introduction
In 2000, AH LS WEDE, etc., for the first time, the network coding theory, and the maximum network traffic can be achieved through network coding. In 2003, I Yeung and CA i prove that linear network coding can realize the maximum stream of the network. Subsequently, T. H O proposed a random network encoding theory, its idea is to participate in the transmission node in the network, and the data transmitted on the output channel is a random linear combination of data transmitted by the point multiple input channels, they also prove The receiving node can correctly restore the information sent by the source with a large probability.
Network coding improves network throughput and reliability, but also brings unloadable security issues, mainly including pollution and stealing two types of problems. T.h O et al. Proposed a network encoding that detects whether pollution attacks are present.
JAGG i's differently designed an adaptive security network encoding for attacker capabilities. Sun Yue, etc. Considering the problem of multicast network failure under network coding.
1 Basic principle of network coding
1. 1 Network Code Definition
Maximum flow minimum sprinkler of network traffic: For known network flow diagrams, the maximum value W of the traffic S to the traffic of the band-kit T is equal to its minimum cut capacity, ie M AX FL OW (S, T) = m in C (s, t). For a network with only one content, you can get the maximum stream on the route.
In order to understand the theoretical basis of network coding, the following will explain the use of network encoding to maximum flow limit by the classic butterfly network as an example.
Be
Be
Figure 1 employs a multicast network with two destination nodes.
Figure 1 is a unit capacity having two destination nodes (the capacity of each side) multicast network. Assume that there is no error in each link and there is no time. The minimum sprinkler of the maximum flow, there is M AXFLO W (S, R L) = 2. At this time, the maximum propagation rate of the source node to the destination nodes R L and R2 cannot exceed 2. If no code is used, as shown in Fig. 1 (a), the links (3, 4) must send a bit to more than the capacity of the side.
However, if the network stream enters the node is allowed to encode, as shown in FIG. 1 (b), a scheme that simultaneously transmits A and B to the destination nodes R 1 and R2, "+" is different or the operation. Such a destination node R1 receives A and A + B, and the A + B can be restored to the A + B. Similarly, the destination node R2 can restore A and B. This allows the butterfly network to get the maximum flow limit.
2 Safety research of network coding in network
Although the original intention of network coding is to improve the throughput of the network, as further research has found that it is also a good way to transfer a secure network. However, in the actual network communication, there is still a security hazard, which mainly manifests the package pollution received by the data forwarding process, and the Byzantian attack is a common means and method for damaging data safe transmission.
2. 1 network encoded packet contamination
Attack of packet contamination is that the popular research objects in the security field are also the most worthwhile research topics. The attacker damages the network node and injects new information to form a new encoded data package. Since each node of network encoding has routing function but also has a routing function The affected honest nodes further affect other honest nodes from a damaged packet, which can cause pollution to spread like plagues. Obviously, even some injected packets, attackers can also reduce their performance. At present, a solution of anti-packet contamination techniques is to use asynchronous high efficiency linear check code, but this technique has seriously affected throughput.
Malicious packet. Attackers fake honest nodes pass incorrect information to neighbor nodes, which cause the original node information encoding that cannot be decoded by the expected next hop node, affecting the correct decoding of the downstream node. Malicious nodes may change the message carried by the data package or the information contained in the packet header leads to the header error. In a packet, some important information (eg, global coding cores in network encoding, location, destination of packets generated by packets, etc.) are recorded in the header, and arbitrary errors that may cause transmission. . If the global coding core changes, it is called a global coding core error, which will affect the decoding of the receiving node; if the information of the destination is changed, it may cause packet loss of the receiving node. In the classification of La M port, most of the errors can be attributed to the Byzantine error. Technology area
485 communication protocol program how to write (51 single-chip 485 communication program case)
How to use 485 communication agreement (typical circuit instance of traditional photoelectric isolation)
MODBUSRTU packet format detailed introduction
Eight modbus RTU data frame formats detailed
Modbus RTU Communication Protocol Format Detailed Description
Our other product:
