Ethernet pon (EPON) and PON technology are the mainstream technologies of FTTH now. What is the difference between ethernet pon (EPON) and PON?
This paper also systematically analyzes the technical realization of ethernet pon (EPON) technology in FTTH; these technologies promote the development and application of FTTH, and the design of FTTH optical network becomes clearer.
As a communication equipment factory with 27 years of production and design experience, Melontel has many years of experience in your FTTH construction and ethernet pon (EPON) technology application. It can provide a strong product guarantee for your FTTH construction. At the same time, Melontel is a TUV certified enterprise, with strong product design and production supervision to ensure the quality of your products.
What is PON?
PON (Passive Optical Network) passive optical network is the main technology to realize FTTH fiber-to-the-home, providing point-to-multipoint fiber access. The ONU (optical network unit) and the ODN (optical distribution network) are composed. Generally, the downlink adopts TDM broadcast mode, and the uplink adopts TDMA (time division multiple access) mode, forming a point-to-multipoint tree topology. The biggest highlight of (Passive Optical Network) as an optical access technology is “passive”. ODN does not contain any active electronic devices and electronic power supplies. It is all composed of passive devices such as optical splitters (Splitter), and the management, maintenance and operation costs are low.
The research on PON (Passive Optical Network) technology originated in 1995. In October 1998, ITU passed the ATM-based PON (Passive Optical Network) technology standard – G.983, which was advocated by the FSAN organization (full service access network). Also known as BPON (Broadband PON). The rate is 155 Mbps, with optional support for 622 Mbps.
EFMA (Ethernet in the First Mile Alliance) proposed the concept of Ethernet-PON (ethernet pon (EPON)) at the end of 2000, with a transmission rate of 1Gbps, and the link layer is based on simple Ethernet encapsulation.
GPON (Gigabit-Capable PON) was proposed by the FSAN organization in September 2002, and in March 2003, the ITU passed the G.984.1 and G.984.2 protocols.
G.984.1 specifies the overall characteristics of the GPON access system; G.984.2 specifies the ODN (Optical Distribution Network) physical medium-related sublayer of GPON; in June 2004, ITU passed G.984.3, which The relevant requirements for the Transmission Convergence (TC) layer are specified.
What is the difference between ethernet pon (EPON) and GPON?
As the two main members of optical network access, ethernet pon (EPON) and GPON have their own advantages, compete with each other, complement each other, and learn from each other. The following is a comparison of them in various aspects:
Ethernet pon (EPON) provides fixed uplink and downlink 1.25 Gbps, using 8b/10b line coding, the actual rate is 1Gbps.
GPON supports a variety of rate levels, can support asymmetric rates of upstream and downstream, downstream 2.5Gbps or 1.25Gbps, upstream 1.25Gbps or 622 Mbps, according to actual needs to determine the upstream and downstream rates, select the corresponding optical module, improve the rate and price ratio of optical devices .
The split ratio is how many ONUs (customer) an OLT port (central office) has. The ethernet pon (EPON) standard defines a split ratio of 1:32.
The GPON standard defines the following split ratios as 1:32; 1:64; 1:128.
In fact, technically, the ethernet pon (EPON) system can also achieve a higher split ratio, such as 1:64, 1:128, and the ethernet pon (EPON) control protocol can support more ONUs. The splitting ratio is mainly limited by the performance indicators of the optical module, and a large splitting ratio will cause the cost of the optical module to increase significantly;
In addition, the PON insertion loss is 15 ~ 18dB, and a large split ratio will reduce the transmission distance;
Excessive user sharing of bandwidth is also the price of a large split ratio.
Maximum transmission distance
The maximum physical distance that the GPON system can support. When the optical splitting ratio is 1:16, it should support a maximum physical distance of 20km; when the optical splitting ratio is 1:32, it should support a maximum physical distance of 10km.
The Ethernet pon (EPON) is the same.
QOS (Quality of Service): ethernet pon (EPON) adds a 64-byte MPCP Multi Point Control Protocol (Multi Point Control Protocol) to the Ethernet header of the MAC layer. MPCP controls the access to P2MP point-to-multipoint topology through messages, state machines and timers structure to achieve DBA dynamic bandwidth allocation.
