Improving the Efficiency of Light Distribution in Passive Optical Network - Article Example

Improving PON Efficiency in FTTx Networks Email Supervisor’s Email Department, Address This article discusses how Ethernet Passive Optical Network (EPON) technology is utilized in implementing fiber to the home/ curb/ building (FTTx) solutions for “last mile” bandwidth bottleneck problems. The fundamental concepts on EPONs are presented at the beginning of the article, followed by discussions on EPON architectures and available upstream channel solutions. Several bandwidth allocation algorithms are then presented at the last part of the report. INTRODUCTION The telecommunications industry has experienced a tremendous growth in bandwidth capacity in the past few years because of developments in fiber-optics technology. This is true in wide area networks (WANs) that provide connectivity between cities and in metropolitan area networks (MANs) that connect telco operators’ nodes within cities. However, most local loop or the “last mile” that serve residential, small business or enterprise users, have not benefited from this. The local subscriber lines for telephone and Internet are still using twisted copper pairs while cable television subscribers are still using copper coaxial cable [1]. With the increasing users’ demands for services such as Internet applications, VoIP, interactive games, high-definition television (HDTV) and video on demand (VOD), the “last mile” connection has become a bandwidth bottleneck. Developments in xDSL and cable TV technologies has in some extent addressed this problem but still not enough to meet the continuously increasing bandwidth demand. A more effective solution is gradually being put in place especially in urban areas by extending the fiber to the user. This technology is called by many names depending on the termination mode – fiber to the home (FTTH), fiber to the curb (FTTC) or fiber to the building (FTTB). All of these FTTx solutions may utilize the Ethernet Passive Optical Network (EPON) distribution technology. This system utilizes bandwidth allocation algorithms to allow efficient sharing of limited upstream channel bandwidth [2]. Different methods of implementing this will be discussed in this article. EPON Architectures Network Architectures An EPON system is a point-to-multipoint fiber optical network with no active elements in the transmission path from the source, an optical line terminal (OLT), to the destination, an optical network unit (ONU). It can use different multipoint topologies, such as bus, ring, and tree. The most typical architecture is based on a tree topology and consists of an OLT, a 1:N passive star coupler (or splitter/combiner), and multiple ONU. The OLT resides in an operator’s central office (CO) that connects the access network to either a metropolitan area network (MAN) or a wide area network (WAN). The OLT is also connected to the passive star coupler through a singe optical fiber. This passive coupler is located farther from the CO but closer to the subscriber premises. Each ONU is located either at curbs or at subscriber premises, and is connected to the passive coupler through a dedicated short optical fiber. The distance between the OLT and each ONU may rangefrom 10 to 20 km [2]. Figure 1 shows the diagram of an EPON system. Figure 1: EPON System (Reproduced with permission of the Metro Ethernet Forum) In the downstream direction, an EPON is a point-to-multipoint network, in which the OLT broadcasts data to each ONU through the 1:N splitter. The value of N is typically between 4 and 64. The media access control (MAC) of each ONU tells it when and how to extract data [3]. Figure 2 shows the EPON downstream operation. Figure 2: EPON Downstream Operation (Reproduced with permission of the Metro Ethernet Forum) In the upstream direction, an EPON is a multipoint-to-point network, in which multiple ONUs transmit data to the OLT through the 1:N passive combiner. All ONUs share the same upstream transmission medium with limited bandwidth so an EPON system must employ a MAC mechanism to arbitrate the access to the shared medium. The purpose of this is to avoid data collisions in the upstream direction and thus efficiently share the upstream transmission bandwidth among all ONUs [4]. Figure 3 shows the EPON upstream operation. Figure 3: EPON Upstream Operation (Reproduced with permission of the Metro Ethernet Forum) A two-stage EPON architecture can also be implemented. In this scheme, an intermediate level of ONU nodes (called sub-OLT) are connected to 1:N splitters. This allows more end-users to share the upstream OLT bandwidth and enables longer access reach/distances