JP4866950B2 - Information task control method using medium access control (MAC) method of optical fiber network time division multiple access - Google Patents

Description

The present invention can effectively and reasonably distribute the media bandwidth access rights of nodes of an optical fiber network to the nodes, and can be applied to an area network, an urban network, a public network, etc. can reduce the construction cost, to a medium access control of the optical fiber network time division multiple access can be improved broadband utilization information task control scheme using (MAC) how (system / method).

  According to the Hartley-Shannon theorem (path / information capacity theorem), the capacity of a path for transmitting media is regulated by the bandwidth of the transmission medium and the signal-to-noise ratio. Thus, a transmission medium having a high bandwidth and a high signal to noise ratio has a larger path capacity. Compared to wire and optical fiber, the optical fiber has much less noise and the optical fiber has a higher bandwidth than the wire, so the optical fiber has a much larger passage capacity than the wire. As a result, the construction of the communication network employs optical fibers instead of wires.

  It is very easy to reduce the construction cost because the bandwidth utilization of the network sharing medium is high and the network topping is very simple. Each type of access control step, for example, carrier sense multiple access (CAMA), carrier sense multiple access with collision detection (CAMA / CD), token ring, A token bus, a time-division multiple access (TDMA), and the like are applied to a network sharing a medium and perform access control of a network node.

  In computer communications, many regional networks currently use a multiple access protocol of carrier sense multiple access with collision detection (CAMA / CD). The applicable medium of this protocol is a wire. At the present time when Internet services are actively developed, it is strongly required to increase the passage capacity. However, in the area network that uses wires, the transmission distance and bandwidth are regulated, and if the passage capacity is increased, the transmission distance is increased. May be shortened and communication quality may be deteriorated. As a result, the construction of communication networks continues to employ optical fibers instead of wires.

  DQBD network (Distributed-queue Dual-bus, IEEE 802.6) protocol was submitted in 1990 to achieve the goal of service alignment, under the demand of digital development and service alignment. LAN (Local Area Networks) or MAN (Metropolitan Area Networks) standards. Such a network is a two-bus toppo and combines time-division multiple access (TDMA) technology to provide a consistent service.

  Such network protocols have the ability to coordinate information tasks that are immediate (eg, voice or multimedia information tasks) and non-immediate (eg, data information tasks). Since the transmission medium is an optical fiber, the code error rate is extremely low, it can be applied to a high-speed communication service, and the quality of service is extremely high. Such networks employ time-division multiple access (TDMA), so packets do not collide, bandwidth utilization is high, and information tasks provided by the network exceed the path capacity. A high bandwidth utilization rate approaching the path capacity can be reached, and since this is a two-bus topo, a full-duplex service can be achieved.

  The DQDB network has characteristics that the above-mentioned area network or urban network does not have, but it distributes the media access rights of the nodes of the network to the medium access control (MAC) for non-immediate information tasks. There was a problem that was unfair. Due to such problems, in the TDMA optical fiber network, the media access right of the upstream node is higher than the media access right of the downstream node. As a result, users who have an independent relationship with each other in the same network have a dispute about the right to use the bandwidth.

  When information (Messages) is generated in each node (Nodes) of the DQDB network, this information is divided into a number of sections, and each section is made a cell (Cells) carrier one by one. Wait for one to be saved in the buffer and transmitted. Buffers in the node are divided into two types, distributed queues (DQ) and local queues (LQ). A buffer that belongs to a distributed queue in a node has a save length of only one cell (53 bits), and the material is temporarily stored in this buffer before the node writes the material to an empty slot on the bus. It is saved and the timing of writing is controlled by DQSM (DQDB State Mechanism). The local queue is a temporary save process when a cell is generated, and the cell that is saved the earliest in the queue is transferred to the distributed queue and waits for transmission. The time from when a cell enters the distributed queue from the local queue to just before writing this cell to an empty slot on the bus is referred to as the distributed queue delay time (DQ delay). The waiting time from when a cell is generated until it enters the distributed queue from the local queue is referred to as the local queue delay time (LQ delay). The DQ average delay and LQ average delay of a node vary depending on the position of the node in the network (distance from the slot generator) and the amount of information task generated. The fact that the cell delay time varies depending on the distance between the node and the slot generator is the above-mentioned unfair phenomenon of access.