The content involved in MPCP includes the allocation of ONU transmission time slots, the automatic discovery and joining of ONUs, and the reporting of congestion to the upper layer for dynamic allocation of bandwidth. MPCP provides basic support for P2MP topology, but the protocol does not classify the priority of services, and all services compete for bandwidth randomly.
GPON has a more complete DBA with excellent QoS service capabilities. GPON divides the service bandwidth allocation method into 4 types. The priority from high to low is Fixed Bandwidth (Fixed), Assured Bandwidth (Assured), Non-Assured Bandwidth (Non-Assured) and Best Effort Bandwidth (Best Effort). DBA A traffic container (T-CONT) is also defined as an upstream traffic scheduling unit, and each T-CONT is identified by an Alloc-ID. each
T-CONT can contain one or more GEM Port-IDs. T-CONT is divided into 5 service types. Different types of T-CONT have different bandwidth allocation methods, which can meet the needs of different service flows for delay, jitter, and packet loss. rate and other different QoS requirements.
T-CONT type 1 is characterized by fixed bandwidth and fixed time slot, corresponding to fixed bandwidth (Fixed) allocation, suitable for services that are sensitive to delays, such as voice services;
Type 2 is characterized by fixed bandwidth but uncertain time slots, corresponding to guaranteed bandwidth (Assured) allocation, suitable for fixed bandwidth services that do not require high jitter, such as video-on-demand services;
Type 3 is characterized by the minimum bandwidth guarantee and the ability to dynamically share the surplus bandwidth, and has the constraint of the maximum bandwidth, corresponding to non-guaranteed bandwidth (Non -Assured) allocation, suitable for services with service guarantee requirements and large burst traffic, such as download services;
Type4 is characterized by best effort (Best Effort), no bandwidth guarantee, suitable for services with low latency and jitter requirements, such as WEB browsing services;
Type 5 is a combined type, after allocating guaranteed and non-guaranteed bandwidth, additional bandwidth requirements are allocated on a best-effort basis.
Operation and maintenance of OAM
The ethernet pon (EPON) does not give too much consideration to OAM, but simply defines ONT remote fault indication, loopback and link monitoring. And it is optional to support GPON. PLOAM (Physical Layer OAM) is defined at the physical layer, and the high layer defines OMCI (ONT Management and Control Interface) manages OAM at multiple levels. PLOAM is used to implement data encryption, status detection, error monitoring and other functions.
The OMCI channel protocol is used to manage the services defined by the high layer, including the functional parameter set of the ONU, the type and quantity of T-CONT services,
QoS parameters, request configuration information and performance statistics, automatically notify system operation events, and realize OLT’s management of ONT configuration, fault diagnosis, performance and security.
Link layer encapsulation and multi-service support
The ethernet pon (EPON) follows the simple Ethernet data format, but adds a 64-byte MPCP point-to-multipoint control protocol to the Ethernet header to realize bandwidth allocation, bandwidth polling, automatic discovery, ranging, etc. in the ethernet pon (EPON) system. Not much research has been done on the support of services other than data services (such as TDM synchronization services). Many ethernet pon (EPON) manufacturers have developed some non-standard products to solve this problem, but they are not ideal, and it is difficult to meet the carrier-class QoS requirements.
GPON is based on a completely new Transmission Convergence (TC) layer. This sublayer can complete the adaptation of high-level diversified services. It defines ATM encapsulation and GFP encapsulation (general framing protocol), and one of the two can be selected for service encapsulation. In view of the current ATM application is not popular, so a GPON.lite device that only supports GFP encapsulation came into being, which removes ATM from the protocol stack to reduce costs.
GFP is a general link layer procedure suitable for multiple services. ITU defines it as G.7041. GPON has made a few modifications to GFP. Port ID is introduced in the header of the GFP frame to support multiple ports. Multiplexing; Fragment (Fragment) segment indication is introduced to improve the effective bandwidth of the system. And only supports variable-length data-oriented data processing mode and does not support data block-oriented data transparent processing mode.
GPON has a strong multi-service bearing capacity.
