Gpon technology. Description of GPON technology

GEPON technology

This material will discuss technology and equipment for organizing passive optical networks - Passive Optical Network, PON. The main differences between PON and classical optical communication channels are the use of passive equipment - optical splitters - for traffic aggregation and high port density.

It is no secret that consumer demands for the speed of information delivery from the Internet are growing exponentially. Today, in large cities, 10 Mbit/s is completely commonplace. The reasons for this process have remained unchanged for a long time - voice and video transmission, multimedia, television (lately also in high-definition versions). But the bitrates are constantly increasing.

A significant part of the costs of any provider project is borne by the cable infrastructure. Moreover, this takes into account not only the cost of the cable, but also its installation, which, if working in an existing infrastructure, can be very high. And of course, I want the investments to last a long time, not require frequent updates, and have a good supply of the required parameters. From this point of view, optical communication channels today are the most productive and “long-range” way to provide network connections between devices. At the same time, the classic architecture assumes a “point-to-point” topology, when each line has its own dedicated ports on each side, and if it is necessary to create “branches,” the installation of active equipment in the node is required. So it can be most successfully used for single long-distance lines.

However, in some situations, a tree topology may be more convenient, which is interesting from the point of view of scalability and a reduced overall length of cables to be laid. PON is just suitable for such projects. In Russia, networks of this type appeared quite a long time ago, more than five years ago.

And the growth in the number of connected users and the launch of the first Russian fiber to the home (FTTH) projects based on PON shows that the technology has taken root here too.

PON network structure

A PON network consists of several elements - a switch at the communication node, communication lines with passive splitters at network nodes and modems on the subscriber side. Each modem receives all packets from the switch, and time frame multiplexing is used during transmission.

Data transmission in forward channel


Data transmission in the backward channel

ZyXEL today offers equipment of the EPON (IEEE 802.3ah) standard, also called GEPON.

Currently, the equipment is involved in several projects, as well as in testing with providers throughout Russia. This is what we will talk about next. Note that other standards of this type of network differ in speed and other technical characteristics.

The switch allows you to connect up to 32 or even 64 subscribers via one fiber (one port). The total data transfer rate (which is divided between subscribers) is 1.25 Gbit/s. Further development of EPON in the coming years also offers a transition to speeds of 10/1 Gigabit/s and 10/10 Gigabit/s. Next year the working version of the 10G EPON standard is expected to be adopted, and the first pilot projects may start in 2010.

With a delay of two to three years, the transition to 10-gigabit speeds and GPON technologies is planned.

For reception and transmission, lasers with different wavelengths are used - 1490 nm for transmission and 1310 for reception. If necessary, it is possible to add analogue cable television channels (100 or more) to the channel, which are modulated by a 1550 nm laser. Depending on the specific network design and equipment used, the total length of the channel can be up to 20 km.


Multiservice network based on GEPON technology

The cable is laid from the switch port in the form of a tree. Splitters installed in nodes are extremely unpretentious - they do not require power supply, configuration and management, heating cabinets, are inexpensive and very compact. This allows them to be placed, for example, in existing telephone distribution cabinets.


Splitter

The simplest end devices are fiber-to-cable converters with a built-in MAC address filter. When using television, another receiver is installed in the modem, and a regular high-frequency cable is output to the TV.

To protect information, it is possible to use encryption (AES128) of all transmitted packets. The technology does not allow direct communication between individual subscribers located on the same switch port - data from one subscriber can reach another only through a GEPON switch, which relays upstream data streams at a wavelength of 1310 nm to a downstream stream at a wavelength of 1490 nm. An additional advantage from a security point of view is the use of exclusively passive equipment on the line, which makes interception difficult.

Among the positive aspects of PON, it should be noted:

  • minimal use of active equipment;
  • minimizing cable infrastructure;
  • low cost of maintenance;
  • possibility of integration with cable television;
  • good scalability;
  • high density of subscriber ports.