  This problem was actively discussed before and after 1990, and many control calculation methods were submitted to improve the above-mentioned phenomenon of unfair access. These control methods were divided into two types depending on the improvement measures used. Broadly divided. One is a countermeasure for slot reuse, and the other is an access control calculation method for improving the distributed queue. The slot reuse strategy removes the read and written slots by changing some of the original passive nodes to the primary node, increasing the available free slots on the nodes downstream of the bus, and The DQ delay time of the node can be shortened. The main purpose of the access control calculation method to improve the distributed queue is to allow the downstream node to obtain an empty slot immediately, and the upstream node is affected by too many transmission requests of the downstream node. Prevent information tasks from getting jammed. In order to achieve this goal, the access control calculation method needs to balance saving bandwidth and wasting bandwidth. The IEEE 802.6 DQDB MAC protocol employs a bandwidth balancing calculation method, which compensates for bandwidth savings, so that downstream node bandwidth access is unfair. Can be improved. IEEE 802.6's DQDB MAC protocol uses bandwidth balancing calculation method to compensate for bandwidth savings, but thorough the problem of unfair bandwidth access to downstream nodes Cannot be improved.

  The problem of unfair bandwidth access of downstream nodes affects the delay time of distributed queues at each node in the DQDB network. The queue consists of multiple buffers. Note that the “distributed queue” is a buffer of cells or packets in the queue distributed to each node. For the DQDB network, each node is a distributed queue. Cell buffer. Each node applies the same MAC protocol and uniquely determines when to write the contents of the cells saved in the distributed queue buffer to an empty slot in the fiber optic bus. Each node decides its own buffer buffer cell write time, but the MAC protocol to be applied must achieve the distributed queue first-come-first-serve (FCFS) service request It is. The network uses this feature to determine the order in which each node transmits data over the bus, that is, the network controls the distributed queue using a distributed control protocol, and the node that applies for transmission of the material first Free slots on the bus can be used more quickly.

  In order to solve the problem of unfair distribution of access rights by understanding the delay characteristics of the distributed queues of the nodes in the DQDB network, several methods of analyzing the DQDB network have been submitted, Based on Bisdikian's an approximate single-node analytical model. In this model, when each node wants to transmit a packet, the “steady-state generation function” generates a “number of transmission request needs” to be transmitted to that node in the downstream node. After this node has obtained the required number of transmission requests, it is possible to easily determine the delay of the distributed queue to which the node's packets are to be transmitted. According to the results of the analysis of such a model, it can be proved that the delay of the distributed queue of nodes in the DQDB network varies with the position of the nodes in the network. In short, conventional analysis methods perform distributed queue delay analysis, taking into account the details of the operation of the MAC protocol. Since the MAC protocol of the DQDB network is very complex, it is very difficult to analyze the queue delay by modularizing this protocol. According to such a method based on the MAC protocol, it is impossible to accurately analyze the DQ delay, and thus the problem of unfair access right distribution cannot be solved.

  From the perspective of the TDMA network, the delay of the distributed queue in the DQDB network is the same as the latency in the TDMA network. The waiting time in the TDMA network indicates the time from when a packet to be transmitted by a node enters a buffer connected to the transmission system of the queued bus to when the packet is written in an empty slot of the bus transmission system. This definition is similar to the distributed queue delay in DQDB networks. Thereby, the delay of the distributed queue in the DQDB network and the waiting time in the TDMA network have the same property. If the latency in a TDMA network varies according to the network topo, the average latency of the nodes in the TDMA network is a function of the position of the nodes, and so the problem with optical fiber TDMA networks is that the distribution of access rights is unfair. It always exists. If the above temporary characteristics actually exist, the problem of unfair distribution of access rights in the DQDB network cannot be solved. On the other hand, if the above-mentioned temporary characteristics do not actually exist, the problem of unfair distribution of access rights in the DQDB network can be solved after analyzing the waiting time in the TDMA network.

  For a stable TDMA network, whatever the MAC protocol is used, the load on the network should be the same as the load on the user's transmission request. Thus, latency analysis in a TDMA network can be completed by observing the slot usage in a TDMA system, without having to consider the details of the MAC protocol control operations. As is apparent from observing the usage of the TDMA system, the length of packet latency at a node is directly affected by the probability of reaching this node in the next available slot. This probability is determined by the distribution of information tasks at the nodes of the network 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 information task amount of the node. When the system capacity of TDMA is extremely large, the average waiting time of the node is inversely proportional to the information task amount of the analyzed node, that is, the larger the information task amount of the node, the shorter the average waiting time of the node. This shows that when the MAC protocol has the ability to control information tasks, the average latency of a node is not related to the location of the node in the network and the distance between the nodes. In other words, in a TDMA fiber optic network, there is no problem of unfair access right distribution when the MAC protocol has the ability to control information tasks.