The TC layer of GPON is essentially synchronous, using a standard 8k Hz (125μm) fixed-length frame, which enables GPON to support end-to-end timing and other quasi-synchronous services, especially TDM services, which is the so-called Native TDM and GPON have “natural” support for TDM services.
ethernet pon (EPON) and GPON have their own strengths and weaknesses. GPON is better than ethernet pon (EPON) in terms of performance indicators, but ethernet pon (EPON) has advantages in terms of time and cost, and GPON is catching up. Looking forward to the future broadband access market, it may not be one who replaces the other, but should coexist and complement each other. For customers with high bandwidth, multi-service QoS and security requirements and ATM technology as the backbone network, GPON will be more suitable. For cost-sensitive, QoS, and low-security customer groups, ethernet pon (EPON) has become the dominant.
FTTX Optical Fiber Network Construction Solution
Uplink Access Technology of ethernet pon (EPON)
The ethernet pon (EPON) uplink access technology can use frequency division multiple access (FDMA), time division multiple access (TDMA), wavelength division multiple access (WDMA), code division multiple access (CDMA) and other methods. The disadvantages of frequency division multiplexing are: multiple repeaters are required in the OLT; initial cost is high; channel capacity is fixed; it is troublesome to increase/decrease users. The disadvantages of code division multiplexing are: interference between channels is determined by the number of ONUs; high-speed devices are required.
The main problem of WDMA is: the cost is relatively high, mainly because the cost of optical devices is still relatively high.
However, if the cheap optical transmitter and optical receiver devices suitable for the WDMA access network are developed, the use of WDMA is the trend of development. Time division multiplexing has the advantages of N ONUs requiring only one OLT transponder and very few wavelengths, but time division multiplexing still requires high-speed electronics.
Based on the above considerations, now we generally use the time division multiplexing multiple access method, and the wavelength division multiplexing multiple access method also has great competitiveness and is the direction of development.
The OLT arranges the time slots in which each ONU is allowed to send uplink signals, and sends out time slot allocation frames. The ONU allocates the frame according to the time slot, and sends its own upstream signal in the time slot allocated to it by the OLT.
In this way, the ONUs can share the upstream channel together, that is, many ONUs share the limited upstream channel bandwidth.
Each ONU occupies one wavelength exclusively, and the detector array receives the signal of each ONU. According to the receiving detector, it can be determined which ONU corresponds to.
Since the ONU exclusively owns one wavelength, it does not need to share the bandwidth with other ONUs, and each upstream channel can achieve a bandwidth symmetrical with that of the downstream channel.
That is to say, in this way, the upstream bandwidth enjoyed by each ONU can be guaranteed, without competing with other ONUs for the limited bandwidth.
FTTH Cables and Their Role in Internet Connectivity
Analysis of key technologies in ethernet pon (EPON)
1. The key technologies of using TDMA in ethernet pon (EPON) The key technologies when ethernet pon (EPON) adopts TDMA are dynamic bandwidth allocation, uplink channel multiplexing implementation technology, framing technology and implementation technology of Ethernet on PON, ranging and delay compensation, optical devices , fast synchronization technology for burst signals, etc.
(1) Dynamic bandwidth allocation The dynamic bandwidth allocation algorithm is a mechanism to change the upstream bandwidth of each ONU of the ethernet pon (EPON) in real time (on the order of ms/μs).
If static bandwidth allocation is used in ethernet pon (EPON), it is not suitable for variable rate services such as data communication. For example, if bandwidth is statically allocated according to the peak rate, the entire system bandwidth will be quickly exhausted, and the bandwidth utilization rate is very low. Bandwidth utilization is greatly improved.
Through DBA, we can improve the PON upstream bandwidth efficiency by dynamically adjusting the bandwidth between ONUs according to the needs of ONU burst services. Due to more efficient use of bandwidth, network administrators can add more users to an existing PON, and end users can also enjoy better services, such as the peak bandwidth available to users can exceed the traditional fixed allocation method. bandwidth.
According to the characteristics of ethernet pon (EPON) and the ITU-T G.983 recommendation, it can be concluded that the specific requirements for dynamic bandwidth allocation design include: service transparency; high bandwidth utilization; low delay and low delay jitter;
Fair allocation of bandwidth, good robustness, and strong real-time performance. The dynamic bandwidth allocation adopts the centralized control method: any ONU sending uplink information must apply for bandwidth from the OLT, and the OLT grants the bandwidth (time slot) occupation authorization according to a certain algorithm according to the request of the ONU, and the ONU sends the information according to the allocated time slot.
