CN101877670B - Optical fiber network media time-division multiple access method and its application traffic control method - Google Patents

Abstract

The invention relates to a time-sharing multiple access method for optical fiber network media and a traffic control mode applied by the same, which mainly enables an MAC protocol to carry out access control on network nodes using the same optical fiber in a traffic control mode so that the access right of the network nodes to the optical fiber bandwidth can be effectively and reasonably distributed among the nodes, the problem of unfair distribution of the access right of the time-sharing multiple access of the optical fiber network is solved, and the optical fiber network is promoted to be suitable for constructing various application networks such as a sub-network of an intercity network, a local area network or a public network, and the like so as to reduce the network construction cost and improve the broadband utilization rate.

Description

Optical fiber network media time-division multiple access method and its application traffic control method

technical field

The present invention relates to a time-division multiple access method for optical fiber network media and its application traffic control method. Improve the problem of unfair distribution of node bandwidth access rights in optical fiber networks, and promote time-division multiple access optical fiber networks to be suitable for constructing local area networks, intercity networks or sub-networks of public networks, etc., to reduce network construction costs and improve broadband usage Rate.

Background technique

According to the Hartley-Shannon theorem (channel/information capacity theorem), the channel capacity of a transmission medium is limited by the bandwidth and signal-to-noise ratio of the transmission medium used. A transmission medium with high bandwidth and high signal-to-noise ratio naturally has a large channel capacity. Comparing the two transmission media of stranded wire and optical fiber, because the signal-to-noise of optical fiber is very low, the bandwidth is much higher than that of stranded wire, resulting in a much higher channel capacity than stranded wire; therefore, in recent years, the transmission construction of communication networks has gradually Fiber optics instead of twisted wires.

Networks for media sharing naturally have high bandwidth utilization, and because their network topology is very simple, it is easy to achieve the goal of reducing construction costs. Various access control steps such as carrier sense multiple access (CSMA), carrier sense multiple access/collision detection (Carrier sense multiple access with collision detection, CSMA/CD), token ring (Token ring), Token bus and Time-division multiple access (TDMA), etc., are applied in the media sharing network as the access control of network nodes.

In terms of computer communication, most of the local area networks currently use the multiple access protocol of Carrier Sense Multiple Access/Collision Detection (CSMA/CD). The physical layer of this protocol is the strand. With the vigorous development of various services on the Internet today, the demand for expanding the channel capacity is bound to increase. The regional network using twisted wires at the physical layer is limited by the transmission distance and bandwidth. The development of channel capacity is not as good as that of optical fibers. Easy, and the increase in channel capacity will lead to shortened transmission distances and degraded communication quality. Therefore, under the strong demand of accelerating the expansion of channel capacity at the physical layer of the regional network in the future, it should be an inevitable trend to replace twisted wires with optical fibers.

Due to the development of digitalization and the demand for service integration, the DQDB network (Distributed-queueDual-bus, IEEE 802.6) protocol was proposed in 1990 as a LAN (Local Area Networks) for service integration purposes. ) or MAN (Metropolitan AreaNetworks) standard. This network is a dual-bus topology combined with the application of Time-division Multiple Access (TDMA) technology to provide integrated services.

This network protocol has the ability to integrate isochronous (such as voice or multimedia traffic) and asynchronous traffic (such as data traffic). Since the transmission medium is optical fiber, it has the characteristics of low bit error rate, which is suitable for the development of high-speed communication services and meets the needs of high service quality. The network adopts the multiple access method of TDMA, which has no packet collision problem and has high bandwidth utilization efficiency. Even when the traffic provided by the network is higher than the channel capacity, it can still achieve high bandwidth utilization efficiency close to the channel capacity, and because of the dual bus topology, this network naturally achieves full-duplex service requirements.

The DQDB network has the use characteristics that other regional networks or metropolitan networks do not have as mentioned above, but because of its Medium Access Control (MAC) for non-real-time (non-isochronous) traffic, it is derived The problem of unfair distribution of media access rights for network nodes. The unfair distribution of access rights between nodes makes the upstream nodes have higher access rights than the downstream nodes on the TDMA optical fiber network. Therefore, there is a dispute over the rights to use bandwidth between users who are under the same network but have independent relationships with each other.

