KR100584420B1 - Dynamic bandwidth allocation method for gigabit ethernet passive optical network - Google Patents

Abstract

The present invention enables a dynamic allocation of bandwidth for data transmission to an optical network unit (ONU) in a Gigabit Ethernet passive optical network (GE-PON). To this end, the present invention first compares an average of the total available bandwidth by the number of ONUs that have requested bandwidth, and compares the average bandwidth with the bandwidth required by each of the ONUs. Only the average bandwidth is allocated to ONUs that require more bandwidth. If all of the ONUs making the bandwidth request have the remaining bandwidth after the bandwidth allocation is completed, the remaining bandwidth is allocated additionally. At this time, the remaining bandwidth is allocated to ONUs that have been allocated only the average bandwidth in proportion to the size of each unallocated request bandwidth, or unallocated from the ONUs in which the required bandwidth is the largest among the ONUs having only the average bandwidth allocated. Allocate additional bandwidth until all remaining bandwidth is allocated for the required bandwidth.

Gigabit Ethernet, Passive Optical Subscriber Network (PON), Gigabit Ethernet Passive Optical Subscriber Network (GE-PON), Dynamic Bandwidth Allocation (DBA).

Description

DYNAMIC BANDWIDTH ALLOCATION METHOD FOR GIGABIT ETHERNET PASSIVE OPTICAL NETWORK}             

1 is a block diagram showing an example of a typical ATM-PON,

2 is a block diagram showing an example of a GE-PON to which the present invention is applied;

3 is a flowchart of dynamic bandwidth allocation according to the first embodiment of the present invention;

4 is a flow diagram of dynamic bandwidth allocation in accordance with a second embodiment of the present invention.

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a passive optical network (hereinafter referred to as "PON"), and more particularly to ONU (in Gigabit Ethernet Passive Optical Network (hereinafter referred to as "GE-PON")). The present invention relates to a method for allocating a bandwidth for data transmission to an optical network unit.

X-Digital Subscriber Line (xDSL), Hybrid Fiber Coax (HFC), Fiber To The Building (FTTB), Fiber To The Curb (FTTC), and Fiber To (FTTH) Various network structures and evolutionary methods such as The Home have been proposed. Among these various network structures, the implementation of FTTx (x = B, C, H) is divided into active FTTx by active optical network (hereinafter referred to as "AON") and passive FFTx by PON configuration. Can be. Among them, PON is proposed as an economical optical subscriber network implementation method in the future due to the network configuration having a point-to-multipoint topology by passive devices.

The PON is referred to as one optical line termination device (hereinafter referred to as "OLT") and a number of optical subscriber network devices, ONUs, referred to as 1 x N passive distribution network (ODN). By using), it is an optical subscriber network structure that forms a distributed topology of a tree structure. In the PON form, ATM-PON was first developed and standardized. Recently, the International Telecommunication Union-Telecommunication section (ITU-T) has standardized the contents of the Asynchronous Transfer Mode Passive Optical Network (hereinafter referred to as "ATM-PON") in ITU-T G.982, ITU. -T G.983.1, ITU-T G.983.3. In addition, the IEEE802.3ah TF of the Institute of Electrical and Electronics Engineers (IEEE) is in the process of standardizing G-bit-based GE-PON systems.

Point-to-point Gigabit Ethernet and Medium Access Control (MAC) technologies for ATM-PON have already been standardized, and are described in IEEE 802.3z and ITU-T G. 983.1. In addition, US Patent No. 5,973,374, which was invented by Gigad Ghaib et al. And issued a patent in the United States on November 2, 1999 under the name "PROTOCOL FOR DATA COMMUNICATION OVER A POINT-TO-MULTIPOINT PASSIVE OPTICAL NETWORK," was issued in the United States. MAC technology in PON is disclosed in detail.

ATM-PON includes one OLT 10 and multiple ONUs 12a, 12b, 12c, as shown in the example shown in FIG. 1, and the OLT 10 and ONUs 12a, 12b, 12c. ) Is connected via ODN 16. The OLT 10 is located at the root of the tree structure and plays a central role in providing information to each subscriber in an access network. The OLT 10 has a tree topology structure and distributes downstream data frames transmitted from the OLT 10 to three ONUs 12a, 12b, and 12c, and vice versa. An ODN 16 is connected which multiplexes upstream data frames from the ONUs 12a, 12b, 12c and transmits them to the OLT 10. The ONUs 12a, 12b, 12c receive downlink data frames and provide them to the end users 14a, 14b, 14c and output the data output from the end users 14a, 14b, 14c as uplink data frames. 16) to the OLT 20. The end users 14a, 14b, and 14c, which show three examples, refer to various types of subscriber network terminators that can be used in a PON including a network terminal (NT).