At the same time, when considering the technology, it is necessary to take into account its features, especially in comparison with point-to-point lines: the bandwidth shared between subscribers, the common environment may not be suitable for the client from a security point of view, passive splitters make it difficult to diagnose an optical line, the influence of a fault may be equipment of one subscriber for the work of the rest, less benefit if sold at the construction stage.

Equipment

ZyXEL's GEPON product line consists of three switches and three modems. The low-end model of the switch has eight GEPON ports and eight corresponding Gigabit Ethernet ports (note that Gigabit devices with lower speeds cannot be connected to them). Up to 32 modems can be connected to each optical port, resulting in 256 subscribers per device. All connectors are located on the front side of the device - 8xPON, 8xGigabit, console, 10/100BaseT off-network control and power. There is also a device reset button here. All ports have a set of indicators to determine the current status. It has a built-in gigabit L2+ switch (non-blocking switching with a throughput of 24 Gbit/s, frame switching speed of 17.8 million packets/s) and four combined 1000Base-T/SFP ports. This option can be used for channel redundancy - when two connectors (SC and RJ45) are connected simultaneously, the optics work, and in the event of a failure in the optical channel, it automatically switches to copper. The power supply and console port for this modification are located on the rear panel. These models are made in a standard 1U case and are recommended for use in fast-growing networks. The most productive model is modular. Its 4.5U chassis provides space for up to sixteen OLC-2301s. Each such linear module has a GEPON port and a combined 1000Base-T/SFP port. The chassis also houses a control module and a dual redundant power supply. Linear modules are hot-swappable, which has a positive effect on the ease of network maintenance and reliability of service provision. Maximum OLT-2300 can support 512 subscribers. All optical modules of the switches are designed for an operating range of 20 km.


OLT-1308

The latest firmware updates for the OLT-1308/OLT-1308H models allow 64 rather than 32 subscribers to operate on one channel, which significantly reduces the cost of one connection. There is no such option for OLC-2301 yet.


Chassis OLT-2300

All GEPON switches support STP/RSTP protocols and mechanisms for prioritizing traffic and organizing virtual networks (including Port Based and 802.1Q). The efficiency of multicast broadcasts is ensured by support for IGMP v.2, IGMP proxy, IGMP snooping and MVR. RS-232 and 10/100Base-TX ports are provided for control. Switches can be configured via the Web interface (SSL is supported, up to five accounts can be installed, examples of screenshots are , , ), telnet, SSH, FTP or the console port. The port numbers of all services can be changed. It is possible to restrict access by IP addresses. The web interface has a built-in help system.

The device automatically finds all connected subscriber modems and allows you to assign specific profiles to them. They include settings for speed, filtering, VLAN, priorities and other parameters. The 802.1x authentication protocol can be used.

Switches also allow you to monitor the physical condition - temperatures, fan speeds, and voltages are checked. For large networks, switches will benefit from SNMP support and compatibility with the NetAtlas EMS management system. In addition, it is possible to combine devices into clusters for general management.

At the moment, ZyXEL does not have models with built-in CATV injectors. However, to mix the TV signal into an optical channel, you can use external splitters and coaxial/optical media converters.




ONU-631HA

The first model of a subscriber GEPON modem is the . It operates in bridge mode, is easy to maintain and is controlled exclusively by the provider using a special protocol. For the user, it offers a standard Gigabit Ethernet port. There are two modifications of modems - with indexes -11 and -12. The first works at distances of up to 10 km, and the second - up to 20 km. The case is made of dark plastic; there are several indicators on the front panel (power, PON, LAN, LAN speed, duplex). On the back side there are two network ports (optical and copper) and a power supply input (12 V 1.5 A). This model is positioned for connecting corporate subscribers and operator network extensions.