  Thus, based on the above analysis, the applicant proposes a MAC protocol for fiber optic TDMA network, which determines the access control of the node by controlling the information task. In other words, the MAC protocol controls the information task to determine when to transmit the saved information bus in the node's queue. Then, the problem of unfair distribution of access rights among nodes in an optical fiber TDMA network can be solved thoroughly.

The main object of the present invention is to control the information task between the nodes in a fiber optic TDMA network by controlling the information tasks and by determining when to transmit the saved information bus in the node queue. It is possible to thoroughly solve the problem of unfair distribution of access rights, and it can be applied to regional network, urban network, public network, etc., it can reduce the construction cost of network and use broadband and to provide a medium access control of the optical fiber network time division multiple access, which can improve the rate (MAC) information task control method that utilizes how (system / method).

In the information task control system according to the first invention of the present application, when the time division multiple access of the optical fiber bandwidth is controlled by the node of the optical fiber network, the information task is performed by the medium access control (MAC) protocol (or method). utilizing medium access control (MAC) method of the control optical fiber network time division multiple access, the slot flow in the optical fiber network into a plurality of frames, each frame includes a 1 / T _B slots, T _B basic is an information task quantity, information tasks of the nodes is an integer multiple of T _B, when one frame reaches the first n nodes, and idle frames of information which the node is waiting for a transmission write to and have slots continuous, maximum number of information that can be written in succession, T (n) / T less than _B, also T (n) / T the same as _B, T (n) is the information task of the n nodes, the maximum number of information that can be written in succession, before the frame reaches, loaded reversed counting counter , when the number of information saved in the queue exceeds T (n) / T _B is the same maximum value of the reverse count counter and T (n) / T _B, the number of information saved in the queue When T (n) / T _B is not exceeded, the maximum value of the counter counter is the same as the number of information in the queue. Every time information is sent, the counter counter decrements by 1 and the counter counter When this node returns to 0, this node stops transmitting information, and the information saved in the queue is then transferred to the next information task control cycle as the next frame arrives. This is an information task control system characterized by restarting.

According to the information task control method using a medium access control (MAC) how the fiber optic network time division multiple access of the present invention, the following effects.
(1) The MAC protocol distributes access rights among nodes in a fiber optic TDMA network by controlling the information tasks and determining when to transmit the saved information bus in the node queue. Can solve the problem of being unfair.

  (2) The present invention can be applied to an area network, an urban network, a public network, and the like, can reduce the construction cost of the network, and can improve the broadband usage rate.

1 is a schematic diagram showing the construction of an optical fiber TDMA network according to an embodiment of the present invention. FIG. T is a curve diagram of the corresponding node average waiting time ₁ (n). To T ₂ (n) is a curve diagram of the corresponding node average latency. T is a curve diagram of the corresponding node average waiting time ₃ (n). It is an effect curve figure with respect to the 1st information task distribution model of the space | interval of a node. It is an effect curve figure with respect to the 2nd information task distribution model of the space | interval of a node. It is an effect curve figure with respect to the 3rd information task distribution model of the space | interval of a node.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
For a TDMA network, a node that wants to transmit information needs to notify the entire network by sending out a “transmission request” first and reserve an empty slot. The number of empty slots to be reserved is the same as the number of “transmission requests” sent out. When the number of empty slots to be reserved increases, the average waiting time of the node is shortened. If the average waiting time of a TDMA node is the standard for assessing slot access rights, nodes that generate a larger amount of information tasks have more access rights. In other words, the TDMA node takes away the access right of the medium due to the large amount of information task generated. The relationship between the average waiting time of a TDMA node and the information task amount of that node is derived as follows.