On each network node (Nodes) of DQDB, when a message (Messages) is generated, the message will be divided into several small pieces, and each small piece will be sequentially regarded as the payload of the unit (Cells), and stored in the buffer one by one , waiting to be sent. The buffers in the nodes are divided into two types: Distributed Queues (DQ) and Local Queues (LQ). The buffer of the distributed queue in the node has only one unit (53 bytes) of storage length, which is the temporary storage position before the node writes the data into the bus empty slot (slot), controlled by DQSM (DQDB State Mechanism) Its write timing. The local queue is the temporary storage location after the unit is generated, and the unit stored in the queue at the earliest will be moved into the distributed queue to wait for transmission. The time from when a unit enters the distributed queue from the local queue to when the unit is written into the empty slot of the bus is called the delay time of the distributed queue (DQ delay). The waiting time between when a unit is generated and when it enters the distributed queue from the local queue is called the delay time of the local queue (LQ delay). The average DQ delay and LQ average delay of a node will vary due to the location of the node in the network (the distance from the slot generator) and the amount of traffic it generates. The situation that the cell delay time varies with the distance between the node and the slot generator is the aforementioned unfair access phenomenon.

This problem has been discussed enthusiastically around 1990, and many control algorithms have been proposed to improve the problem of unfair access. These control methods can be divided into two categories according to the improvement strategies they use. The first type is the strategy of slot reuse, and the second type is to improve the access control algorithm of distributed queues. The slot reuse strategy needs to change some of the original passive nodes into active nodes to erase the slots that have been read and written, so as to increase the number of available empty slots for the downstream nodes of the bus, and shorten the DQ delay time of the downstream nodes . The main purpose of the second type of improved access control algorithm is to enable downstream nodes to obtain empty slots in real time, and at the same time, it is necessary to prevent upstream nodes from being affected by excessive transmission requests from downstream nodes, resulting in traffic congestion. To achieve this goal, the access control algorithm needs to strike a balance between bandwidth reservation and bandwidth waste. The DQDB MAC protocol of IEEE 802.6 adopts the algorithm of Bandwidthbalancing, and compensates in the way of bandwidth reservation, so as to improve the problem of unfair bandwidth access of downstream nodes. Although the IEEE 802.6 MAC protocol uses the Bandwidth balancing algorithm to compensate for reserved bandwidth, the problem of unfair access has not been completely resolved, and it cannot provide a fair and reasonable distribution of access rights. independent user.

The unfair distribution of this access right affects the delay of the distributed queue of each node in the DQDB network. The queue is composed of multiple buffers, and the so-called "distributed queue" means that the buffers of the units or packets in a queue are scattered among the nodes. For the DQDB network, each node has a distributed queue unit buffer. Using the same MAC protocol, each node independently determines the timing when the unit content stored in its distributed queue buffer is written into the empty slot on the optical fiber bus. Although each node independently determines the writing time of its cell buffer belonging to the distributed queue, the applied MAC protocol must meet the service requirement of the distributed queue First-Come-Fist-Serve (FCFS). The network uses this feature to determine the order in which nodes use the bus to transmit data. That is to say, the network provides a set of distributed control protocols to control the distributed queue, so that the nodes that apply for data transmission earlier can use the bus earlier. empty slot.

In order to understand the characteristics of the distributed queue delay of DQDB nodes and solve the problem of unfair distribution of access rights, several methods have been used for the analysis of DQDB networks. These methods are based on the approximate single-node analysis model proposed by Bisdikian (an approximate single-node analytical model). In this mode, each node uses a steady-state generation function (thesteady-state generation function) to generate the number of "transmission requests" sent to the node by its downstream nodes when there are packets to be transmitted. After the node obtains the demand number, it can easily determine the distributed queue delay of the packet to be transmitted in the node. The analysis results of this model prove that the distributed queue delay of each node in the DQDB network will vary with the location of the node in the network. In short, the existing analysis method is to consider the operation details of the MAC to analyze the delay of the distributed queue. Since the MAC protocol of the DQDB network is very complex, it is quite difficult to model this protocol and complete the queue delay analysis. It is an impossible task to obtain accurate DQ delay analysis based on the operation of the MAC protocol. Therefore, this method of analysis cannot contribute to solving the problem of unequal distribution of access rights.

