four-layer load balancing, a load balancing service can only use one TCP or UDP port. For applications that need to occupy multiple ports, the load balancing scheduler can only establish these ports in different load balancing services. That is, when providing load balancing for a multi-port application, multiple load balancing services need to be established, which takes up a lot of resources, and the load balancing forwarding efficiency is low.

In order to solve the above problems, an embodiment of the present invention provides a data forwarding method. The specific process of using this method to realize four-layer load balancing is shown in Figure 1, including:

Step 101, binding at least one port between the same server and the load balancing scheduler;

The application environment of the embodiment of the present invention is as shown in Figure 2, comprising a load balancing scheduler 201 and at least one server 202 connected to the load balancing scheduler, each server is connected to the load balancing scheduler 201 through at least one port, The port can be a TCP or UDP protocol port; in this application environment, the load balancing scheduler 201 acts as a forwarding platform, forwarding the received data between the client and the server 202 .

In this step, firstly, in the fourth layer (transport layer) of the OSI (Open System Interconnection, Open System Interconnection) seven-layer model, a mark is assigned to all ports that can be used for services, specifically through the iptables mark mechanism for the port The following mark adding operation is realized, and the added mark is a four-layer mark, and the load balancing scheduler working on the fourth layer can recognize this four-layer mark.

When allocating the four-layer mark, you can assign the same mark to multiple ports and correspond to the same IP according to the actual situation. In this way, it is equivalent to binding the ports with the same mark. For the load balancing scheduler working at the transport layer, these ports are the same.

After allocating the four-layer mark, locate the different four-layer marks on different servers, so that each server establishes a corresponding relationship with a four-layer mark, that is, at least one corresponding to the four-layer mark A corresponding relationship is established between the ports, so that the data forwarded from at least one of the above ports can be received. The corresponding relationship between the server and the four-layer mark, and the corresponding relationship between the four-layer mark and the IP are all stored in the load balancing scheduler.

Step 102, the load balancing scheduler receives the client request, and determines the server corresponding to the client request;

In this step, the load balancing scheduler receives a client request, indicating that a client requests data of a service. The client request carries the IP of the target server, and the load balancing scheduler determines the layer-4 mark corresponding to the client request according to the IP of the target server.

Step 103, the load balancing scheduler selects a port from at least one port with the four-layer mark, and forwards the client request to the server through the port;

In this step, the load balancing scheduler first obtains at least one layer-4 marked port determined in step 103, and selects one of these ports, and forwards the client request to the target server through the port.

Optionally, certain port selection rules may also be adopted, such as using each port in rounds.

Embodiments of the present invention also provide a load balancing scheduler, the structure of which is shown in Figure 3, including:

A mark determining module 301, configured to receive a client request, and determine a four-layer mark corresponding to the client request;

The forwarding module 302 is configured to select any port from at least one port with the four-layer mark, and forward the client request to the server.

Further, the above load balancing scheduler is shown in Figure 4, and also includes:

A binding module 303, configured to bind at least one port between the same server and the load balancing scheduler.

The embodiment of the present invention also provides a load balancing system, including a load balancing scheduler, there is at least one port between the load balancing scheduler and each server, and at least one port of the same server has the same four-layer identification;

The load balancing scheduler is configured to receive a client request, determine a Layer 4 mark corresponding to the client request, select any port from at least one port with the Layer 4 mark, and forward the client request to a server.

The above load balancing scheduler and load balancing system can be combined with a data forwarding method provided by an embodiment of the present invention. After the load balancing scheduler receives a client request, it determines the corresponding four-layer mark according to the client request, from Select any port from at least one of the four-layer marked ports, and forward the client request to the server through this port, realizing forwarding data to the same server through multiple ports, and solving the problem of low data forwarding efficiency in four-layer load balancing .

Those of ordinary skill in the art can understand that all or part of the steps carried by the methods of the above embodiments can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium, and the program can be executed when executed , including one or a combination of the steps of the method embodiment.

In addition, each functional unit in each embodiment of the present invention may be implemented in the form of hardware, or may be implemented in the form of software functional modules. If the integrated modules are realized in the form of software function modules and sold or used as independent products, they can also be stored in a computer-readable storage medium.

The storage medium mentioned above may be a read-only memory, a magnetic disk or an optical disk, and the like.

The above is only a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Anyone skilled in the art can easily think of changes or substitutions within the technical scope disclosed in the present invention. Should be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope described in the claims.

Claims (8)

1. a data forwarding method is characterized in that, comprising:

Load balance scheduler receives client requests, determines four layers of mark of this client requests correspondence;

Described load balance scheduler is selected arbitrary port from least one port with described four layers of mark, to the described client requests of server forwards.

2. data forwarding method according to claim 1 is characterized in that, described load balance scheduler receives client requests, determines also to comprise before the step of four layers of mark of this client requests correspondence:

At least one port between same server and load balance scheduler is bound.

3. data forwarding method according to claim 2 is characterized in that, described at least one port between same server and load balance scheduler is bound comprises:

The interpolation mark that described at least one port carries out under the iptables tagging mechanism is operated, for described at least one port adds the four layer mark of identical mark as port;

Described four layers of mark are navigated on the described server.

4. data forwarding method according to claim 3 is characterized in that, also comprises:

Store the corresponding relation of described server and described four layers of mark.

5. data forwarding method according to claim 1 is characterized in that, described port is Transmission Control Protocol port or udp protocol port.

6. a load balance scheduler is characterized in that, comprising:

The mark determination module is used to receive client requests, determines four layers of mark of this client requests correspondence;

Forwarding module is used for selecting arbitrary port from least one port with described four layers of mark, to the described client requests of server forwards.

7. load balance scheduler according to claim 6 is characterized in that, also comprises:

Binding module is used at least one port between same server and load balance scheduler is bound.

8. a SiteServer LBS is characterized in that, comprises load balance scheduler, respectively has at least one port between described load balance scheduler and between each server, and at least one port of same server has four layers of same sign;

Described load balance scheduler is used to receive client requests, determines four layers of mark of this client requests correspondence, selects arbitrary port from least one port with described four layers of mark, to the described client requests of server forwards.

Images