  The derivation of the average latency of the TDMA node is based on three network operating conditions. For one, the network being analyzed is in full load. Second, it is possible to reserve an empty slot with only one “transmission request”. Third, the access control of the node does not regulate the dispatch quantity of “transmission request”. Under these three network operation conditions, all nodes can freely compete and can take away the access right of an arbitrary slot. For a TDMA network that is stable and includes N nodes, R is the speed of the slot on the network bus, T (n) is the amount of information tasks generated by the nth node in the network, and n is the node And 0 ≦ n <N. Then T (n) is

r（n）は、第n個のノードに取られ情報を伝送するスロットの数量である。
第１個の操作条件と公式（１）とにおいて、全てのT（n）の和は、

r (n) is the number of slots taken by the nth node to transmit information.
In the first operating condition and formula (1), the sum of all T (n) is

  When viewed from the angle of the nth node, the slot in the transmission medium is divided into three types. These are used slots, reserved slots, and free slots. The used slot received upstream node information. The reserved slot transmits downstream node information. The free slot is a slot in which the nth node can be used. These free slots can be reserved for the nth node and may be reserved. In a stable network, the probability that a free slot appears in the nth node is the same as the information task amount of the nth node or larger than the information task amount of the nth node. Since the information task amount of the nth node is T (n) and this network is fully loaded, the probability that the nth node takes a free slot is T (n).

For the data section generated by the nth node, this data section enters the buffer of the connected bus and waits for transmission, and then the i-th slot of this node becomes a free slot. Temporary. Let p _w (n, i) be the probability that this source section will be written into the i-th slot, i = 1,2, ..., R, and p _w (n, i) is

  Let M (n) be the maximum waiting time of the nth node. Since the nth node uses r (n) slots per second, M (n) is

公式（４）に公式（１）のr（n）を代入すると、M（n）は、

Substituting r (n) of formula (1) into formula (4), M (n) becomes

  When μ (n) is the average waiting time of the nth node, μ (n) is

公式（６）に公式（５）のM（n）を代入すると、μ（n）は、

Substituting M (n) of formula (5) into formula (6), μ (n) becomes

  If the slot speed R of the network is constant, the formula (7) shows that the average waiting time of the nth node changes according to only the information task amount of the nth node, and the location of the node and each node It shows that it is not related to the distance between, and is not related to the MAC protocol that promotes the share of the medium.

  For high-speed networks, R is close to infinity. As a result, μ (n) in a high-speed network is

公式（８）によると、高速TDMAノードの平均待ち時間は、ノードの情報タスク量と逆比例する。簡単に言うと、高速TDMAネットワークにとって、ノードの情報タスク量が多いほど、平均待ち時間が短い。すなわち、高速TDMAネットワークにおいて、ノードの平均待ち時間はネットワークのトポと関係しない。言い換えると、高速TDMAノードの媒体アクセス権は、ノードの情報タスク量に従って変化する。このように、使用するMACプロトコルが適当な情報タスク制御能力を含めば、このネットワークの媒体アクセス権は、ノードらに有効で合理に分配されることができる。

According to the formula (8), the average waiting time of the high-speed TDMA node is inversely proportional to the information task amount of the node. Simply put, for high-speed TDMA networks, the higher the amount of information tasks in a node, the shorter the average waiting time. That is, in a high-speed TDMA network, the average waiting time of the node is not related to the network topology. In other words, the medium access right of the high-speed TDMA node changes according to the information task amount of the node. Thus, if the MAC protocol used includes the appropriate information task control capability, the media access rights of this network can be effectively and reasonably distributed to the nodes.

  Since the optical fiber network is a high-speed network having a high capacity, the characteristic that the average waiting time of the high-speed TDMA node is inversely proportional to the information task amount of the node is extremely natural for the TDMA optical fiber network. Thus, if the MAC protocol of the TDMA optical fiber network includes appropriate information task control capability, the media access rights of this network can be distributed fairly and reasonably to the nodes.

  In order to demonstrate the above-described invention, the present invention will be described by taking an example of applying access control by applying a MAC having information task control capability to a time division multiple access network.

  In order to test the effectiveness of the MAC control method of a TDMA network node in a fiber optic network, a temporary provision of some network operating conditions is submitted, which makes it convenient to simulate a fiber optic TDMA network. FIG. 1 is a schematic diagram showing the construction of an optical fiber TDMA network according to an embodiment of the present invention. The medium between the slot generators and the slot terminators is a single optical fiber. The slot flow in the optical fiber is pumped out of the slot generator and sinks into the slot terminator. All networks have N nodes. All nodes are numbered sequentially from left to right. The number of each node is also set as the position of the topo node in all networks. The time interval for sending one slot into the optical fiber by the slot generator is called one slot time. In addition, the interval related to the adjacent node and the operating conditions such as the length of information and the distribution of information tasks among the nodes will be described in detail below.