From the perspective of a TDMA network, the distributed queue delay on the DQDB network is equivalent to the waiting time on the TDMA network. The waiting time on the TDMA network refers to the period from when the packet to be transmitted in the node enters the buffer connected to the bus transmission system in the queue to when the packet is written into the idle slot of the bus transmission system. This definition is the same as the latency of distributed queues on the DQDB network. Therefore, the delay of the distributed queue on the DQDB network has the same property as the waiting time on the TDMA network. If the waiting time on the TDMA network will change with the network topology, that is, the average waiting time of the nodes in the TDMA network is a function of the node position, then the optical fiber TDMA network will be inherently accompanied by the problem of unfair distribution of access rights. are indivisible. If the properties of the above assumptions can be confirmed to exist, it will be impossible to completely eliminate the problem of unfair distribution of access rights on the DQDB network. On the contrary, if the nature of the above assumptions and their existence cannot be verified, then after the latency on the TDMA network is analyzed, the problem of unfair distribution of access rights on the DQDB network should be able to obtain an appropriate solution.

For a stable TDMA network, no matter which MAC protocol is used, the load borne by the network should be equal to the load requested by the user. Therefore, the analysis of the latency on the TDMA network can be done by observing the status of the slots being used on the TDMA system, without considering the details of the control operation of the MAC protocol. From the observation of the usage status of this TDMA system, it can be seen that the waiting time of a packet on a certain node is directly affected by the probability that the next available slot will arrive at this node. And this probability can be determined by the traffic distribution of the network nodes and the capacity of the transmission medium.

The analysis results show that the average waiting time of a TDMA node is a function of the traffic volume of the node. When the TDMA system capacity is very large, the average waiting time of its nodes is inversely proportional to the traffic volume of the analyzed node, that is, the larger the traffic volume transmitted by a node, the shorter its average waiting time. According to this inference, if the MAC protocol has the ability of traffic control, the average waiting time of nodes will have nothing to do with the position of nodes in the network and the distance between nodes. In other words, if the MAC of the TDMA optical fiber network has the operation capability of traffic control, the problem of unfair allocation of access rights will not exist.

In view of the above analysis situation, the present inventor proposes that the MAC protocol of the optical fiber TDMA network needs to determine the access control of nodes through the operation of traffic control. In other words, the MAC protocol determines the timing when the information stored in the node queue can be transmitted to the bus through the operation of traffic control, so as to completely solve the problem of unfair distribution of access rights between nodes in the optical fiber TDMA network.

Contents of the invention

The main purpose of the present invention is to provide a method for time-division multiple access of optical fiber network media and the traffic control method of its application. Media access rights can be effectively and fairly distributed among nodes to improve the problem of unfair distribution of access rights to nodes in optical fiber TDMA networks, so that the optical fiber TDMA network can be used to construct sub-networks of regional networks, metropolitan networks or public networks And so on, in order to reduce the cost of network construction, improve broadband utilization and network service quality.

Description of drawings

Figure 1 is a schematic diagram of the architecture of an optical fiber TDMA network.

Fig. 2 is a curve diagram of the average waiting time of nodes corresponding to T ₁ (n).

Fig. 3 is a curve diagram of the average waiting time of nodes corresponding to T ₂ (n).

Fig. 4 is a graph of the average waiting time of nodes corresponding to T ₃ (n).

FIG. 5 is a graph showing the effect of node spacing on the first traffic allocation mode.

FIG. 6 is a graph showing the effect of node spacing on the second traffic allocation mode.

Fig. 7 is a graph showing the effect of node spacing on the third traffic allocation mode.

Detailed ways

For a TDMA network, a node that wants to transmit information must first send a "transmission request" to notify the entire network to reserve a free slot. The number of reserved free slots corresponds to the number of sent "transmission requests". When the number of reserved idle slots increases, the average waiting time of nodes will decrease accordingly. Therefore, if a node has higher traffic, its average waiting time will be smaller. If the average waiting time of a TDMA node is used as a criterion for evaluating the slot access rights it has, then the nodes that generate a larger amount of traffic will have more access rights. In other words, TDMA nodes compete with each other according to the amount of traffic they generate, so as to gain access to the media. The relationship between the average waiting time of a TDMA node and its node traffic can be derived as follows.