ONU-634HA

The second model is more interesting for connecting home users - it has a built-in centrally managed 4-port switch with VLAN 802.1Q binding to Fast Ethernet ports. Like the 631, it is fully configured by the provider, which reduces maintenance costs. There are also now ONU-634FA samples - four network ports and a cable TV output, which allows you to directly connect a regular TV to a GEPON modem.




ONU-634FA

Recommended prices for GEPON equipment
ModelCost ($)Cost per subscriber ($)
ONU-631HA-11/12 372/454 372/454
ONU-634HA-11/12 388/502 388/502
OLT-1308 23 939 47
OLT-1308H 23 283 46
OLT-2300M/OLC-2301HA-12 1 317/2 670 90 (for 512 subscribers)

To build a network, you will also need splitters (the approximate cost is from 400 rubles for 1x2 to 4000 rubles for 1x8, there are also 1x32 models), a single-mode optical cable (the cost is equal to the price of a UTP cable: prices for fiber cable start at 7-8 rubles per meter) and connectors (from 100,140 rubles per connection).

Testing of the described equipment as part of the OLT-1308 switch and ONU-631A modems was carried out on the ZyXEL test site using the Ixia Chariot test package. The results for simultaneous operation of one, two and three clients are shown in the table (packets of maximum size, Mbit/s). Modems were connected to one of the switch ports through one splitter. It can be seen that in the case of maximum load, the speeds are evenly distributed across all clients. We also note the high efficiency of data transfer, including the operation mode of several clients - the total speed practically coincides with the maximum possible.

In general, it can be noted that the technology is not difficult to set up and operate and works according to specifications. Speeds correspond to those familiar from gigabit copper networks.

conclusions

GEPON technology can be successfully used to organize optical communication channels to the subscriber and is especially effective if there are restrictions on cable laying and installation of active equipment on the line. The effectiveness of this solution depends on many factors and it is certainly impossible to say unequivocally that this is the best option; everything is determined by the specific requirements of the customer. However, the estimates made allow us to conclude that even today, in some cases, the cost of connecting home subscribers via fiber optics may not exceed $500.

As for the equipment described, ZyXEL today offers a full line of GEPON devices that allows you to create optical networks of any scale with all the necessary control systems and technologies to improve reliability.


all about passive optical networks (PON)

A couple of years ago, we already published a short introduction to passive optical networks (PON). However, at that time, the market was just taking a closer look at this relatively young technology - the first installations of PON networks were just appearing in the world, and their number was just few. There was no talk of PON coming to Belarus at that time. Today the situation has changed: PON has proven itself excellent in large operator networks around the world, and is gradually spreading to the masses, becoming an affordable and attractive last-mile solution for smaller providers as well.
There has also been progress in Belarus - Solo has taken over PON equipment produced by Terawave Communications. Which I gladly reported at the seminar held in Minsk on August 9.
Here is a good reason for a large, detailed and intelligible technical material on PON, the introduction to which you are now reading :)
We will tell you about the equipment in the next issues, keep an eye on the hardware section.

