CN1848802A - The Method of Realizing IPv6 High Performance Interconnection Based on P2P on IPv4 - Google Patents

Abstract

A method for realizing IPv6 high-performance interconnection on IPv4 based on P2P includes connecting boundary gateways of IPv6 ' islet '' to be a overlapped network of P2P on IPv4 network and using it to interact reachable information to automatically set up IPv6 -in-IPv4 tunnel between Ipv6 ' islets ', using tunnel between ' islets ' to retransmit it preferentially when external IPv6 data packet being sent to ' islets ' is received by boundary gateway.

Description

The Method of Realizing IPv6 High Performance Interconnection Based on P2P on IPv4

technical field

A method for realizing IPv6 high-performance interconnection on IPv4 based on P2P belongs to the field of Internet technology, and in particular relates to a tunnel technology for interconnecting IPv4 and IPv6 networks.

Background technique

The IPv4 protocol has a serious defect of insufficient address space. For this problem, traditional solutions such as network address translation (NAT) and other technologies can temporarily slow down the impact of address consumption, but it destroys the end-to-end characteristics of the network. Fundamentally solve the problem of address allocation. With the rise of P2P applications in recent years, users need a unique IP address, and traditional NAT devices can no longer meet the needs of users. The Internet Engineering Task Force (IETF) began to conduct IPv6 research in the mid-1990s. Compared with the IPv432-bit address space, the IPv6128-bit space can allocate globally unique IP addresses to all users, thereby providing good support for end-to-end applications. In addition, IPv6 can also provide better support for security, mobility, and service quality. Therefore, the transition of the Internet from the existing IPv4 to IPv6 has become an inevitable trend of network development. Since the existing IPv4 network has formed a certain scale, in order to protect existing equipment and investment, it is impossible to replace the network to IPv6 at one time. The transition will be a long-term process. Only by doing a good job in the integration of IPv4 and IPv6 networks can a smooth transition be achieved. .

The specific transition technology can be divided into three types according to the type: dual-stack technology, protocol conversion technology and tunnel technology. The disadvantage of the dual-stack technology is that it cannot be deployed on a large scale at one time, and the protocol conversion technology has not been recommended by the IETF due to defects such as destroying the end-to-end feature. Therefore, the most widely used technology is the tunnel technology. As shown in Figure 1 (IPv6-in-IPv4 tunnel), if the data packets in the left IPv6 network need to be transmitted to the right IPv6 network, IPv6-in-IPv4 encapsulation needs to be performed at the left dual-stack router, that is, the original Some IPv6 data packets are used as the payload of the encapsulated IPv4 packet. At the same time, the destination address of the encapsulated IPv4 packet is set as the IPv4 address of the right dual-stack router and forwarded to the IPv4 network. When the encapsulated IPv4 packet passes through the IPv4 When the network reaches the right dual-stack router, the router decapsulates the packet, restores it to an IPv6 data packet and forwards it to the right IPv6 network, thereby realizing the interconnection of IPv6 networks that cannot be directly connected.

RFC3574, RFC3750, RFC4029, and RFC4057 divide existing networks into four types according to different network scopes: ISP Networks (ISP Networks), 3GPP Networks (3GPP Networks), Enterprise Networks (Enterprise Networks) and Unmanaged Networks (Unmanaged Networks) . The ISP network is the backbone network range controlled by the operator, and is generally built using dual-stack or pure IPv6 technology, while the 3GPP network is the standard for 3G mobile.

For unmanaged networks, that is, simple Office/Home networks, NAT is generally used to solve the problem of not having a global IPv4 address. Users inside NAT can use IPv6-in-UDP-in-IPv4 tunnel technology such as Teredo to traverse NAT, but the tunnel configuration is complicated. At the same time, this technology destroys the original IPv4 security model and easily causes the internal information leakage of NAT. Therefore, it is a better choice to directly upgrade the gateway to provide IPv6 support internally. Since the surrounding IPv4 network is not directly connected to the IPv6 backbone network, these unmanaged networks with upgraded gateway devices have become IPv6 "isolated islands" distributed in the IPv4 network.

For enterprise networks, RFC4057 proposes three deployment schemes for enterprise network transition: Wide-Scale Dual-Stack Deployment, Sparse Dual-Stack Deployment, and IPv6 Dominant Network Deployment ), for the first and third schemes, if the ISP has not provided IPv6 support, the enterprise network will also become an IPv6 "island"; for the second scheme, in order to prevent internal information leakage of the enterprise, IETF recommends that IPv6 access nodes be placed on Inside the enterprise network, in this way, from the outside of the enterprise network, it is still an "isolated island" of IPv6.