In the ATM-PON as described above, downlink and uplink transmission is performed in the form of a data frame in which ATM cells having a size of 53 bytes are bundled in a predetermined size. In the tree-shaped PON structure as shown in FIG. 1, the OLT 10 properly inserts a downlink cell to be distributed to each of the ONUs 12a, 12b, and 12c in the downlink frame. In addition, in case of uplink transmission, the OLT 10 accesses data transmitted from the ONUs 12a, 12b, and 12c in a time division multiplexing (TDM) manner. At this time, the ODN 16 connected between the OLT 10 and the ONUs 12a, 12b, and 12c is a passive element. Accordingly, the OLT 10 prevents data from colliding with the passive element ODN 16 through virtual distance correction using an algorithm called ranging. In addition, when the OLT 10 transmits downlink data to the ONUs 12a, 12b, and 12c, the OLT 10 and the ONUs 12a, 12b, and 12c are encrypted keys for encryption and maintenance for confidentiality. It is designed to send and receive OAM (Operations, Administration and Maintenance) messages for each other. To this end, up / down frames have corresponding data fields in dedicated ATM cells or general ATM cells that can send and receive messages at regular intervals.

On the other hand, with the development of Internet technology, subscribers require more bandwidth, which requires relatively expensive equipment, bandwidth limitations (up to 622Mbps), and ATMs that need to segment Internet Protocol (IP) packets. The goal is to end-to-end transmission in Gigabit Ethernet technology, which is relatively inexpensive and can achieve high bandwidth (around 1Gbps). Therefore, the PON structure of subscriber networks has come to require Ethernet rather than ATM.

In the case of Gigabit Ethernet, however, the point-to-point and collision-type MAC protocols have already been standardized and the MAC controller chip has been commercialized, but the point-to-multipoint GE-PON architecture includes MAC scheduling. (scheduling) procedures are currently being standardized in the IEEE 8023.ah Ethernet in the First Mile (TFM) TF.

Meanwhile, ATM-PON performs Dynamic Bandwidth Allocation (DBA) for data transmission of ONUs. In such ATM-PON, ONUs have four independent queues according to service classes defined according to parameters such as Variable Bit Rate (VBR), Constant Bit Rate (CBR), and Real-time. This queue ensures the quality of service (QoS) by storing the data traffic coming into the ONU and making dynamic bandwidth allocation in consideration of this class of service.

However, in the case of GE-PON, unlike ATM, the protocol base is Ethernet, so there is no defined service class. In addition, since the packet size of Ethernet, which is an underlying technology, is variable, it must be distinguished from the bandwidth allocation method in ATM-PON based on ATM having a cell having a fixed packet size. Bandwidth allocation scheduling in GE-PON is problematic when considering that it is a PON with a point-to-multipoint structure that has not been used in Ethernet-based networks. The downlink traffic from the OLT to the ONU is broadcasting, so there is no difference compared to conventional Ethernet. However, since traffic from each ONU to the OLT is multiplexed and arrives at the OLT, in order to avoid collisions, the OLT must allocate time to transmit data at different times to each ONU. In addition, since the amount of traffic that ONUs want to send is different, dynamic bandwidth allocation is required to effectively transmit to the OLT without incurring much delay.

Accordingly, an object of the present invention is to provide a bandwidth dynamic allocation method essential for GE-PON implementation.

Another object of the present invention is to provide a dynamic bandwidth allocation method that can improve the utilization of network resources while considering the required bandwidth of the ONU.

It is another object of the present invention to provide a bandwidth dynamic allocation method that can guarantee the minimum amount of transmission by guaranteeing fairness between ONUs.