The basic idea of the allocation grant algorithm is: each ONU uses the uplink divisible time slot to reflect the time distribution of cell arrivals and request bandwidth, and the OLT allocates the bandwidth fairly and reasonably according to the request of each ONU, and at the same time considers handling overload and channel errors. , Handling of cell loss, etc.
ONUs may be disconnected frequently (especially in the case of FTTH), and disconnected ONUs should no longer consume network bandwidth. The solutions to this problem are: make new requests arrive within the RTT Dt time after the permission signal; if the request (k) is lost, it means that the ONU has been disconnected; wait for 1 minute before querying this ONU next time.
This allows disconnected ONUs to consume only about 0.0005% of the PON (Passive Optical Network) bandwidth. In the PON, the traffic container is used on the transmission convergence layer to store and manage the service flow of the upstream bandwidth allocation, so that the bandwidth utilization rate of the PON (Passive Optical Network) segment is improved. Analysis shows that with DBA, the maximum bandwidth utilization can reach 80%, compared to 40% without DBA. The average transmission delay is 100ms without DBA and <10ms with DBA.
(2) The time division multiple access multiplexing method of ethernet pon (EPON) uplink implementation technology can also adopt fixed time slot time division multiple access multiplexing, statistical time division multiple access multiplexing, random access and other methods.
The shortcomings of the fixed time slot time division multiplexing method are: when some of the time slots are not used, the same bandwidth is still occupied; the adaptability to high burst rate services is not strong; the ONU needs to be synchronized.
There is no definite access time for random access;
Due to the use of CSMA/CD, there is a limitation of transmission distance. Statistical time-division multiplexing can overcome the shortcomings of the first two, so statistical time-division multiplexing is generally chosen.
(3) The uplink signal transmission sends the Ethernet frame in the time slot allocated by the ONU, and the statistical multiplexing is realized by changing the time slot size by the size of the provided data amount. The implementation method of TDMA is the key point, that is, how to use the TDMA method to make the bandwidth utilization, delay and delay jitter of the uplink channel meet the requirements.
Among them, issues such as the allocation method of the upstream bandwidth, the fixed or variable ONU sending window, the size of the largest ONU sending window, the interval of the ONU sending window, and whether the Ethernet frame is cut need to be studied and determined.
Since the data/video traffic flow is self-similar, there is no fixed time slot of the best size, so the traffic aggregation can not play much role. The burst size distribution has a long correlation (large tail), that is, most burst packets are relatively small, but most of them have a large number of burst packets appearing at the same time. Realize statistical time division.
Issues to consider are:
- The burst time and the size of the business flow are unpredictable, so there must be feedback;
- It is better to use the Roll-call polling method, but it requires ONUs to listen to each other, that is, the PON (Passive Optical Network) can only be used as a broadcast star or passive ring (this is too strict). Hub polling can also be used, but the work time is very long.
The solution is to use interleaved polling routing. Upstream signals should use a flexible request/grant approach.
For example, it can flexibly provide permission signals for each user and each service in an ONU through “logical ports”, realize flexible operation, or adopt “multi-request mode” to shorten the permission period and have high TCP traffic.
2. The key technology of using TDMA in ethernet pon (EPON) Many technologies when ethernet pon (EPON) adopts WDMA and TDMA, such as dynamic bandwidth allocation, uplink channel multiplexing implementation technology, framing technology and implementation technology of Ethernet on PON, fast burst signal Synchronization technology and so on are no longer required, thus simplifying the design of the overall system.
The key technologies when using WDMA are: low-cost wavelength-stable optical transmitter modules, low-cost optical receiver modules, low-cost DWDM, ranging, and so on. For low-cost wavelength-stable optical transmitter modules and low-cost optical receiver modules, the Optical Fiber Teaching and Research Office of Huazhong University of Science and Technology has basically developed them.
3. Other key technologies Other key technologies include: Ethernet framing technology and implementation technology on PON; ranging and delay compensation technology; optical devices supporting high-speed burst signals; fast synchronization of burst signals, etc.
In addition, downlink channel security, how to implement QoS, how to implement VLAN and network management, etc. are also issues that affect the application prospects of ethernet pon (EPON), which must be considered. It can be seen that as long as the cost of the required optical devices is reduced, the ethernet pon (EPON) using the WDMA access method has a great advantage over the TDMA method.