Ecological environment fuzzy networking system protection based on optical fiber intelligent network

2.1 Optical fiber access network technology

The access network refers to all the equipment between the backbone network and the user terminal. Its length is generally several hundred meters to several kilometers, so it is vividly called the “last kilometer”. The access methods of the access network include digital subscriber line (copper wire) access, optical fiber/coaxial hybrid network access (HFC), and optical fiber access.

The optical splitter is divided from the network structure and can be divided into feeder fiber, optical splitter and branch line. Each of these components is composed of different passive optical devices. The optical distribution network (ODN) realizes the bidirectional transmission of optical signals between an OLT (optical line terminal) and multiple ONUs (optical network units).

In the tree structure, one OLT corresponds to multiple ONT/ONUs, and an optical splitter is provided between one OLT and multiple ONT/ONUs. The advantage of the tree structure is that due to fewer jumpers, when the optical signal is transmitted in the optical cable line, its attenuation is small and the failure rate is low. And because this structure has fewer jumpers, database management is also more convenient. The shortcoming of the tree structure is that the number of optical cables required behind the optical splitter is large, and the number of pipes required increases accordingly. This situation is very obvious when the optical splitter is installed centrally.

Located between the network access point and the optical distribution point is called the distribution optical cable. When the users are densely distributed, the distribution optical cable is required to enter the room directly. At this time, the distribution optical cable should have both indoor and outdoor functions. If it is necessary to share the distribution optical cable, it is required that the distribution optical cable should have a strong midway fiber splitting function. When the optical splitter is installed in the corridor, the number of optical fiber cable cores should be considered as twice that of the optical splitter; when the optical splitter is installed outside the corridor, the number of optical fiber cable cores to the optical splitter should be twice that of the optical splitter. The number of cores of the distribution optical cable from the optical splitter to the corridor should be configured according to the number of owners of the corridor. Figure 1 shows the intelligent architecture of the optical fiber intelligent network communication network link.

OPEN IN VIEWER

Fig. 1

Intelligent architecture of optical fiber intelligent network access communication network link.

The PON technology can select a suitable method to ensure the reliability of the line according to the specific conditions of a series of project requirements such as the business type, cost and user type of the cell. The PON method can provide a lot of line reliability according to the actual needs and characteristics of the project. The transmission medium used in the PON mode is optical fiber, which to a certain extent provides the guarantee of the line material, while the common DSL mode uses copper cable for transmission. This transmission medium is subject to quality, distance and different line signals. The reliability of the transmission is reduced due to the crosstalk between them. The LAN mode adopts the Ethernet transmission protocol mode, which can achieve redundancy protection, because it detects the link through the response packet, and if a problem is found, the connection can be quickly restored. However, this method will increase the occupancy of equipment ports, thereby increasing the number of equipment ports, which not only poses a greater challenge to the energy and space of the computer room, but also increases capital investment.

The IP multicast monitoring function can manage multicast traffic in the second layer of the device. In this way, the data flow will only flow to those host interfaces that have participated in the multicast, and will not flow to those host interfaces that have exited the multicast, thereby greatly reducing the sending range and saving network traffic. When the IP multicast monitoring function is enabled on the switch, the switch will create a multicast forwarding table for each VLAN. When the switch receives a join message from a host, it automatically adds the port number of the host in the multicast forwarding table, and when the switch receives a leave message from a host, it automatically enters the multicast forwarding table to delete the port number where the host is located. Especially in video multicast, this function is very useful. By setting this function, the video server does not need to send a video stream for each video-on-demand, but completes the playback of the video through the multicast technology of layer-by-layer replication. Therefore, the use of network traffic can be greatly reduced.

Through the network technology of multicast, a unicast source or a multicast source can send a single data message to multiple subscribers at the same time. In the multicast technology, no matter how many receiving hosts, there is only one data packet transmitting on any link of the network, and the data packet will not be sent to all the hosts on the link. Thus, the data transmission speed is greatly improved, and the waste of network resources and the occurrence of network congestion are avoided.