In order to achieve the above objects, the bandwidth dynamic allocation method according to the present invention requires a bandwidth equal to or smaller than the average value, by first comparing the average of the total available bandwidth by the number of ONUs for which the bandwidth is required and the bandwidth required by each of the ONUs. An ONU allocates all the requested bandwidth, and an ONU that requires more than the average bandwidth allocates only the average bandwidth. If all of the ONUs making the bandwidth request have the remaining bandwidth after the bandwidth allocation is completed, the remaining bandwidth is allocated additionally. At this time, the remaining bandwidth is allocated to ONUs that have been allocated only the average bandwidth in proportion to the size of each unallocated request bandwidth, or unallocated from the ONUs in which the required bandwidth is the largest among the ONUs having only the average bandwidth allocated. Allocate additional bandwidth until all remaining bandwidth is allocated.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the following description, detailed descriptions of well-known functions and configurations that may unnecessarily obscure the subject matter of the present invention will be omitted.

FIG. 2 is a block diagram of a GE-PON to which the present invention is applied. The ODN 20 which is a central station, the optical splitter which is a passive element, and the three ONUs 22a are illustrated. 22b and 22c and end users 24a, 24 and 24c are connected like the ATM-PON of FIG. 1 described above, and use the TDM scheme for uplink transmission according to the tree structure of the point-to-multipoint connection. . However, unlike ATM-PON of FIG. 1, uplink and downlink frames are configured based on a variable length Ethernet frame. The variable length Ethernet frame format structure and the GE-PON functions related to the variable length Ethernet frame, such as initial ONU registration, periodic ONU registration, ranging, and bandwidth dynamic allocation, were determined by the applicant in January 2002. Patent application No. 2765 No. 2765, a Korean patent application dated 17, discloses a method for implementing an operation in a Gigabit Ethernet passive optical subscriber network system and its Ethernet frame structure.

The operation of the GE-PON of FIG. 2 with reference to 2002 Patent Application No. 2765 is as follows. First, the ONUs 22a, 22b, and 22c register to the OLT 20 to inform their location and existence, and are assigned respective ONU IDs. Then, ranging, which is a distance correction procedure, is performed. The reason is that the ONUs 22a, 22b, and 22c compensate for synchronization errors with respect to the up and down time delays during the registration process, but are not precisely corrected for errors that may be caused by other external variables such as temperature. Because it is. The ONUs 22a, 22b and 22c that have completed registration and distance correction are given an opportunity for the OLT 20 to transmit data through an uplink data grant opportunity frame. 22c) measures the amount of data in the buffer it has and transmits this queue value to the OLT 20 in the band allocation request frame. The uplink data transmission opportunity grant frame is a downlink packet used when the OLT 20 wants to give one of the ONUs 22a, 22b, and 22c an opportunity to transmit data upward, and a bandwidth allocation request frame. Is an uplink packet used when one of the ONUs 22a, 22b, 22c makes a band allocation request to the OLT 20 with the permission of the OLT 20. The OLT 20 receives bandwidth requests for a period of time in one time slot and then allocates the appropriate data transmission bandwidth. The result is included in the uplink data transmission opportunity grant frame of the next time slot and transmitted to the ONUs 22a, 22b, and 22c. At this time, the allocation information is composed of a time to start transmission and a time to maintain the transmission, and the ONUs 22a, 22b, and 22c receiving the data transmit data to the OLT 20 as much as the time given at the allocated time.

3 shows that the OLT 20 according to the first embodiment of the present invention corresponds to the bandwidth requirement of the ONUs 22a, 22b, and 22c as described above, according to the minimum-guaranteed portional scheduling algorithm. A process flow for dynamic bandwidth allocation as "MGPSA" is shown in steps 100-122. Definitions of the parameters used in FIG. 3 and FIG. 4 described below are as follows. " BW_TOT_AVAIL " means the total bandwidth available for data transmission of the scheduling target band, that is, the GE-PON as shown in FIG. " N " is the number of ONUs that are required for bandwidth as the number of ONUs in a normal state that completes registration and distance correction and prepares data transmission among all ONUs 22a, 22b, and 22c. " BW_AVG " means an average of total available bandwidth BW_TOT_AVAIL divided by N , that is, an average of required bandwidths of ONUs for which bandwidth is requested. " BW_REQ _i " means the bandwidth requested by a particular ONU, " BW_ALLOC _i " means the bandwidth allocated by the OLT 20 to a specific ONU, " BW_REMAIN " is the minimum of the total available bandwidth, as described below It means the remaining bandwidth after performing primary allocation to guarantee bandwidth. " i " means a specific ONU, and from 1 to the total number of ONUs, that is, an example of three in FIG. , For example, up to 32. " N1 " refers to the number of ONUs to be additionally allocated after primary allocation, as described below.