than the usual 20 km maximum distance between the ONUs and the CO [5]. Channel Separation In order to increase transmission efficiency in an EPON system, the upstream and downstream transmission channels should ideally be separated. One simple solution is to utilize space division multiplexing, where two separate optical fibers and passive couplers are used for upstream and downstream transmissions. However, there is a more cost-effective solution. This is by using different wavelengths for upstream and downstream transmissions. Typically, a 1550 nm wavelength is used for downstream transmission and a 1310 nm wavelength is used for upstream transmission [2]. Multiple Access CSMA/ CD In the upstream direction of an EPON system, multiple ONUs transmit data packets to the OLT through a common passive combiner and share the same optical fiber from the combiner to the OLT. The passive combiner has a directional property so the data packets from an ONU can only reach the OLT but not the other ONUs. Conventional contention-based multiple access such as the carrier sense multiple access with collision detection (CSMA/CD) protocol, is difficult to implement because the ONUs are unable to easily detect a collision that may occur at the OLT. Although the OLT is able to detect a collision and inform the ONUs by sending a collision message, the transmission efficiency is reduced because of considerable propagation delay between the OLT and the ONUs. To address this problem, an optical looping-back technique was proposed in order to achieve high channel efficiency with CSMA/CD [6]. With this looping-back technique, a portion of the upstream signal power transmitted by each ONU is looped back to the other ONUs at using a 3:N coupler and connecting two ports of the coupler together through an isolator. However, to implement the optical CSMA/CDprotocol, each ONU has to use an additional receiver operating at the upstream wavelength and a carrier sensing circuit, which would largely increase the network cost. This solution also is unable to provide guaranteed bandwidth to each ONU and thus is difficult to support any form of quality of service (QoS). WDM Another possible solution is to use wavelength division multiplexing (WDM) technology and allow each ONU to operate at a different wavelength, thus avoiding interference with the transmissions of the other ONUs. This technology is called WDM EPON. It is simple to implement, but requires a receiver array at the OLT to receive the data transmitted in multiple channels. It also requires each ONU to use a fixed transmitter operating at a different wavelength, which would result in an expensive solution [7]. TDM The third and probably the most practical solution is time division multiplexing (TDM) on a single wavelength for the upstream transmission. With TDM, each ONU is allocated a timeslot or transmission window for data transmission, which is performed by the OLT. Ethernet packets received from one or more users are buffered in an ONU until the timeslot for that ONU arrives. Upon the arrival of its timeslot, the ONU will send out its buffered packets at the full transmission rate of the upstream channel. Accordingly, TDM avoids data collisions from the different ONUs. This article focuses on bandwidth allocation in TDM EPONs since this is the most practical and cost-effective solution [2]. Bandwidth Management in EPONs In this section, we discuss the major issues that are related to bandwidth management in TDM EPON systems. Bandwidth management is a critical issue for efficiently utilizing the bandwidth of the shared upstream wavelength. Bandwidth management involves two main issues: bandwidth negotiation and bandwidth allocation. Bandwidth Negotiation Bandwidth negotiation is the process of exchanging information between the OLT and each ONU in order for each ONU to report its bandwidth demand to the OLT and for the OLT to send its bandwidth allocation decision to each ONU. For this purpose, IEEE 802.3ah defines a multipoint control protocol (MPCP) to support bandwidth negotiation between the OLT and ONUs in EPON. This includes two 64-bytes MAC control messages: REPORT and GATE. The REPORT message is generated by each ONU to report its queue status to the OLT. The OLT allocates bandwidth for each ONU based on the queue status information contained in the received REPORT message, and uses the GATE message to deliver its bandwidth allocation decision to each ONU [8]. Bandwidth Allocation To allocate bandwidth or a timeslot for each ONU, the OLT needs to perform a bandwidth allocation algorithm based on the bandwidth requests from each ONU as well as some allocation policy and/or service