The derivation of the average waiting time of TDMA nodes is based on three network operating conditions. The first operating condition indicates that the analyzed network is fully loaded. Another operating condition restricts each "transmission request" to reserve only one free slot. The third operating condition assumes that the node's access control operation cannot limit the number of its "transfer request" sent. Because of these three operating conditions, all nodes can freely compete for access to any channel. For a stable TDMA network with N nodes, let R represent the speed of the slots on the network bus, T(n) represent the traffic generated by the nth node in the network, where n is the serial number of the node, and 0≤n<N. Then T(n) can be expressed as follows:

T(n)=r(n)/R, (1)

In the formula, r(n) is the number of slots captured by the nth node and used to transmit information every second.

Due to the first operating condition and formula (1), the sum of all T(n) can be expressed as follows:

$\underset{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\mathrm{<span\; class="notranslate">\; T</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$

From the point of view of the nth node, the slots on the transmission medium can be divided into three types. These three types are used slots, reserved slots and free slots. The used slots carry the information of the upstream nodes, the reserved slots are used to transmit the information of the downstream nodes, and the free slots are the available slots for the nth node. These free slots may or may not have been reserved by the nth node. In a stable network, the probability of a free slot appearing at the nth node should be equal to or greater than the traffic volume of the nth node. Since the traffic volume of the nth node is T(n), and the network is fully loaded, the probability of the nth node capturing a free slot is also T(n).

As for a data segment generated by the nth node, assuming that the data segment enters the buffer of the connection bus and enters the state of waiting to be transmitted, the ith slot of this node is a free slot. Let p _w (n, i) represent the probability that this data segment is written into the i-th channel, where i=1, 2, . . . , R, then p _w (n, i) can be expressed as follows:

p _W (n, i)=T(n)(1-T(n)) ^i-1 (3)

Let M(n) represent the maximum waiting time of the nth node. Since the nth node uses r(n) slots per second, M(n) can be expressed as follows:

M(n)＝R-r(n)+1 (4)

Substituting r(n) of formula (1) into formula (4), M(n) can be organized as follows:

M(n)＝R(1-T(n))+1 (5)

If μ(n) is used to represent the average waiting time of the nth node. This average waiting time can be calculated as follows:

$\mathrm{<span\; class="notranslate">\; \&mu;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; ip</span>}}_{\mathrm{<span\; class="notranslate">\; W</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span>{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; T</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}$

$<span\; class="notranslate">\; =</span><span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; -</span>{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; T</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; T</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span><span\; class="notranslate">\; /</span>\mathrm{<span\; class="notranslate">\; T</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 6</span>}<span\; class="notranslate">\; )</span>$

Substituting M(n) of formula (5) into formula (6), the average waiting time of the nth node can be rearranged as follows:

μ(n)=[1-(1-T(n)) ^R(1-T(n))+1 (1+(R(1-T(n))+1)T(n))]/ T(n)(7)

If the slot rate R of the network is fixed, equation (7) shows that the average waiting time of the nth node only changes with the traffic volume of the nth node, and has nothing to do with the location of the node and the distance between the nodes , and no correlation exists with the MAC protocol that enables media sharing.

For high-speed networks, R approaches infinity. Then μ(n) on the high-speed network can be deduced as follows:

$\underset{\mathrm{<span\; class="notranslate">\; R</span>}<span\; class="notranslate">\; \&Right\; Arrow;</span><span\; class="notranslate">\; \&infin;</span>}{\mathrm{<span\; class="notranslate">\; lim</span>}}\mathrm{<span\; class="notranslate">\; \&mu;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; R</span>}<span\; class="notranslate">\; \&Right\; Arrow;</span><span\; class="notranslate">\; \&infin;</span>}{\mathrm{<span\; class="notranslate">\; lim</span>}}<span\; class="notranslate">\; \{</span><span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; -</span>{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; T</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; R</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; T</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; +</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; R</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; T</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; T</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span><span\; class="notranslate">\; /</span>\mathrm{<span\; class="notranslate">\; T</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \}</span>$

$<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; /</span>\mathrm{<span\; class="notranslate">\; T</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 8</span>}<span\; class="notranslate">\; )</span>$

Equation (8) clearly shows that the average waiting time of a high-speed TDMA node is inversely proportional to the traffic volume of the node. In short, for a high-speed TDMA network, the greater the traffic volume of a node, the smaller its average waiting time will be. This property strongly shows that on high-speed TDMA networks, the average waiting time of a node is independent of the network topology. In other words, the media access right of a high-speed TDMA node varies with the traffic volume of the node. Therefore, if the MAC protocol used contains appropriate traffic control capabilities, the medium access rights of the network can be effectively and reasonably allocated among nodes.