To sum up, in the early stage of transition, IPv6 "isolated islands" will exist in the existing IPv4 network on a large scale. They have the following characteristics:

(1) The address structure is simple and can be represented by a unified address prefix;

(2) There is no direct link to IPv6, and it is generally connected to the IPv6 backbone network through an IPv6-in-IPv4 tunnel;

According to the principle that IPv6 address allocation is easy to converge, the first one is easy to meet. Therefore, even in the ISP network, the customer network that is upgraded before the Backbone, that is, the backbone network, can also be regarded as an IPv6 "island".

On the other hand, with the implementation of China's Next Generation Internet Demonstration Project, China has begun to build several IPv6 test networks including CERNET2. These networks can be regarded as IPv6 "big islands". There is a high-speed interconnection between other IPv6 backbone networks (dual-stack or pure IPv6 can be used), which can be considered as an IPv6 "continent".

In order to allow IPv6 "isolated island" users to connect to the backbone network, IETF generally recommends the use of configuration tunnel technology to solve the connectivity problem between the "isolated island" and the IPv6 "continent". The isolated island is connected to the IPv6 "mainland" through an IPv6-in-IPv4 configuration tunnel. In order to reduce the burden of tunnel management, auxiliary means such as Tunnel Broker can be used. However, no matter which method is used, there will be one or A small number of IPv6 access nodes are responsible for forwarding traffic from IPv6 "islands" to access other IPv6 networks. In this way, whether the "island" accesses the interior of the IPv6 "continent" or the traffic between "islands" accesses each other, it needs to be forwarded through the IPv6 access node. On the other hand, most of the IPv6 "islands" are formed by upgrading the original IPv4 network, and the original applications will be upgraded to support IPv6, including various services. At the same time, because the IPv6 "continent" has just been formed, there are relatively few internal applications That is to say, the business is concentrated in the original network range upgraded from the IPv4 network, and the IPv6 network accessed by "isolated island" users may be mainly distributed in the address range of other "isolated islands". Therefore, a large part of the traffic forwarded by the IPv6 access node leads to other "isolated islands", which may cause the IPv6 access node to be overloaded and become the bottleneck of the entire network structure. Then, how to divert the traffic between IPv6 "isolated islands" in the above scenario and provide better access services for IPv6 "isolated island" users is the main problem to be solved in the present invention.

Contents of the invention

The purpose of the present invention is to provide a method for automatically establishing IPv6-in-IPv4 tunnels between IPv6 "isolated islands" based on the existing IPv4 network, in order to share the traffic of ordinary IPv6 tunnel access nodes and improve the access rate. into the quality.

The idea of the method provided by the present invention is: all IPv6 "isolated island" border gateways (IPv4/IPv6 dual-stack routers or IPv4/IPv6 dual-stack three-layer switches) utilize the existing IPv4 network to connect into a P2P overlapping network, and the interaction can reach Information, so as to form a one-way tunnel table on each border gateway; after the border gateway receives the IPv6 data packet sent to the outside of the "isolated island", it first queries the tunnel table, performs IPv6-in-IPv4 encapsulation, and uses it only if the query fails The default tunnel forwarding; after the encapsulated data packet reaches the destination end of the tunnel through the IPv4 network, the border gateway of the destination end decapsulates it, restores it to an IPv6 packet, and forwards it to the destination IPv6 "island". Realize the function of diverting traffic between IPv6 "islands", as shown in Figure 3.

The present invention is characterized in that: the method is a method for automatically establishing IPv6-in-IPv4 tunnels between IPv6 "isolated islands" using P2P technology on the basis of the existing IPv4 network, and the IPv6 "isolated islands" refer to the surrounding The IPv4 network is not yet a simple Office/Home network or enterprise network connected to the IPv6 backbone network, and the tunnel establishment method contains the following steps in turn:

Step 1, the border gateways of all IPv6 "isolated islands" obtain the IPv6 global address range and access tunnel information from the ISP;

Step 2, each IPv6 "isolated island" described in step 1 sets the IPv6 "isolated island" border gateway IPv4 address, IPv6 address and reachable IPv6 address range according to the content obtained in step 1, and configures the default IPv6 access tunnel;

Step 3. The border gateways of all IPv6 "isolated islands" form a P2P overlay network on the basis of the IPv4 network according to the following steps, so as to exchange reachable information through the overlay network, including their own and other node address information;

Step 3.1, registration: the border gateways of each IPv6 "isolated island" register with the registration server after startup, and obtain other P2P node information from the registration server;

Step 3.2, establish adjacency relationship: each IPv6 "isolated island" border gateway selects some nodes from all other P2P nodes to establish an overlapping adjacency relationship according to its own needs;

Step 3.3, exchange reachability information: Obtain reachability information of all other border gateways from the adjacent nodes described in step 3.2, and spread its own address information to the entire P2P network through neighbor nodes;