When implementing multicast technology, two types of address support are required, namely, IP multicast address and Ethernet multicast address. A multicast group is identified by an IP multicast address. The reason why the Ethernet multicast address is needed is because the IP data packets transmitted on the network are all encapsulated in Ethernet frames. Because the host is required to receive both unicast and multicast packets at the same time, multiple IP addresses and Ethernet addresses are required. In IPv4, class D addresses are multicast addresses, which can be divided into reserved addresses, local link addresses, and management authority addresses according to different address ranges.

When a host wants to join a multicast group, it first sends a “Membership Report” message to its IP subnet multicast router, and at the same time, prepares its own IP module to receive the multicast group. If this host is the first machine in its IP subnet to apply for joining the multicast group, a piece of information about the multicast router and other related information will be established in the multicast distribution tree.

Through port aggregation (trunking) technology, load balancing among multiple ports can be realized, avoiding network congestion, and achieving the function of multiple low-bandwidth switching ports to achieve high-bandwidth links. The port aggregation function implemented on the device can bundle multiple physical ports together as a logical port, thereby providing higher communication bandwidth.

Another function of port aggregation is to enhance the reliability of connections between devices. When a port fails or a line is interrupted, data packets can continue to be forwarded through other ports, thereby effectively ensuring smooth communication. No matter how many ports there are in port aggregation, they are regarded as one port externally. Among the many ports that make up the port trunk, one port is designated as the master port (masterport). Because all member ports in port aggregation work in the same way, when you need to set the ports in it, you only need to set the main port. All settings for the main port, such as VLAN, STP and other functions, will be applied to other member ports at the same time.

2.3 Security analysis of optical fiber access network design

A large number of services in the PON network are carried by means of IP. Therefore, the security of the network should be paid attention to designing the scheme. If the network has security risks, it may cause the theft of user services and the destruction of user data. The network information is encrypted to prevent data from being stolen during network transmission. Since the equipment in the PON network uses the IP protocol in Ethernet for transmission, the equipment may be unstable due to interference from external factors or other environmental factors.

But PON network is not a panacea, it cannot solve all security problems. For example, web pages on the public network, even if the PON network is secure, if the web page program itself has loopholes or other security risks, it may be used by hackers to attack and steal data. Security issues like this should be resolved at the application layer. For those network security issues that may occur in the way of carrying services by IP, the PON network should have the corresponding bearing capacity to establish a strong security mechanism covering the entire network to prevent the emergence of various network security issues.

PON system refers to a system that adopts PON technology. As mentioned earlier, the network topology adopted by PON technology is a point-to-multipoint structure. That is, when the data is downlink, usually the same link is used to broadcast data packets to many receivers. The data packet sent by the OLT can be received by any ONT/ONU in the network, which poses a security risk. In the PON system, possible network security issues include embezzlement or abuse of user accounts, counterfeiting of IP addresses or MAC addresses, denial of service (Do S, Deny of Service) attacks, etc. In order to ensure the security of user information in the PON system, isolation technology, such as Layer 2 data isolation in a virtual local area network (VLAN), is required to ensure the security of user information in the system.

In the PON system, security measures should be taken below the MAC layer, because the security measures above the MAC are transparent to the transmission of frame header information including the MAC address. For other ONT/ONU MAC addresses, illegal ONT/ONU can still be obtained through illegal means. The use of security measures below the MAC layer can effectively prevent the MAC address from being stolen illegally. For the source address, destination address, length, type and other information, the security measures below the MAC layer can be processed by appropriate processing methods to ensure that this information is not stolen from illegal ONT/ONUs.

2.4 Design of optical fiber access networking scheme

FTTH adopts PON technology, and PON technology adopts passive optical splitting network to achieve full transparent transmission. At the same time, the FTTH method uses optical fiber to reach the home directly, and connects to the user access equipment ONT through the optical cable. The ONT is deployed in the user’s home or office, and the OLT realizes remote connection control. The ONT transmits voice and data services to the central office, which is processed by the OLT, and then transmitted through the central office network.

The terminal host and access router protocol stack are shown in Fig. 2.

OPEN IN VIEWER

Fig. 2

Protocol stack of terminal host and access router.