PON network architecture

The development of the Internet, including the emergence of new communication services, contributes to the growth of data flows transmitted over the network and forces operators to look for ways to increase the capacity of transport networks. When choosing a solution you need to consider:
- diversity of subscriber needs;
- potential for network development;
- efficiency.
In the developing telecommunications market, it is dangerous both to make hasty decisions and to wait for more modern technology. Moreover, in the opinion of the authors, such a technology has already appeared - this is the technology of passive optical networks PON (passive optical network).
A PON access distribution network based on a fiber cabling tree architecture with passive optical splitters at the nodes may be the most cost-effective and capable of supporting broadband transmission of a variety of applications. At the same time, the PON architecture has the necessary efficiency in increasing both network nodes and throughput, depending on the present and future needs of subscribers.
The construction of access networks is currently mainly proceeding in four directions:
- networks based on existing copper telephone pairs and xDSL technology;
- hybrid fiber-coaxial networks (HFC);
- wireless network;
- fiber-optic networks.
Using ever-improving xDSL technologies is the easiest and most inexpensive way to increase the capacity of an existing twisted-pair copper cabling system. For operators when it is necessary to provide speeds of up to 1-2 Mbit/s, this path is the most economical and justified. However, transmission speeds of up to tens of megabits per second on existing cable systems, taking into account long distances (up to several km) and low quality copper, seem to be a difficult and quite expensive solution.
Another traditional solution is hybrid fiber-coaxial networks (HFC, Hybrid Fiber-Coaxial). Connecting multiple cable modems to one coaxial segment reduces the average cost of building network infrastructures per subscriber and makes such solutions attractive. In general, the design limitation on bandwidth remains here.
Wireless access networks can be attractive where there are technical difficulties in using cable infrastructures. Wireless communication by its nature has no alternative for mobile services. In recent years, along with traditional solutions based on radio and optical Ethernet access, WiFi technology has become increasingly widespread, allowing for a total bandwidth of up to 10 Mbit/s and in the near future up to 50 Mbit/s.
It should be noted that for the three listed areas, a further increase in network capacity is associated with great difficulties that are not present when using a transmission medium such as fiber.
Thus, the only way to ensure that the network is able to handle new applications that require ever-increasing transmission speeds is to lay an optical cable from the central office to the home or corporate client. This is a very radical approach. And just 5 years ago it was considered extremely expensive. However, nowadays, thanks to a significant reduction in prices for optical components, this approach has become relevant. Today, laying OK for organizing an access network has become beneficial both when updating old and when building new access networks (last miles). There are many options when it comes to choosing fiber optic access technology. Along with the traditional solutions based on optical modems, optical Ethernet, and Micro SDH technology, new solutions using PON passive optical network architecture have emerged.

basic topologies of optical access networks

There are four main topologies for constructing optical access networks: “point-to-point”, “ring”, “tree with active nodes”, “tree with passive nodes”.

point-to-point (P2P)

The P2P topology (Fig. 1) does not impose restrictions on the network technology used. P2P can be implemented for any network standard, as well as for non-standard (proprietary) solutions, such as optical modems. From the point of view of security and protection of transmitted information, a P2P connection ensures maximum security for subscriber nodes. Since CC needs to be routed individually to the subscriber, this approach is the most expensive and is attractive mainly for large subscribers.

Rice. 1. Point-to-point topology.

ring

Ring topology (Fig. 2.) based on SDH has proven itself in urban telecommunication networks. However, not all is well in access networks. If, when building a city highway, the location of nodes is planned at the design stage, then in access networks it is impossible to know in advance where, when and how many subscriber nodes will be installed. With random territorial and temporary connection of users, the ring topology can turn into a heavily broken ring with many branches; new subscribers are connected by breaking the ring and inserting additional segments. In practice, such loops are often combined in one cable, which leads to the appearance of rings that look more like a broken line - “collapsed” rings, which significantly reduces the reliability of the network. In fact, the main advantage of the ring topology is minimized.


Rice. 2. Ring topology.

tree with active nodes

A tree with active nodes (Fig. 3.) is an economical solution in terms of fiber use. This solution fits well within the framework of the Ethernet standard with a hierarchy of speeds from the central node to subscribers 1000/100/10 Mbit/s (1000Base-LX, 100Base-FX, 10Base-FL). However, each tree node must contain an active device (in relation to IP networks, a switch or router). Optical Ethernet access networks, predominantly using this topology, are relatively inexpensive. The main disadvantage is the presence of active devices at intermediate nodes that require individual power supply.


Rice. 3. Topology "tree with active nodes".

tree with passive optical fan-out PON (P2MP)

Solutions based on PON architecture (Fig. 4.) use the point-to-multipoint logical topology P2MP (point-to-multipoint), which is the basis of PON technology; an entire fiber optic segment of the tree architecture can be connected to one port of the central node, covering dozens of subscribers. At the same time, compact, completely passive optical splitters are installed in the intermediate nodes of the tree, which do not require power or maintenance.