Because the optical fiber network is a high-speed network with high capacity, the above-mentioned characteristic that the average waiting time of high-speed TDMA nodes is inversely proportional to the traffic volume of the nodes is an inherent characteristic of TDMA optical fiber network. Therefore, if the MAC protocol of the TDMA optical fiber network includes appropriate traffic control capabilities, the media access rights of this network can be effectively distributed among nodes fairly and reasonably.

In order to prove and illustrate the above-mentioned invention, the present invention takes a MAC with traffic control capability to be applied to a time-division multiple access optical fiber network for access control as an example for illustration and verification.

In order to test the inventive effect of the node MAC control method in optical fiber TDMA network, some assumptions of network working conditions are put forward in order to carry out the simulation of optical fiber TDMA network. Figure 1 shows the architecture of the optical fiber TDMA network. In the figure, the medium between slot generators and slot terminators is an optical fiber. The slot flow on the optical fiber is sent out by the slot generator, and finally sinks into the slot terminator. The entire network has N nodes. All nodes are numbered sequentially from left to right. The number of each node can also represent its node position in the entire network topology. The time interval for the slot generator to send a slot to the optical fiber is called a slot time. The distance that a slot extends on an optical fiber is called a slot length. In addition, those working conditions related to the distance between adjacent nodes, the length of messages and the distribution of traffic between nodes are described later.

In an analog embodiment, the interval between each pair of adjacent nodes is equal, and the interval is an integer multiple of the slot length. In order to confirm whether the average waiting time of nodes will change with the interval between nodes, the interval between nodes in the simulated embodiment will be changed with the embodiment, but the original traffic distribution of nodes will remain unchanged. In all simulated embodiments, the length of the messages remains constant, and the length of each message is equal to the length of a slot payload. The traffic distribution of the nodes affects the change of the average waiting time of the nodes, so the traffic distribution will change with the embodiment to test the characteristics of the "ideal fair behavior" of the optical fiber TDMA network.

In order to easily complete traffic control, a basic traffic volume will be used as the basis for traffic allocation. The basic traffic volume is expressed in _TB . The amount of _TB is determined by the traffic allocation of each embodiment. In an embodiment, the traffic volume of any node is an integer multiple of _TB , in other words, the smallest possible node traffic volume in the embodiment is _TB . Based on the reference of _TB , it is not only easy to define the traffic distribution on various optical fiber TDMA networks, but also the traffic control operation can be easily completed in conjunction with various traffic distributions.

An example of MAC operation that can accomplish traffic control is described below. MAC operations for traffic control are done using slot frames. The slot stream on the fiber is divided into several slot frames. Each slot frame contains 1/T _B slots. When a slot frame arrives at the nth node, the node can continuously write the information waiting to be transmitted into the free slot in the frame. The maximum number of messages that can be written continuously must be less than or equal to T(n)/ T _B , where T(n) is the traffic volume of the nth node. The maximum number of messages that can be written continuously will be loaded into a countdown counter before the slot frame arrives. When the number of messages stored in the queue is greater than T(n)/T _B , the maximum value of the countdown counter is equal to T(n)/T _B . Otherwise, the maximum value of the countdown counter shall be equal to the number of messages in the queue. Then, every time a message is sent, the countdown counter is decremented by one. Until the countdown counter returns to zero, the node must immediately stop the transmission of information. At this time, the information still stored in the queue needs to wait for the arrival of the next slot frame before restarting the next traffic control cycle as described above.

The purpose of this network simulation is mainly to test whether the optical fiber TDMA network conforms to the characteristics of "ideal fair behavior". The inspected optical fiber TDMA network needs to use the MAC protocol with traffic control capability for access control. "Ideal fair behavior" means that the average waiting time of nodes in the optical fiber TDMA network will not change with the relative position of the nodes in the network without changing the traffic volume of the nodes. This relative position includes the position of the nodes and the spacing between nodes. Therefore, the traffic distribution and spacing between nodes must be changed according to the simulated embodiment.