First, the OLT 20 divides the total available bandwidth BW_TOT_AVAIL by N in step (100) to obtain an average value of the required bandwidths of the ONUs that requested the bandwidth, and then in step (102), among the ONUs that make the bandwidth request with i as 1 in step (102). After specifying the first ONU, perform step (104). In step 104, the required bandwidth RW_REQ _i is compared with the average value BW_AVG . Ten thousand and one bandwidth request RW_REQ if _i is greater than the average value BW_AVG, that is, the average value of only BW_AVG bandwidth in 108 steps assigned when the ONU requested a larger bandwidth than the average BW_AVG. When alternatively or less which is a required bandwidth BW_REQ _i is the average value BW_AVG, that is the average value if equal to the BW_AVG or the one requiring less bandwidth ONU has both a bandwidth BW_ALLOC _i assign a bandwidth BW_REQ _i as required in the 106 step is assigned, and the average value BW_AVG The difference between the bandwidth and the requested bandwidth BW_REQ _i , that is, BW_AVG-BW_REQ _i , is updated as BW_REMAIN in addition to BW_REMAIN . The initial value of BW_REMAIN is 0. Next, i is increased by 1 in step (110), and then it is checked whether N is exceeded in step (112). If i does not exceed N , go to step 104 above and repeat steps 104 through 112 for the next ONU among the ONUs that requested the bandwidth. On the contrary, if i exceeds N , the primary allocation is completed for ONUs that have requested bandwidth, and therefore, the process proceeds to step 114.

Accordingly, when an amount less than or equal to the average value BW_AVG is requested, all of the requested amount is allocated. Otherwise, as much as the average value BW_AVG is allocated first, the minimum bandwidth is guaranteed. This is a way to guarantee minimum bandwidth to all ONUs so that ONUs that require relatively small amounts of bandwidth are not harmed.

After allocating the bandwidth as described above, it is checked whether there is remaining bandwidth BW_REMAIN in the total available bandwidth BW_TOT_AVAIL in step 114. If there is no remaining bandwidth BW_REMAIN , band allocation is terminated, but if there is remaining bandwidth BW_REMAIN , the allocation is additionally allocated to ONUs receiving only the average value BW_AVG in the first allocation in steps 116 to 122. In step 116, i is set to 1, and then, in step 118, the bandwidth BW_ALLOC _i to be allocated is determined as shown in Equation 1, and then additionally allocated.

As the above-mentioned equation (1) is re-allocated BW_REQ _i -BW_ALLOC of targeted ONU _i, that is set by the one (106) Step a requested bandwidth allocated bandwidth BW_ALLOC BW_REQ _{i i} in proportion to the size of BW_REQ _i -BW_AVG The remaining bandwidth BW_REMAIN is allocated dynamically. Then, in step 120, i is increased by 1, and then it is checked whether N1 is exceeded in step 122. If i does not exceed N1 , the process proceeds to step 118 described above and repeats steps 118 to 122 for the next ONU among the ONUs to be additionally assigned. On the contrary, when i exceeds N1 , all additional allocations are completed for ONUs that are additional allocation targets, and thus, the process ends.

According to the MGPSA as described above, if a quantity that is less than or equal to an average value is firstly requested, all of the required amount is allocated. Otherwise, as much as the average value is firstly allocated, a relatively small amount of band is required. This ensures minimum bandwidth for all ONUs so that they do not suffer damage. Thereafter, if there is remaining bandwidth, the remaining bandwidth is allocated to ONUs having only the average bandwidth allocated in proportion to the size of each unallocated request bandwidth.

On the other hand, there is a burst property as a characteristic of the traffic of the ONU. On the other hand, even if the required bandwidth of one particular ONU is significantly larger than the other ONUs, according to the MGPSA, only a little larger than the average value will be allocated, so that the overall traffic is uniform and thus the bursty characteristics of the original traffic can be achieved. This may result in a failure to reflect. Therefore, if the traffic characteristics are bursting, it is desirable to perform dynamic bandwidth allocation by reflecting this.