  In a simulation script, the spacing between each pair of adjacent nodes is the same, and this spacing is an integer multiple of the slot length. In order to check whether the average waiting time of the node changes according to the node interval, the interval between the nodes in the simulation script changes depending on the script, but the distribution of the information task of the original node does not change. In all simulation scripts, the length of information is constant and the length of each information is the same as the carrying length of one slot. Node information task distribution affects changes in node average latency, so information task distribution varies from script to script, thereby examining the “ideal fair behavior” characteristics of fiber optic TDMA networks Can do.

In order to easily complete control of information tasks, one basic information task amount is used as a basis for distributing information tasks. Let T _B be the basic information task amount. The amount of T _B is determined by the distribution of information tasks in each script. In each script, information tasks of any node is an integer multiple of T _B, in other words, the smallest possible node information task quantity in screenplay is T _B. To quote T _B, the distribution of information tasks in each type of optical fibers TDMA network can be easily defined, control information tasks can also be adjusted easily to the distribution of each type of information task.

Examples of MAC operations that can control information tasks are described below. Information task control MAC operations are completed by frame. The slot flow in the optical fiber is divided into a plurality of frames. Each frame includes 1 / T _B slots. When one slot reaches the nth node, information that this node is waiting for transmission is continuously written in empty slots in the frame, and the maximum information quantity that can be continuously written is T (n) / It should be smaller than T _B or equal to T (n) / T _B, and T (n) is the information task amount of the nth node. This maximum continuously writable information quantity is loaded into the inverse counting counter before the frame arrives. When saved Qty queue is greater than T (n) / T _B, the maximum value of the reverse count the counter is the same as T (n) / T _B. When the information quantity saved in the queue is not greater than T (n) / T _B , the maximum value of the counter counter should be the same as the information quantity in the queue. Each time one piece of information is sent out, the inverse counting counter is decremented by one. When the reverse count counter reaches 0, this node immediately stops transmitting information. At this time, the information saved in the queue is the information task control as described above when the next frame arrives. Restart the cycle.

  The purpose of simulating this network is to test whether an optical fiber TDMA network meets the characteristics of "ideal fair behavior". The fiber optic TDMA network to be inspected needs to control access through a MAC protocol that has the ability to control information tasks. “Ideal fair behavior” is based on the premise that the information task amount of a node does not change, and the average waiting time of the node of the optical fiber TDMA network changes according to the related position in the network of the node. The related position includes the position of the node and the interval between the nodes. Thus, the distribution and spacing of information tasks among nodes should change according to the simulation script.

On the other hand, as shown in formula (8), the node average waiting time obtained by the analysis is used for confirming the result of the simulation. If D _rms is the effective value of the error in the average node waiting time between analysis and simulation, then D _rms is

公式におけるμ_s(n)とμ(n)とは、それぞれシミュレーションと分析とによる第n個のノードの平均待ち時間とされる。

Μ _s (n) and μ (n) in the formula are the average waiting time of the nth node by simulation and analysis, respectively.

  In the drawings showing the simulation results below, the numbers on the horizontal axis are the series numbers of the nodes. Since the node series numbers are discrete numbers, all the curves in the drawing are composed of short straight lines. The unit of average node waiting time on the vertical axis is indicated by slot time.

In order to understand the influence of the information task amount of the node of the optical fiber TDMA network on the average node waiting time and the “ideal fair behavior”, three types of information task models are used for simulation. In this three information task model, information tasks of the nodes is an integer multiple of T _B. Since the optical fiber transmits information in one direction and the information of all nodes is not sent out from this simulation network, the (N-1) th node does not generate information task amount in all information task distribution models. . In the first information task distribution model, if T ₁ (n) is the nth node as the information task amount,

  In the above definition, the node approaching the slot generator has the maximum information task amount. For other nodes, the larger the node number, the smaller the information task amount. Based on this definition, the maximum number of slots that can be taken by the nth node in one frame is (N−n−1). Since this network is full loaded,

In accordance with the first information task distribution model, the number of slots included in one frame is (F ₁ ).

  The result of the simulation based on the first information task distribution model is shown in FIG. This simulation network includes 40 nodes. In FIG. 2, a solid line indicates a change in the average node waiting time due to the analysis, and a broken line indicates a change in the average node waiting time due to the simulation. The two curves almost completely overlap. The effective value of the error of the average node waiting time shown by the two curves is

  The effective value of this error is almost zero, which confirms that the inference of the node average latency of the present invention is accurate and that the simulation for the fiber optic TDMA network is very accurate.