Step 3.4, forming a one-way tunnel table: all IPv6 "island" border gateways use the obtained reachability information to form tunnel tables leading to other "islands";

Step 4, in the IPv6 "isolated island" border gateway described in step 1, after a border gateway judges that the data packet received is a data packet that needs to be forwarded, as required, selectively carry out from the following operations:

If: the data packet is an IPv6 packet sent to the outside of the "isolated island", then the border gateway of the "isolated island" queries the tunnel table obtained in step 3:

If the query is successful, the tunnel endpoint information is obtained, and then IPv6-in-IPv4 encapsulation is performed on the IPv6 packet, the source address is the IPv4 address of the local border gateway, and the destination address is the IPv4 address at the other end of the tunnel;

If the query fails, the default tunnel is used for IPv6-in-IPv4 encapsulation, the source address is the IPv4 address of the border gateway, and the destination address is the IPv4 address of the other end of the default tunnel;

If: when the belonging data packet is an IPv4 packet, the "isolated island" border gateway checks whether the IPv4 header is followed by the IPv6 header, which is an IPv6-in-IPv4 encapsulated packet, and if so, the data packet is decapsulated;

In step 5, the "isolated island" border gateway mentioned in step 4 queries the dual-stack routing and forwarding table and forwards the data packet finally formed in step 4 to complete the tunnel forwarding operation.

On the basis of the existing IPv4 network, the method of using P2P technology to automatically establish tunnels between IPv6 "isolated islands" proposed by the present invention can be widely used in the early stage of IPv4/IPv6 transition, which greatly alleviates IPv6 tunnel access equipment (such as Tunnel Broker, etc.) pressure, improving the performance of IPv6 access. Another advantage of the present invention is that it is suitable for incremental deployment, plug and play, and can be popularized by gradually upgrading IPv6 "isolated island" border network management software. Therefore, the present invention can be promoted by gradually upgrading the software of IPv6 "isolated island" border network equipment, or by special network equipment. The invention has been applied in the network equipment jointly developed by Tsinghua University and Biwei Network Technology Co., Ltd., and is planned to be popularized in CERNET2 and China Next Generation Network CNGI.

Description of drawings

Figure 1. Schematic diagram of IPv6 in IPv4 tunnel;

Figure 2. Schematic diagram of the initial stage of IPv4/IPv6 transition:

表示配置隧道；

Indicates the configuration tunnel;

Figure 3. Schematic diagram of the P2P overlay network scheme:

表示配置隧道

表示P2P邻居

Indicates the configuration tunnel

Indicates a P2P neighbor

Indicates automatic tunnel Indicates registration to join the P2P network

Tunnel Table is the tunnel table of each IPv6 "isolated island" border gateway, recording the IPv4 address and reachable IPv6 address range of other "isolated island" border gateways;

Figure 4. The border gateway joins the P2P network process;

Figure 5. The processing process of the forwarded data by the border gateway.

Detailed ways

At the control level, the registration server maintains P2P node information and provides access services. The process of adding an IPv6 "isolated island" to a P2P network is shown in Figure 4, and it needs to go through the following steps:

(1) Registration: After the IPv6 "isolated island" border router is started, it will register with the registration server and obtain other P2P node information from the registration server;

(2) Establish an adjacency relationship: select a part from the above-mentioned nodes to establish an adjacency relationship of the Overlay;

(3) Exchange reachability information: Obtain reachability information of all other border gateways from neighbors, and spread its own address information to the entire P2P network through neighbors;

(4) Form a tunnel table: use the obtained reachability information to form a tunnel table leading to other IPv6 "islands".

When the P2P network node information changes, the information diffusion is completed through the update mechanism of the P2P network, the tunnel table is updated in time, and the validity of the boundary gateway tunnel information is maintained.

At the data level, the data forwarding process is shown in Figure 5. If the border gateway receives an IPv6 packet that needs to be forwarded to the outside of the "isolated island", it queries the tunnel table maintained at the control plane. If the query is successful, the tunnel endpoint information can be obtained. For IPv6 Packets are encapsulated in IPv6-in-IPv4, the source address is the IPv4 address of the border gateway, and the destination address is the IPv4 address of the other end of the tunnel; if the query fails, the default tunnel is used for IPv6 in IPv4 encapsulation, the source address is the IPv4 address of the border gateway, and the destination address is The IPv4 address of the other end of the default tunnel; if the data packet received by the border gateway is an IPv4 packet, check whether the IPv4 header is followed by the IPv6 header, and if so, decapsulate the packet; after the above encapsulation/decapsulation processing, Query the dual-stack routing table to determine the next hop and forward it. The encapsulation mechanism in the technical invention can adopt the IPv6-in-IPv4 configuration tunnel packet encapsulation mechanism defined in RFC2893.