FIG. 4 illustrates that the OLT 20 minimizes the maximum bandwidth guarantee by reflecting the case where the traffic characteristic is burst according to the second embodiment of the present invention in response to the bandwidth requirements of the ONUs 22a, 22b, and 22c as described above. The process flow for performing dynamic bandwidth allocation as the minimum-guaranteed maximum scheduling algorithm (hereinafter referred to as "MGMPSA") is shown in steps (100) to (112) and (200) to (204). Steps (100) to (112) of FIG. 4 are the same as steps (100) to (112) of FIG. Therefore, steps 100 through 112, which guarantee minimum bandwidth by primary allocation, are the same as the MGPSA described above.

However, when additionally allocating the remaining bandwidth BW_REMAIN after the first allocation, the allocation method is different as in steps 200 and 204. In this case, the additional bandwidth is allocated until the remaining bandwidth BW_REMAIN is allocated from the ONUs in which the sizes of the unallocated request bandwidths BW_REQ _i -BW_AVG are allocated among the requested bandwidths of the ONUs to be allocated additionally. First, in step 200, it is checked whether there is remaining bandwidth BW_REMAIN , and if not, it ends. However, if there is the remaining bandwidth BW_REMAIN , in step 202, MAX (BW_REQ _k -BW_ALLOC _k ) , that is, BW_REQ _k -BW_ALLOC _k selects the ONU with the maximum value, and performs additional allocation to this ONU in step (204). In step 204, the bandwidth BW_ALLOC _k to be allocated is additionally determined by Equation 2 below, and BW_REMAIN is updated as shown in Equation 3 below.

As shown in Equation 2, the remaining bandwidth BW_REMAIN is allocated to the ONU having the largest unassigned request bandwidth. When BW_REMAIN is smaller than the ONW 's BW_REQ _i -BW_AVG , all of the BW_REMAIN are allocated. If there is still bandwidth after this allocation, proceed to step (200) again to provide ONU with the next highest unallocated bandwidth until all remaining bandwidth BW_REMAIN is allocated. Will be allocated.

Therefore, the minimum bandwidth is guaranteed by the primary allocation to all ONUs that require bandwidth, and when there is remaining bandwidth, the traffic characteristics burst by allocating the maximum bandwidth as much as possible from the ONU with the largest bandwidth demand. This also reflects the dynamic bandwidth allocation.

As described above, the present invention proposes a dynamic bandwidth allocation scheme essential for implementing a GE-PON, so that even when the characteristics of the GE-PON having a variable packet length and the traffic characteristics are burst, the dynamic allocation is effectively performed to provide the characteristics of the data. It can conserve and increase the utilization of limited resources. Accordingly, in the GE-PON, which has a point-to-multipoint characteristic different from the conventional Ethernet-based network, bandwidth can be allocated according to the requirements of the ONU without data collision.

Claims (2)

In a Gigabit Ethernet Passive Optical Network (OLT) in which one optical line terminal (OLT) and a plurality of optical network units (ONUs) are connected through an optical distribution network (ODN), the OLT is data of the ONUs. A method of allocating bandwidth to each ONU in response to a bandwidth request for transmission, The average of the total available bandwidth divided by the number of ONUs that made the bandwidth request is compared with the bandwidth required by each of the ONUs, and all the required bandwidths are allocated to the ONUs that requested the bandwidth below the average value, and the average value is larger than the average value. Allocating only the average bandwidth to the ONU requesting the bandwidth; If there are remaining bandwidths among the total available bandwidths after completing allocating the bandwidths to all ONUs that have requested the bandwidths, the remaining bandwidths are allocated to the ONUs that have been allocated only the average bandwidth, and the size of each unallocated request bandwidth. Dynamic bandwidth allocation method comprising the step of additionally in proportion to. In a Gigabit Ethernet Passive Optical Network (OLT) in which one optical line terminal (OLT) and a plurality of optical network units (ONUs) are connected through an optical distribution network (ODN), the OLT is data of the ONUs. A method of allocating bandwidth to each ONU in response to a bandwidth request for transmission, The average of the total available bandwidth divided by the number of ONUs that made the bandwidth request is compared with the bandwidth required by each of the ONUs, and all the required bandwidths are allocated to the ONUs that requested the bandwidth below the average value, and the average value is larger than the average value. Allocating only the average bandwidth to the ONU requesting the bandwidth; If all of the ONUs that have requested the bandwidth have the remaining bandwidth among the total available bandwidth after completing the bandwidth allocation, unassigned from the ONUs in the order of the larger requested bandwidth among the ONUs allocated only the average bandwidth. And allocating the remaining bandwidth until all the remaining bandwidth is allocated.

Images