Rice. 4. Topology "Tree with passive optical branching".

It is well known that PON allows you to save on cable infrastructure by reducing the total length of optical fibers, since only one fiber is used in the section from the central node to the splitter. Less attention is paid to another source of savings - a reduction in the number of optical transmitters and receivers in the central node. Meanwhile, the savings of the second factor in some cases turn out to be even more significant. Thus, according to NTT estimates, a PON configuration with a splitter in the central office in close proximity to the central node turns out to be more economical than a point-to-point network, although there is practically no reduction in the length of the optical fiber! Moreover, if the distances to subscribers are not great (as in Japan), taking into account operating costs (in Japan this is a significant factor), it turns out that PON with a splitter in the central office is more economical than PON with a splitter close to the subscriber nodes.
Advantages of PON architecture:
- absence of intermediate active nodes; fiber saving;
- saving of optical transceivers in the central node;
- ease of connecting new subscribers and ease of maintenance (connection, disconnection or failure of one or more subscriber nodes does not in any way affect the operation of the others).
The P2MP tree topology allows you to optimize the placement of optical splitters based on the actual location of subscribers, the cost of laying cables and operating the cable network.
Disadvantages include the increased complexity of PON technology and the lack of redundancy in the simplest tree topology.

operating principle of PON

The main idea of ​​the PON architecture is the use of just one transceiver module in the OLT to transmit information to multiple ONT subscriber devices and receive information from them. The implementation of this principle is shown in Fig. 5.
The number of subscriber nodes connected to one OLT transceiver module can be as large as the power budget and maximum speed of the transceiver equipment allows. To transmit information flow from OLT to ONT - direct (downstream) flow, as a rule, a wavelength of 1550 nm is used. On the contrary, data streams from different subscriber nodes to the central node, which together form the reverse (downstream) stream, are transmitted at a wavelength of 1310 nm. OLT and ONT have built-in WDM multiplexers that separate outgoing and incoming streams.


Rice. 5. Basic elements of PON architecture and operating principle

direct flow

The direct stream at the level of optical signals is broadcast. Each ONT, reading the address fields, selects from this general stream a part of the information intended only for it. In fact, we are dealing with a distributed demultiplexer.

reverse flow

All ONT subscriber nodes transmit in the reverse stream on the same wavelength, using the TDMA (time division multiple access) concept. In order to eliminate the possibility of signals from different ONTs crossing, each of them has its own individual data transmission schedule, taking into account an adjustment for the delay associated with the removal of this ONT from the OLT. The TDMA MAC protocol solves this problem.

PON standards

The first steps in PON technology were taken in 1995, when an influential group of seven companies (British Telecom, France Telecom, Deutsche Telecom, NTT, KPN, Telefonica and Telecom Italia) created a consortium in order to implement the ideas of multiple access over a single fiber. . This informal organization, supported by ITU-T, is called FSAN (full service access network). Many new members - both operators and equipment manufacturers - joined in the late 90s. The goal of FSAN was to develop common guidelines and requirements for PON equipment so that equipment manufacturers and operators could coexist together in a competitive PON access system market. Today FSAN has 40 operators and manufacturers and works closely with standards organizations such as ITU-T, ETSI and the ATM Forum.

Some ITU-T standards governing xPON technology.