In the second information task distribution model, a node has a larger amount of information task as the series number of the node is larger. If T ₂ (n) is the information task amount of the nth node,

In this definition, the node approaching the slot generator has the smallest amount of information task. The other nodes have a larger amount of information tasks as the node number is larger. Based on this definition, the maximum number of slots that can be taken by the nth node in one frame is (n + 1). The slot quantity F ₂ included in one frame is the same as F ₁ , and both are (N (N−1)) / 2. FIG. 3 is a schematic diagram of changes in the average node waiting time of the network of the second information task distribution model. In FIG. 3, a solid line and a broken line indicate changes in the node average waiting time due to analysis and simulation, respectively. The two curves almost completely overlap. The effective value of the node average waiting time error indicated by the two curves is 0.047958. If n ₁ is the series number of the node in FIG. 2 and n ₂ is the series number of the node in FIG. 3, comparing the changes between FIG. 2 and FIG. 3, the result is whether n ₁ is the same as n _2. Nevertheless, as long as T ₁ (n ₁ ) is the same as T ₂ (n ₂ ), the node average waiting time of n ₁ and n ₂ is always the same. As is clear from this phenomenon, if the MAC protocol of the fiber optic TDMA network has the ability to control information tasks, the average latency of each node is only related to the node's own information task amount, and in the node's network Regardless of position.

According to the comparison of the results of the above two simulations, when the distribution of information tasks in the network is average, it can be predicted that the average waiting time of each node is the same. In order to test this prediction, it is assumed that the third information task distribution model is uniformly distributed. In the three information task distribution model, if T ₃ (n) of information tasks of the n _{nodes, T 3 (n) = 1} / (N-1) = a T _B. Each frame includes K (N-1) slots, where K is a natural number. In one frame, the maximum number of slots that each node can take is K. In this simulation, the selected K is 2. FIG. 4 shows the result of the simulation of information task distribution distributed on the average.

  In FIG. 4, a solid line and a broken line indicate changes in the node average waiting time due to analysis and simulation, respectively. The two curves almost completely overlap. The effective value of the error of the node average waiting time indicated by the two curves is 0.003766. The average waiting time of all nodes is almost the same. This is consistent with the prediction results.

  The following simulation examines the effect of the spacing between adjacent nodes on the node average latency. In the above three simulation scripts, the spacing between nodes is the length of one slot. In the following simulation scripts, the distance between nodes is the length of three slots, but the other operating conditions are the same as the above three simulation scripts. FIG. 5 is an effect curve diagram for the first information task distribution model of the node spacing, FIG. 6 is an effect curve diagram for the second information task distribution model of the node spacing, and FIG. 7 is the node spacing. It is an effect curve figure with respect to the 3rd information task distribution model. The distance between nodes is simulated using the length of one slot and the length of three slots.

  As shown in FIG. 5 to FIG. 7, the solid line is the result of simulation using the length of one slot for the interval between adjacent nodes, and the broken line is the slot for three slots between adjacent nodes. This is a result of simulation using the length of. The two curves almost completely overlap. The effective value of the node average waiting time error indicated by each curve in FIG. 5 is 0.001235, the effective value of the node average waiting time error indicated by each curve in FIG. 6 is 0.03755, and the node indicated by each curve in FIG. The effective value of the average waiting time error is 0.000176. The effective value of each error is almost zero. This indicates that the node average latency is not affected by the spacing between nodes.

  The present invention can be applied to an optical fiber network.

Claims (1)

Optical fiber network time division multiple access medium access control (MAC) for controlling information tasks with a medium access control (MAC) protocol (or method) when controlling optical fiber bandwidth time division multiple access by a node of an optical fiber network )
The slot flow in an optical fiber network is divided into a plurality of frames, each frame includes 1 / T _B slots, T _B is the basic information task amount, and the information task amount of each node is an integer multiple of T _B ,
When a frame reaches the nth node, information that this node is waiting for transmission is continuously written in the idle slot of the frame, and the maximum number of information that can be continuously written is T (n ) / T _B or the same as T (n) / T _B , T (n) is the information task amount of the nth node,
The maximum number of information that can be written continuously is loaded into the counter counter before the frame arrives,
When the number of information saved in the queue exceeds T (n) / T _B , the maximum value of the counter counter is the same as T (n) / T _B.
If the number of information that is saved in the queue does not exceed the T (n) / T _B is the same as the number of information of the maximum value in the queue for reverse counting counter, every time, every time to feed the information, reverse counting The counter decrements 1
When the backcount counter returns to 0, this node stops transmitting information, and at this time the information saved in the queue is the next information task as described above when the next frame arrives. An information task control system characterized by restarting a control cycle.

Images