On the other hand, the average waiting time of nodes generated by the analysis, represented by Equation (8), will be used to confirm the simulation results. If D _rms is used to represent the root mean square error between the average waiting time of the analyzed and simulated nodes, then D _rms can be defined as follows:

${\mathrm{<span\; class="notranslate">\; D.</span>}}_{\mathrm{<span\; class="notranslate">\; rms</span>}}<span\; class="notranslate">\; =</span>{<span\; class="notranslate">\; (</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}\underset{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&mu;</span>}}_{\mathrm{<span\; class="notranslate">\; the\; s</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; \&mu;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; /</span>\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 9</span>}<span\; class="notranslate">\; )</span>$

In the formula, μ _s (n) and μ (n) represent the average waiting time of the nth node obtained from simulation and analysis respectively.

In all the following figures showing simulation results, the number on the horizontal axis represents the ordinal number of the node. Because the serial numbers of nodes are discrete numbers, all curves in the graph are composed of small line segments. The unit of the average waiting time of nodes on the vertical axis is expressed in slot time.

In order to understand the traffic volume of the nodes in the optical fiber TDMA network, the impact on the average waiting time of the nodes, and the "ideal fair behavior", three traffic distribution modes are used in the simulation. In the three traffic allocation modes, the traffic volume of each node is an integer multiple of _TB . Because the optical fiber transmits information in one direction, and the information of all nodes is not sent out of the simulated network, in all traffic distribution modes, the (N-1)th node does not generate traffic. In the first traffic allocation mode, let T ₁ (n) represent the traffic volume of the nth node, then T ₁ (n) can be defined as follows:

T ₁ (n)=(Nn-1)T _B , n=0, 1, . . . , N-2(10)

In this definition, the nodes closest to the slot generator have the greatest amount of traffic. The traffic volume of other nodes decreases as the node number increases. Based on this definition, within a slot frame, the maximum number of slots that the nth node can capture is (N-n-1). Because the network is fully loaded, the

$\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; T</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; T</span>}}_{\mathrm{<span\; class="notranslate">\; B</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; T</span>}}_{\mathrm{<span\; class="notranslate">\; B</span>}}<span\; class="notranslate">\; /</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 11</span>}<span\; class="notranslate">\; )</span>$

Therefore, corresponding to the first traffic allocation mode, the number of slots (F ₁ ) contained in a slot frame is:

F ₁ =1/T _B =N(N-1)/2 (12)

Based on the first traffic allocation model, the corresponding simulation results are shown in Fig. 2. This simulated network contains forty nodes. The solid line in the figure represents the variation of the average waiting time of nodes obtained from the analysis and calculation, and the dotted line represents the variation of the average waiting time of nodes obtained from the simulation. Both curves are almost completely heavy base. The root mean square error of the average waiting time of nodes represented by the two curves is calculated as follows;

${\mathrm{<span\; class="notranslate">\; D.</span>}}_{\mathrm{<span\; class="notranslate">\; rms</span>}}<span\; class="notranslate">\; =</span>{<span\; class="notranslate">\; (</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}\underset{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&mu;</span>}}_{\mathrm{<span\; class="notranslate">\; the\; s</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; \&mu;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; /</span>\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0.073485</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 13</span>}<span\; class="notranslate">\; )</span>$

The root mean square error value is close to zero. This result not only confirms the correctness of the deduction of the average waiting time of nodes in this paper, but also confirms the accuracy of the simulation of the optical fiber TDMA network.

In the second traffic allocation mode, the traffic volume of nodes increases sequentially with the node serial number. Let T ₂ (n) represent the traffic volume of the nth node, and the volume of T ₂ (n) can be defined as follows:

T ₂ (n)=(n+1)T _B , n=0, 1, . . . , N-2 (14)

In this definition, the nodes closest to the slot generator have the least amount of traffic. The traffic volume of other nodes increases with the increase of the node serial number. Based on this definition, within a slot frame, the maximum number of slots that the nth node can capture is (n+1). The number of slots F ₂ contained in a slot frame is the same as that of F ₁ , which is (N(N-1))/2. Fig. 3 shows the variation of the average waiting time of the nodes for the network with the second traffic allocation mode. The solid line and dotted line in the figure respectively represent the change of the average waiting time of the nodes in the analysis and simulation. Both curves are almost completely heavy base. The root mean square error between the average waiting time of nodes represented by the two curves is 0.047958. If n ₁ and n ₂ represent the node numbers in Figure 2 and Figure 3 respectively, then compare the changes in Figure 2 and Figure 3 . The comparison results show that no matter whether n ₁ is equal to n ₂ or not, as long as T ₁ (n ₁ ) is equal to T ₂ (n ₂ ), the average waiting time of nodes n ₁ and n ₂ must be equal. This phenomenon shows that when the MAC protocol of the optical fiber TDMA network has traffic control capabilities, the average waiting time of each node is only related to the traffic volume of the node itself, and has nothing to do with the network location of the node.