APON/BPON

In the mid-90s, the generally accepted view was that only the ATM protocol was capable of guaranteeing acceptable quality of QoS communication services between end subscribers. Therefore, FSAN, wanting to provide transport of multiservice services through the PON network, chose ATM technology as a basis. As a result, in October 1998, the first ITU-T G.983.1 standard appeared, based on the transport of ATM cells in the PON tree and called APON (ATM PON). Then, over the course of several years, many new amendments and recommendations appeared in the G.983.x series (x=1–7), the transmission speed increased to 622 Mbit/s. In March 2001, recommendation G.983.3 appeared, adding new entities to the PON standard:
- transmission of various applications (voice, video, data) - this actually allowed manufacturers to add appropriate interfaces on the OLT for connecting to the backbone network and on the ONT for connecting to subscribers;
- expansion of the spectral range – opens up the possibility for additional services on other wavelengths under the same PON tree, for example, broadcast television on a third wavelength (triple play).
The APON standard extended in this way is given the name BPON (broadband PON).
APON today allows dynamic bandwidth allocation (DBA) between different applications and different ONTs and is designed to provide both broadband and narrowband services.
APON equipment from different manufacturers supports trunk interfaces: SDH (STM-1), ATM (STM-1/4), Fast Ethernet, Gigabit Ethernet, video (SDI PAL), and subscriber interfaces E1 (G.703), Ethernet 10/100Base -TX, telephony (FXS).
Due to the broadcast nature of the forward stream in the PON tree and the potential for unauthorized access to data by an ONT to which the data is not intended, APON provides the ability to forward stream data using encryption techniques with public keys. There is no need to encrypt the return stream since the OLT is located on the operator's premises.

PON G.983.1 Standard Basics

In November 2000, the LMSC (LAN/MAN standards committee) IEEE created a special commission called “Ethernet in the first mile” (EFM, Ethernet in the first mile) 802.3ah, thereby realizing the wishes of many experts to build a PON network architecture that is as close to to the currently widespread Ethernet networks. In parallel, the EFMA alliance (Ethernet in the first mile alliance) is being formed, which was created in December 2001. In fact, the EFMA alliance and the EFM commission complement each other and work closely on the standard. While EFM focuses on technical issues and standard development within the IEEE, EFMA studies more the industrial and commercial aspects of using new technology. The goal of joint work is to achieve consensus between operators and equipment manufacturers and develop IEEE standard 802.3ah, fully compatible with the developing IEEE 802.17 backbone packet ring standard.
The EFM 802.3ah Commission must standardize three types of access network solutions:
EFMC (EFM copper) – point-to-point solution using twisted copper pairs. To date, work on this standard has been almost completed. Of the two alternatives between which the main struggle unfolded - G.SHDSL and ADSL+ - the choice was made in favor of G.SHDSL.
EFMF (EFM fiber) – solution based on point-to-point fiber connection. Here it is necessary to standardize various options: “duplex over one fiber, at the same wavelengths”, “duplex over one fiber, at different wavelengths”, “duplex over a pair of fibers”, new options for optical transceivers. Similar solutions have been offered by a number of companies as “proprietary” for several years. It's time to standardize them.
EFMP (EFM PON) – solution based on a point-to-multipoint connection over fiber. This solution, which is essentially an alternative to APON, received the similar name EPON.
Currently, the development of 802.3ah standards, including EFMP, is in its final stages, and adoption is expected this year. Arguments in favor of EPON technology are strengthened by the Internet's focus solely on the IP protocol and Ethernet standards.

GPON

The architecture of the GPON access network (Gigabit PON) can be considered as an organic continuation of APON technology. At the same time, both an increase in the bandwidth of the PON network and an increase in the efficiency of transmission of various multiservice applications are realized. GPON standard ITU-T Rec. G.984.3 GPON was adopted in October 2003.
GPON provides a scalable frame structure at transmission rates from 622 Mbps to 2.5 Gbps, supports both symmetric and asymmetric bit rates in the PON tree for downstream and upstream and is based on the ITU-T G.704.1 GFP standard ( generic framing protocol), providing encapsulation into a synchronous transport protocol of any type of service (including TDM). Studies show that even in the worst case of traffic distribution and flow fluctuations, bandwidth utilization is 93% compared to 71% in APON, not to mention EPON.
If in SDH the band division occurs statically, then GFP (generic framing protocol), while maintaining the structure of the SDH frame, allows for dynamic band allocation.

comparison of APON, EPON, GPON technologies

The table provides a comparative analysis of these three technologies.