level agreement (SLA). There are several bandwidth allocation algorithms available, which can be classified into two broad categories: static bandwidth allocation (SBA) and dynamic bandwidth allocation (DBA). With SBA, each ONU is allocated a timeslot with a fixed length, which does not require bandwidth negotiation and is thus simple to implement. However, due to the bursty nature of the network traffic, it may result in a situation in which some timeslots are overflowed even under very light load while other timeslots are not fully used even under very heavy traffic. This results to the upstream bandwidth being underutilized. To increase bandwidth utilization, the OLT must dynamically allocate a variable timeslot to each ONU based on the instantaneous bandwidth demand of the ONUs. To implement DBA, polling is used that flexibly arbitrate the transmissions of multiple ONUs. This significantly increase bandwidth utilization and improve network performance [2]. Dynamic Bandwidth Allocation Algorithms for EPONs In this section, we present a number of proposed DBA algorithms for EPONs. QoS is a main concern in EPONs so we classify these algorithms into DBA with QoS support and DBA without QoS support. DBA without QoS Support Interleaved Polling with Adaptive Cycle Time (IPACT) IPACT is the first DBA algorithm proposed for EPON. In this solution, the OLT polls ONUs and grants timeslots to each ONU in a round-robin fashion. The timeslot granted to an ONU is determined by the queue status reported from that ONU. The OLT is able to know the dynamic traffic load of each ONU and allocate the upstream bandwidth in accordance with the bandwidthdemand of each ONU. IPACT also checks the service level agreement (SLA) of end users to limit the allocated bandwidth to each ONU [9-12]. Estimation-Based Dynamic Bandwidth Allocation In estimation-based DBA algorithm, the queue length of each ONU is reduced by estimating the packets arriving at an ONU during the waiting time and incorporating the estimation in the grant to the ONU. In this algorithm, a control gain is used to adjust the estimation that is based on the difference between the departed and arrived packets in the previous transmission cycle. This results to a reduced average packet delay as compared to IPACT [13]. Interleaved Polling with Adaptive Cycle Time with Grant Estimation (IPACT-GE) This solution is a combination of IPACT and estimation-based DBA. With IPACT-GE, the amount of packets arriving to an ONU between two consecutive pollings is estimated based on the similarity characteristic of network traffic. The OLT then decides the granted transmission size for the ONU based on the estimated packet amount as well as the amount requested in the previous polling cycle [14]. Bandwidth Guaranteed Polling (BGP) Bandwidth guaranteed polling (BGP) is a DBA algorithm proposed to provide bandwidth guarantees in EPONs. In this scheme, ONUs are divided into two groups: bandwidth guaranteed and bandwidth non-guaranteed. The OLT performs bandwidth allocation by using a couple of polling tables. The first polling table divides a fixed-length polling cycle into a number of bandwidth units and each ONU is allocated a certain number of such bandwidth units. The number of bandwidth units allocated to an ONU is determined by the bandwidth demand of that ONU, which is given by its SLA with a service provider [15]. Multi-Thread Polling This algorithm allows each ONU to send its REQUEST message before the previous GATE message is received from the OLT. This a new “thread” of signaling between an ONU and the OLT. This algorithm is implemented by allowing the OLT to maintain a polling table wherein each ONU has an entry that records the ONU’s Round-Trip Time (RTT) and its most recent requests in each thread. In each thread, the OLT performs bandwidth allocation and distributes the GATE messages to all ONUs. An important procedure in this algorithm is setting the proper initial thread interval and tuning threshold, which can improve the performance of a single-thread polling algorithm in terms of average packet delay and throughput under heavy traffic load [16]. DBA with QoS Support Dynamic Bandwidth Allocation with Multiple Services (DBAM) Dynamic Bandwidth Allocation with Multiple Services accommodates different types of traffic in EPON networks. Instead of providing multiple services among ONUs and among end users separately, DBAM incorporates both of them into the REPORT/GATE mechanism with class-based bandwidth allocation. It also utilizes limited bandwidth allocation to arbitrate bandwidth allocation among ONUs in order to prevent aggressive bandwidth scrambling. In addition, it employs