The specific application examples are as follows. In Figure 3, it is assumed that the P2P network already includes border gateways Router A and Router B of "IPv6 island A" and "IPv6 island B". The server obtains the IPv4 address information of Router A and Router B, selects Router B as a P2P neighbor and establishes an adjacency relationship with it, and obtains links to 2001:250:f001:f002:28ba∷/80 and 2001:250 from Router B: The tunnel information of f001:f002:20a∷/80 and the tunnel information up to 2001:250:f001:f002:fe30∷/80 are updated to other P2P nodes through Router B, so that all P2P nodes have updated their own tunnel information. Information about IPv6-in-IPv4 tunnels to other "islands".

When Router C receives the IPv6 data packet whose destination address is 2001:250:f001:f002:20a::2005, it first queries the tunnel table and finds out that the tunnel endpoint is Router A, and then performs IPv6-in-IPv4 encapsulation on the IPv6 packet , the destination address is the IPv4 address of Router A: 202.112.49.247, the source address is the IPv4 address of Router C: 59.66.116.43; the encapsulated packet is forwarded to the IPv4 network through the dual-stack gateway forwarding mechanism; when the encapsulation After the packet arrives at Router A, Router A detects that the IPv4 packet header is followed by the IPv6 packet header, that is, the IPv6in IPv4 encapsulated packet, then decapsulates it, and then forwards it to the inside of the IPv6 "isolated island" through the dual-stack gateway forwarding mechanism , and finally sent to the destination host.

Through the maintenance and update of the above P2P network, all IPv6 "isolated island" border gateways connected to the P2P network can obtain and update the IPv4 in IPv6 tunnel information leading to other "isolated islands" in a timely manner, and the traffic between "isolated islands" is preferentially used These tunnels thus divert the traffic between IPv6 "islands" and greatly relieve the pressure on the default access tunnel. It can be seen that the present invention has achieved the intended purpose.

Claims (1)

1. The method for realizing IPv6 high-performance interconnection on IPv4 based on P2P is characterized in that: the method is a method of using P2P technology to automatically establish IPv6-in-IPv4 tunnels between IPv6 "isolated islands" on the basis of existing IPv4 networks Method, described IPv6 " isolated island " refers to surrounding or IPv4 network, the simple Office/Home network or enterprise network that has not yet been connected with IPv6 backbone network, described tunnel establishment method contains following steps successively: Step 1, the border gateways of all IPv6 "isolated islands" obtain the IPv6 global address range and access tunnel information from the ISP; Step 2, each IPv6 "isolated island" described in step 1 sets the IPv6 "isolated island" border gateway IPv4 address, IPv6 address and reachable IPv6 address range according to the content obtained in step 1, and configures the default IPv6 access tunnel; Step 3. The border gateways of all IPv6 "isolated islands" form a P2P overlay network on the basis of the IPv4 network according to the following steps, so as to exchange reachable information through the overlay network, including their own and other node address information; Step 3.1, registration: the border gateways of each IPv6 "isolated island" register with the registration server after startup, and obtain other P2P node information from the registration server; Step 3.2, establish adjacency relationship: each IPv6 "isolated island" border gateway selects some nodes from all other P2P nodes to establish an overlapping adjacency relationship according to its own needs; Step 3.3, exchange reachability information: Obtain reachability information of all other border gateways from the adjacent nodes described in step 3.2, and spread its own address information to the entire P2P network through neighbor nodes; Step 3.4, forming a one-way tunnel table: all IPv6 "island" border gateways use the obtained reachability information to form tunnel tables leading to other "islands"; Step 4, in the IPv6 "isolated island" border gateway described in step 1, after a border gateway judges that the data packet received is a data packet that needs to be forwarded, as required, selectively carry out from the following operations: If: the data packet is an IPv6 packet sent to the outside of the "isolated island", then the border gateway of the "isolated island" queries the tunnel table obtained in step 3: If the query is successful, the tunnel endpoint information is obtained, and then IPv6-in-IPv4 encapsulation is performed on the IPv6 packet, the source address is the IPv4 address of the local border gateway, and the destination address is the IPv4 address at the other end of the tunnel; If the query fails, the default tunnel is used for IPv6-in-IPv4 encapsulation, the source address is the IPv4 address of the border gateway, and the destination address is the IPv4 address of the other end of the default tunnel; If: when the belonging data packet is an IPv4 packet, the "isolated island" border gateway checks whether the IPv4 header is followed by the IPv6 header, which is an IPv6-in-IPv4 encapsulated packet, and if so, the data packet is decapsulated; In step 5, the "isolated island" border gateway mentioned in step 4 queries the dual-stack routing and forwarding table and forwards the data packet finally formed in step 4 to complete the tunnel forwarding operation.

Images