3.1 Fuzzy logic

The starting point of fuzzy logic is the concept of fuzzy sets. In general, the transition from membership to non-membership in fuzzy concentration is gradual rather than sudden. Fuzzy modeling is the most important issue in fuzzy logic theory. There are many explanations for fuzzy modeling. For example, a fuzzy set can be regarded as a fuzzy model of human concepts. Fuzzy modeling can also be simply understood as a method of forming a system model based on a description language with fuzzy levels. In a broader sense, fuzzy modeling can be interpreted as a qualitative modeling scheme, through which natural language is used to qualitatively describe system behavior. Fuzzy modeling in a narrow sense is a system description with fuzzy quantity. The amount of fuzzy is represented by fuzzy numbers or fuzzy sets related to language tags.

The difference from classical set theory is that in fuzzy set theory, elements have the concept of membership. By defining set membership as a probability distribution, fuzzy set theory can represent incomplete or inaccurate information. Based on fuzzy set theory, fuzzy logic deals with approximate concepts rather than exact factors. For example, you define a person’s height as 0.6 “high” and 0.4 “low” instead of “completely high” or “completely low”. Since fuzzy logic can handle approximate reasoning similar to human reasoning, it has been widely accepted by the industry and used in many applications. Compared with numerical values in mathematics, fuzzy logic uses non-numeric language variables to express facts. The fuzzy membership function is used to express the degree of the value belonging to the language tag.

Generally, a fuzzy logic-based system includes three steps: input, processing and output steps. The input step converts the input value into a language variable. The processing steps collect logical rules defined in the form of IF-THEN statements, and apply the rules to obtain results in language format. The output step converts the language results into numerical values.

As mentioned above, in the fuzzy logic-based routing algorithm, each node evaluates its neighbor’s location, mobility and channel conditions by exchanging hello messages. When a node must send a packet, the node uses fuzzy logic to calculate the average relay suitability of each neighbor based on the neighbor’s location, mobility, and channel conditions. Then, the node selects a relay node for each required broadcast area. For each neighbor, the calculation steps are as follows:

First, you use pre-defined language variables and membership functions to convert the location, mobility and signal strength of neighbor nodes into fuzzy values. The second is the mapping and combination of IF / THEN rules.

The third is defuzzification. You use the predefined output membership function and the proposed defuzzification method to convert the fuzzy value obtained in the previous step into a quantitative value.

After calculating the number of all members in the set, the member with the largest number is selected as the final result of the fuzzy logic algorithm.

3.2 Fuzzy logic routing

In order to improve the packet receiving rate of the sender and the receiver, the design adopts the ecological environment fuzzy networking opportunistic routing based on fuzzy logic. Considering that the distance from the sender node to the receiver node needs to use multi-hop to spread data, the problem of relay node selection becomes the focus of this article.

In the selection of relay nodes, the vehicle location, mobility and signal strength are considered. These three indicators are contradictory. If the node closest to the destination is selected as the relay node, it will minimize the number of relay and forwarding nodes. However, the relay node may lose data packets because the signal is weak. In addition, because the vehicle moves randomly, the relay vehicle may move out of the transmission range of the sender. These conflicts will be affected by mobility, distribution and signal fading. Therefore, the mathematical model of the optimal relay problem is very complicated to derive, and the solution based on it is too expensive for practical applications. Fortunately, fuzzy logic can deal with inaccurate and uncertain information. In this algorithm, a fuzzy logic-based method is used to identify the relay node that will produce the best result. The basic scheme is shown in Fig. 3.

OPEN IN VIEWER

Fig. 3

Routing process based on fuzzy logic.

1) Calculate fuzzy logic factors

In this routing algorithm, when receiving a greeting message from a neighbor, the node evaluates the neighbor based on the location of the vehicle, mobility and signal strength. Thus, by exchanging hello messages, each sending node maintains the evaluation results of each of its neighbors. When selecting relay nodes, you use these evaluation results. If the node does not receive any hello message from the neighbor in the time interval T, which is defined as 3 times the hello interval (1 s), the node initializes the neighbor’s evaluation value (including all factors that will be given). The neighbor will also be deleted from the neighbor list.