Notes:
1 – discussed in the project.
2 – the standard allows network expansion up to 128 ONT.
3 – transmission is allowed in the forward and reverse directions at the same wavelength.
4 – carried out at higher levels.

more about APON

And now - some purely technical specifics about how PON networks work. The APON variety is taken as an example.
The interaction of the subscriber node with the central one begins with the establishment of a connection. After which the data is transferred. All this is done according to the APON MAC protocol. During the connection establishment process, a ranking procedure is launched, which includes: distance ranking, power ranking and synchronization. The central node, like a conductor, ensures the coordinated work of all subscriber nodes - orchestra members.

APON MAC - protocol for interaction between the central node and subscriber

The MAC protocol for APON access systems solves three problems:
- elimination of collisions between transmissions in the reverse flow;
- clear, efficient, dynamic division of the return flow strip;
- Maintaining the best possible negotiation for transport of end-user initiated applications.
The APON MAC protocol is based on a request/grant mechanism. The main idea is to send requests from the ONT to the required band. Based on knowledge of how the reverse flow is loaded, and what services are a priori assigned to a particular ONT, the OLT makes a decision on processing these requests.

ranking procedures

The PON network initialization is based on three procedures: determining distances from OLT to different ONTs (distance ranging); synchronization of all ONTs (clock ranging); and determining, when receiving at the OLT, the intensities of optical signals from different ONTs (power ranging).

ranking by distance

Distance ranging - determining the time delay associated with the removal of ONT from OLT - is performed at the stage of registration of subscriber nodes, and is required in order to ensure collision-free transport and create a unified synchronization in the reverse flow.
First, the network administrator enters data about the new ONT into the OLT, it serial number, parameters of services provided by ONT. Then, after physically connecting this subscriber node to the PON network and turning on its power, the central node begins the ranking process. Ranking with ONT, which is registered in the OLT registry, occurs every time the ONT is turned on. When power is turned off and on at the OLT, ranging occurs with all registered ONTs.
The OLT, sending a signal to the ranked ONT, listens to the response from it and, based on this, calculates the time delay on the double trip RTT (round trip time), then transmits the calculated value to the ONT in the forward stream. Based on this, the ONT subscriber node introduces an appropriate delay, which precedes the start of sending the frame in the reverse stream. Subscriber nodes located at different distances will introduce different delays. In this case, the sum of the introduced hardware delay and the delay in propagation of the light signal along the optical path from ONT to OLT will be the same for all subscriber nodes.
Taking into account the fact that OLT-ONT distances can vary within wide limits (the G.983.1 standard defines the range 0-20 km), let's estimate possible delay variations. If we take into account that the speed of light in a fiber is 2*105 km/s, then an increase in the OLT-ONT distance by 1 km will correspond to an increase in the delay time on a double path by 10 μs. And for a distance of 20 km, the RTT will be 0.2 ms. In fact, this is the minimum theoretical time it takes for an OLT to perform ranking with a single ONT. Ranging by distance of a larger number of subscriber nodes occurs sequentially and requires a proportional increase in the total ranking time. During this time, the reverse flow cannot be used for data transmission by other ONTs.
After the distance ranking is completed, the OLT, based on the prescribed services for each ONT and using the MAC protocol, decides which subscriber node to transmit in each specific time slot.
Note that the total delay when sending a frame to the reverse stream is introduced not only by the finite time of signal propagation along the fiber, but also by the OLT and ONT electronics elements. The delay on the part of the latter may experience slight drift, for example due to fluctuations in equipment temperature. Therefore, at the data transmission stage, the OLT informs the ONT about small adjustments to the delay introduced into the reverse flow - micro ranging. As a result, the accuracy with which sent frames from different ONTs are stabilized is 2–3 bits.