class-based traffic prediction to take into account the traffic that arrives during the waiting period, which ranges from sending the queue status report to sending the traffic buffered in each ONU. Such prediction is based on the actual traffic received in the previous waiting period. The OLT serves all ONUs in a fixed round-robin fashion in order to facilitate traffic prediction [17]. Two-Layer Bandwidth Allocation Two-layer bandwidth allocation (TLBA) is a hierarchical allocation algorithm that allocates bandwidth in two layers. In the first layer, the transmission cycle is partitioned or the upstream bandwidth is allocated among differentiated service classes that is called class-layer allocation. In the second layer, the partition or bandwidth allocated to each class is distributed to all ONUs within the same class based on a max-min fairness policy called ONU-layer allocation. To avoid any class from monopolizing the available bandwidth in a cycle, a per-class threshold is introduced. This guarantees a minimum bandwidth for each class under high traffic load. Any excess bandwidth from the classes that need less than their thresholds is distributed among the classes that need more than their thresholds [18]. Queue-Based Dynamic Bandwidth Allocation Queue-based DBA is an OLT-centric algorithm based on individual requests from the service queues in ONUs for a QoS-aware EPON. To prevent high-priority traffic from completely taking away bandwidth from lower-priority traffic, this scheme utilizes queue scheduling and makes use of the excess bandwidth of lightly-loaded queues to meet the bandwidth demand of heavily-loaded queues. It incorporates an efficient polling mechanism to solve the idle period problem and employs a different-cycle policy to reduce the scheduling overhead that selectively allocates bandwidths to different service classes based on their delay bounds. This DBA algorithm outperforms the IPACT in terms of average packet delay and bandwidth utilization [19]. QoS-Aware Dynamic Bandwidth Allocation The Dynamic Credit Distribution (D-CRED) is a QoS-aware DBA algorithm designed for gigabit EPONs. In this scheme, unused slot remainders are eliminated by introducing a dynamic queue threshold technique that allows each ONU to have only one threshold, and the OLT to dynamically change the threshold value (called credit). This results to higher throughput efficiency and more precise bandwidth allocation. D-CRED can achieve a 6.4% higher bandwidth utilization compared to IPACT with limited service, and 99% of the theoretical maximum utilization. D-CRED can also be extended to QoS support. For this purpose, a concept of fairness is defined and based on this, D-CRED calculates a degree of satisfaction of an ONU regarding credit allocation in each cycle and keeps track of the degree during the entire busy periods [20]. Fine Scheduling The fine scheduling algorithm consists of an inter-ONU scheduler at the OLT and an intra-ONU scheduler at each ONU. In the inter-ONU scheduling, a DBA algorithm fairly allocates the bandwidth for upstream data transmission and optimizes the upstream bandwidth utilization. Unlike most other DBA algorithms that only report the buffered data size at each ONU, fine scheduling DBA allows each ONU to send two requests in its REPORT messages: a maximum window size and a minimum or guaranteed window size. After receiving all the REPORT messages in a cycle, it performs bandwidth allocation and allocates a bandwidth to each ONU based on the following fairness criteria: 1) The minimum bandwidth demand is always guaranteed; 2) The excess bandwidth is allocated to a unit (an ONU or a queue) according to its weight; 3) No unit is allocated a bandwidth more than its maximum bandwidth demand. This algorithm also eliminate unused slot remainders without causing transmission delay by sending the REPORT message ahead of the data stream. It also informs the OLT of the requested maximum and minimum bandwidths for the next cycle as well as the total actual bandwidth of each ONU for this cycle. This algorithm also employs hierarchical intra-ONU scheduler to realize fine granularity scheduling to support QoS for traffic of each individual user [21-23]. CONCLUSION Ethernet Passive Optical Network or EPON is an efficient distribution solution for FTTx networks. It combines low-cost point-to-multipoint optical infrastructure with low-cost high-bandwidth Ethernet. However, there are several concerns in implementing the upstream channel from the ONUs to the OLTs considering that this is a shared channel with limited bandwidth capacity. There are several solutions developed to address this issue of bandwidth bottleneck