The interval of sending hello messages will have a certain impact on the performance of routing algorithms. If the greeting interval is too long, the neighbor information may be inaccurate and useless. Conversely, if the greeting interval is too short, the increased message overhead may cause performance degradation. The optimal greeting interval should be determined by the node density, node speed and a series of other factors. In this routing algorithm, this article uses a greeting interval of 1 s. According to the experience of previous research, this value is suitable for most scenarios.

When a node receives a hello message from neighbor X, it calculates the location factor (PF) according to the following formula. d(X) is the distance between the current sending node and node X. R is the maximum distance that can provide stable communication. This article assumes that each node has the same transmission power and the transmission power is constant, so that the transmission range of each node is the same. PF = { d ( X ) / R & R > d ( X ) 1 & R ⩽ d ( X )(1)

When receiving a hello message from neighbor x, the node calculates the mobility factor (MF) as shown in the following formula. MF ( s , x ) = | v ( x ) | - min | v ( y ) | max | v ( y ) |(2)

MF represents the mobility level of neighbor nodes. The higher the value of MF, the lower the mobility of the node. MF is initialized to 0. In the following formula, an exponential moving average is used because it wants to eliminate short-term errors. α is the smoothing factor, and its value is set to 0.7. After a lot of experiments and analysis by the researchers, this paper chooses 0.7 as the most suitable value. MF ( s , x ) ⇒ α MF i ( s , x ) + ( 1 - α ) MF i - 1 ( s , x )(3)

When receiving a hello message from neighbor X, the node calculates the channel condition indicator (CF), as shown in the following formula. It represents the average channel condition from the sending node to its neighbor nodes. Here CF is initialized to 0. RxPr is the received signal power, and RXThresh is the receive threshold. The RXThresh value is defined according to the received power. CF ( X ) = 1 - RXThresh / Rx Pr(4)

2) Obfuscation

In fuzzy logic, “fuzzification” is defined as the process of converting numerical values into fuzzy values using fuzzy membership functions.

The sender node uses the membership function and the previously calculated position factor to determine the degree to which the position belongs to {good, medium, bad}. When the position factor is 0.2, the vertical line showing the position factor meets “bad” and “medium” at (0.2, 0.6) and (0.2, 0.4) respectively. Therefore, the fuzzy value {good: 0, medium: 0.4, bad: 0.6} can be obtained. Note here that the membership function given above and other membership functions given later are defined based on simulation results.

3) IF/THEN rule mapping

Once the fuzzy value of the location factor, mobility factor and channel condition factor is calculated, the sender node uses the IF / THEN rule to calculate the node’s rank. The linguistic variable of rank is defined as {perfect, good, okay, no, bad, very bad}. If the location is good, the mobility is slow, and the channel conditions are good, then the level is perfect.

In an IF/THEN rule, the IF part is called the “cause” and the THEN part is called the “result”. If multiple rules are applied at the same time, their assessment results must be combined. The Min-Max method is used here. In the Min-Max method, for each rule, the minimum value of the antecedent is used as the finality. When combining different rules, the maximum value of the result needs to be used.

As shown in Fig. 4, assuming that the neighbor’s location, mobility and signal strength factor belong to the corresponding language variables {good: 1, medium: 0, bad: 0}, {slow: 0.7, medium: 0.2, fast: 0}, {Good: 0.5, Medium: 0.5, Bad: 0}. All rules are combined to give fuzzy results.

OPEN IN VIEWER

Fig. 4

Example of fuzzy logic rule mapping.

4) Defuzzification

Defuzzification is a process of generating numerical results based on the output membership function and the corresponding membership level qualifications. Here, this article uses the center of gravity (COG) method to defuzzify the results of the mapping in the previous section. More specifically, you can cut the output membership function with a horizontal line according to the corresponding degree, and remove the top. For the example given above, the level {Yes} is 0.25, the level {good} is 0.5 and the level {perfect} is 0.5.

Then, you calculate the centroid of this shape. The abscissa of the centroid will be the defuzzification value. If u(x) is used to represent the result function and x is used to represent the horizontal axis, the center of gravity will be calculated as follows: COG = ∫ u ( x ) xdx / ∫ u ( x ) dx(5)

Here COG indicates the suitability of the neighbor as a relay node. The higher the value, the better the neighbor node. In this routing algorithm, for each broadcast area, the sender node calculates the fitness value of each neighbor node, and then selects the node with the largest fitness value as the relay node.