From the comparison of the above two simulation results, it can be predicted that when the network traffic is evenly distributed, the average waiting time of each node will be equal to each other. To test this prediction, assume that the third traffic allocation pattern is a uniform distribution. Let T ₃ (n) represent the traffic volume of the nth node in the third traffic allocation mode, then for all nodes, T ₃ (n)=1/(N−1)=T _B . Each slot frame contains K(N-1) slots, where K is a positive integer. In a slot frame, the maximum number of slots that each node can capture is K. In this simulation, K was chosen equal to 2. Figure 4 shows the simulation results of evenly distributed traffic distribution.

In Fig. 4, the solid line and the dotted line represent the variation of the average waiting time of nodes in analysis and simulation respectively. Both curves are almost completely heavy base. The root mean square error between the average waiting time of the nodes represented by the two curves is 0.003766. The average waiting time for all nodes is almost the same. This is consistent with the predicted results.

The following simulation is used to examine the effect of the distance between adjacent nodes on the average waiting time of nodes. In the previous three simulation examples, the interval between nodes is the length of one slot. In the subsequent simulation embodiments, the interval between nodes is increased to three slot lengths, while other working conditions remain the same as the previous three embodiments. Fig. 5, Fig. 6 and Fig. 7 are simulation results respectively corresponding to three traffic distribution modes, and the interval between nodes is one and three slot lengths.

In Figure 5, Figure 6 and Figure 7, the solid line represents the interval between adjacent nodes, which is the simulation result of one slot length, and the dotted line represents the interval between adjacent nodes, which is the simulation result of three slot lengths . Both curves are almost completely heavy base. Corresponding to Figure 5, Figure 6 and Figure 7, the root mean square error between the average waiting time of nodes represented by the two curves in each figure is 0.001235, 0.03755 and 0.000176 respectively. Each root mean square error value tends to zero. This represents the average waiting time of nodes and will not be affected by the size of the node interval.

The above simulation results show that the average waiting time of nodes has nothing to do with the network topology in the optical fiber TDMA network with the traffic control capability of the MAC protocol. Therefore, through the method of the present invention, the media access rights of the optical fiber network can be effectively and reasonably allocated among nodes, thereby improving the problem of unfair access time of the optical fiber network. The improved optical fiber TDMA network is suitable to be used to construct sub-networks of urban networks, regional networks or public networks to reduce network construction costs, improve broadband utilization and communication quality.

To sum up, the embodiment of the present invention can indeed achieve the expected use effects, and the specific functions disclosed by it have not only been seen in similar products, but also have not been disclosed before the application, and have fully complied with the provisions of the Patent Law. Therefore, I filed an application for an invention patent in accordance with the law, and I sincerely appreciate the convenience of reviewing and granting the patent.

Claims (1)

1. time-division multiple access method of optical fiber network media; It is characterized in that; Said method is to make the mode of MAC agreement with news affair control; To the control that conducts interviews of fiber network node, the be engaged in mode of control of wherein said news comprises: the news concentrated flow on the fiber optic network is divided into several news groove frames, and each news groove frame comprises 1/T _BIndividual news groove, wherein T _BThe basic news of representative affair amount, and the news affair amount of each node all is T _BIntegral multiple; When a news groove frame arrives n node, this node with its etc. transferred information, the free time that writes continuously in the frame is interrogated groove, the maximum information number that this writes out continuously must be less than or equal to T (n)/T _B, wherein T (n) is the news affair amount of n node; The said maximum information number that writes out continuously is loaded into a reciprocal counter before news groove frame arrives; Information number in being stored in formation is greater than T (n)/T _BThe time, the maximum of reciprocal counter will equal T (n)/T _BOtherwise the maximum of reciprocal counter should equal the information number in the formation; Then, whenever see an information off, reciprocal counter subtracts one; Make zero up to reciprocal counter, this node is the transmission of Stop message immediately, the information that still stores in formation this moment; Next news groove frame such as need to arrive, just can restart the control of being engaged in of next news.

Images