in EPON upstream channels. Based on the discussions, dynamic bandwidth allocation (DBA) based on TDM access method could be the most practical solution. TDM only requires a single transceiver at the OLT unlike the other access methods that require multiple transmitters or receivers. This is complemented by DBA’s ability to dynamically allocate variable timeslots to the ONUs that results to increased bandwidth utilization and improved network performance. REFERENCES [1] J. Zheng and H. T. Mouftah, Optical WDM Networks: Concepts and Design Principles, Wiley-IEEE Press, Hoboken, New Jersey, 2004 [2] G. Kramer and G. Pesavento, “Ethernet passive optical network (EPON): building a next- generation optical access network,” IEEE Communications Magazine, vol. 40, no. 2, Feb. 2002, pp. 66-73 [3] C. Foh, L. Andrew, E. Wong, and M. Zukerman, “FULL-RCMA: a high utilization EPON,” IEEE Journal on Selected Areas in Communications, vol. 22, no. 8, Oct. 2004, pp. 1514-1524 [4] S. Sherif, A. Hadjiantonis, G. Ellinas, C. Assi, and M. Ali, “A novel decentralized Ethernet-based PON access architecture for provisioning differentiated QoS,” IEEE/OSA Journal of Lightwave Technology, vol. 22, no. 11, Nov. 2004, pp. 2483-2479 [5] A. Shami, X. Bai, N. Ghani, C. Assi, and H. Mouftah, “QoS control schemes for two-stage Ethernet passive optical access networks,” IEEE Journal on Selected Areas in Communications, vol. 23, no. 8, Aug. 2005, pp. 1467-1478 [6] Chang-Joon Chae et al., “Optical CSMA/CD media access scheme for Ethernet over passive optical network,” IEEE Photonics Technology Letters, vol. 14, no. 5, May 2002, pp. 711-713 [7] B. N. Nesai et al., “An optical implementation of a packet-based (Ethernet) MAC in a WDM passive optical network overlay,” in Proc. of OFC’01, vol. 3, 2001, pp. WN5-1- WN5-3 [8] Y. Luo, S. Yin, N. Ansari, and T. Wang, “Resource management for broadband access over time-division multiplexed passive optical networks,” IEEE Network, vol. 21, no. 5, Sept./Oct. 2007, pp. 20-27 [9] G. Krammer, B. Mukherjee, and G. Pesavento, “IPACT: a dynamic protocol for an Ethernet PON (EPON),” IEEE Communications Magazine., vol. 40, no. 2, Feb. 2002, pp. 74-80 [10] IEEE 802.3ah Ethernet in the First Mile Task Force, Retrieved April 25, 2009, from http://www.ieee802.org/3/efm/index.html [11] J. Zheng and H. T. Mouftah, “Media access control for Ethernet passive optical networks (EPONs), an overview,” IEEE Communications Magazine, vol. 43, no. 2, Feb. 2005, pp. 145-150 [12] J. Zheng and H. T. Mouftah, “Adaptive scheduling algorithms for Ethernet passive optical networks,” IEE Proceedings-Communications，vol. 152, no. 5, Oct. 2005, pp. 643-647 [13] H.-J. Byun, J.-M. Nho, and J.-T. Lim, “Dynamic bandwidth allocation algorithm in Ethernet passive optical networks,” IEEE Electronics Letter, vol. 39, no. 13, Jun. 2003, pp. 1001-1002 [14] Y. Zhu and M. Ma, “IPACT with grant estimation (IPACT-GE) scheme for Ethernet passive optical networks,” IEEE/OSA Journal of Lightwave Technology, vol. 26, no. 14, Jul. 15, 2008, pp. 2055-2063 [15] M. Ma, Y. Zhu, and T. Cheng, “A bandwidth guaranteed polling MAC protocol for Ethernet passive optical networks,” in Proc. of IEEE INFOCOM’03, vol. 1, Mar./Apr. 2003, pp. 22-31 [16] H. Song, B.-W. Kim, and B. Mukherjee, “Multi-Thread Polling: A Dynamic Bandwidth Distribution Scheme in Long-Reach PON”, IEEE Journal on Selected Areas in Communications, vol. 27, no. 2, Feb. 2009, pp. 134-142 [17] Y. Luo and N. Ansari, “Bandwidth allocation for multiservice accesss on EPONs,” IEEE Communications Magazine, vol. 43, no. 2, Feb. 2005, pp. S16-S21 [18] J. Xie, S. Jiang, and Y. Jiang, “A dynamic bandwidth allocation scheme for differentiated services in EPONs,” IEEE Communications Magazine, vol. 42, no. 8, Aug. 2004, pp. S32- S39 [19] P. K. Choudhury and P. Saengudomlert, “Efficient queue based dynamic bandwidth allocation scheme for Ethernet PONs,” in Proc. of IEEE GLOBECOM’07, Washington D.C., Nov. 2007, pp. 2183-2187 [20] H. Miyoshi, T. Inoue, and K. Yamashita, “QoS-aware dynamic bandwidth allocation scheme in Gigabit-Ethernet passive optical networks,” in Proc. of IEEE ICC’04, Paris, France, Jun. 2004, pp. 90-94 [21] B. Chen, J. Chen, and S. He, “Efficient and fine scheduling algorithm for bandwidth allocation in Ethernet passive optical networks,” IEEE Journal of Selected Topics in Quantum Electronics, vol. 12, no. 4, Jul./Aug. 2006, pp. 653-660 [22] N. Ghani et al. “Quality of service in Ethernet passive optical networks,” in Proc. of 2004 IEEE Sarnoff Symposium, Princeton, NJ, Apr. 2004, pp. 161-165 [23] Ethernet Passive Optical Network – A Tutorial, Retrieved April 25, 2009, from http://www.metroethernetforum.org Read More