4.1 Reliability simulation under different networking modes

According to the given reliability parameters, the common networking mode of distribution network communication EPON is simulated. Assume that the number of ONUs connected to the backbone fiber in the three networking modes is all 10. Due to the limitation of optical power loss, all ONUs cannot be cascaded on the same fiber core. The ONUs needs to be grouped: starting from the first ONU, every 5 ONUs that are geographically adjacent. Reliability simulation is carried out according to the EPON optical fiber wiring diagram of the three networking modes. Figure 5 is the reliability simulation result.

OPEN IN VIEWER

Fig. 5

Reliability simulation results of different networking.

Regardless of the networking mode, all ONUs can communicate normally, that is, the profust reliability when the system is in a fuzzy state is the highest.

When the fuzzy success state of the system is in the range of [0,0.8], the profust reliability of the single-point radiation network and the double-point double T network in these states is much greater than the profust reliability of the hand-in-hand protection network in this state. This is because the hand-in-hand protection of the networked optical cable is in different directions and can resist the failure of a single point of failure. When a certain section of the optical cable fails, the ONUs of the single-point radiation network and the double-point double-T network cannot communicate normally after the failure point. The actual data illustrates the superiority of hand-in-hand protection networking.

When the fuzzy success status of the system is 0, that is to say, when all ONUs cannot communicate normally, the profust reliability of the single-point spoke network in this state is much higher than that of the double-point double-T network. The double-T network backs up the OLT equipment, so it is less likely to happen that all ONUs fails.

According to the given reliability parameters, the common networking mode of distribution network communication EPON is simulated. Assume that the number of ONUs connected to the backbone fiber in the three networking modes is 1-10 respectively. Reliability simulation is carried out according to the optical fiber wiring diagram under different number of ONUs in three networking modes. We observe the reliability of normal communication of all ONUs, and the results are shown in Fig. 6.

OPEN IN VIEWER

Fig. 6

Reliability simulation results under different networking scales.

It can be seen from Fig. 6 that under the same protection mode, when the number of ONUs connected in the system is the largest, the availability of the system is not the highest. This method is more in line with the actual situation, because the more ONUs that are connected, the more components in the system, and as a whole, the probability of component failure increases. At the same time, the network scale has a smaller impact on the reliability of the hand-in-hand protection network than on the single-point radiation network and the double-point double T network. Although full link protection is performed on the entire network, the optical direction of the hand-in-hand protection network is different, which is more able to resist the failure of a single node.

In order to observe the influence of component failure rate changes on the reliability of the networking, the number of ONUs connected to the backbone fiber in the three networking modes is 10, and the maintenance time parameter MTTR of the components is 8 h. We observe the reliability of normal communication of all ONUs, and the results are shown in Fig. 7.

OPEN IN VIEWER

Fig. 7

The influence of component failure rate on network reliability.

At the same time, in order to understand the impact of component repair time changes on the reliability of the network, under the same conditions, the repair time of each component is controlled to be 2 h, 5 h, 8 h, 11 h, and 14 h. We observe the reliability that all ONUs can communicate normally, that is, the system is in a fuzzy state, and the result is shown in Fig. 8.

OPEN IN VIEWER

Fig. 8

The impact of component repair time on network reliability.

As shown in Figs. 7 and 8, we have studied the influence of component failure rate and repair time on the reliability of different networking. In the definition of profust reliability of EPON in this article, the normal communication of ONU is taken as the reference object. With the change of POS failure rate and maintenance time, the reliability of the three networking modes basically changes. This is because POS has a low failure rate compared to other components due to its passive characteristics. Therefore, fluctuations in the normal range of POS failure rate and repair time have little impact on overall reliability.

The significance of the above results for the project is:

Because the POS failure rate and the fluctuation of the normal range of maintenance time have little effect on the three networking modes, all when the ONU needs to be grouped based on fiber core resources due to optical power loss or future expansion restrictions, only a suitable group can be configured, without considering the influence of the grouping on the reliability.