ranking by power

Power ranging - changing the discrimination threshold of a photodetector in order to increase the sensitivity of the photodetector or to avoid its unwanted saturation. Since ONTs are located at different distances from the OLT, the insertion losses in optical signals propagating through the PON tree will be different. This can lead to malfunction of photodetectors due to signal weakness or due to overload.
There are two possible ways out of this situation - either adjust the power of the ONT transmitters, or adjust the response threshold on the OLT photodetector. The second option was chosen as more reliable.
The OLT photodetector threshold is adjusted each time a new ATM packet is received from the reverse preamble flow based on the measurement of the integral power in the packet preamble.
Power adjustment is also required on all ONTs. This is done in a similar way, but only once before synchronizing the receiver to work with the synchronous TDM stream from the OLT. Then the integrated power on the ONT is continuously calculated, and the discrimination threshold of the photodetector is smoothly adjusted.

synchronization

Synchronization or phase ranging is necessary for both forward and reverse flow.
ONT subscriber nodes are synchronized at the beginning of their initialization and then maintain synchronization all the time, adjusting to continuous TDM traffic from the OLT, and carrying out what is commonly called synchronous data reception.
On the contrary, the central OLT node is synchronized each time according to the preamble of a newly arriving ATM packet. Knowing the time delay calculated at the distance ranking stage on the part of the ONT that sent this packet is not enough here - greater accuracy is required. The method of receiving data with preamble synchronization is usually called asynchronous. Preamble synchronization is similar to the solution in ten-megabit Ethernet technology with a preamble size of 64 bits (8 bytes). However, keeping the preamble size the same for a relatively small ATM packet (in the upstream) would be a huge inefficient use of bandwidth. For APON technology, a new synchronization technique has been developed, based on the CPA (clock phase alignment) method, which allows the necessary synchronization to be carried out after receiving only three bits! The larger upstream ATM packet preamble size was chosen because the preamble also serves the function of providing a power ranking procedure.

APON frame structure for forward and reverse stream

To manage the request/grant mechanism, FSAN has defined an APON frame structure for the forward and reverse flows. This format has been standardized by ITU-T in recommendation G.983.1. In Fig. Figure 6 shows the APON frame format for symmetrical traffic mode 155/155Mbit/s. A downstream frame consists of 56 ATM cells of 53 bytes. The upstream frame consists of 52 ATM packets of 56 bytes each and one MBS slot with a total length of also 56 bytes, discussed below.


Rice. 6. ITU G.983 frame format - forward and reverse stream frame structure.

direct flow

Transmission permissions are sent in bursts in special ATM service cells - two per frame, which are called PLOAM (physical layer operation and maintenance) cells. They follow strictly regularly, alternating with 27 data cells. One PLOAM cell contains 26 permissions for ONT, each for transmitting just one (!) ATM packet. The remaining 54 cells in the forward frame carry data and are not used for the request/permit mechanism.

reverse flow

The reverse flow represents a collection of data bursts from different ONTs. The subscriber node can transmit data only after receiving the appropriate permission read from the PLOAM cell. Data packets from ONT to APON are transmitted in ATM packets. The only difference between an ATM packet and a cell is that the packet has an additional 3 byte preamble. Thus, the length of an ATM packet is 56 bytes. The preamble is not needed for cells in the forward stream due to the synchronous data reception mode, as discussed above. The first two bits of the preamble do not contain an optical signal, which is sufficient to eliminate the overlap of packets from different ONTs - slight fluctuations in the delay during signal propagation are inevitable in the line.
If we take into account that a transmission grant is required for each ATM packet, then the total number of grants written in PLOAM cells over a long period of time must correspond to the number of ATM packets emitted by all ONTs during this time. Why does PLOAM fit 26 permits? Two PLOAM cells can grant permission to transmit 52 ATM packets, the same number as there are in the upstream ATM frame.

MBS slot

The MBS (multi burst slot) slot in the reverse stream is a service slot. It informs the OLT about the nature of the transmission requests from the ONT. This slot has 8 subfields or minislots corresponding to different ONTs (Fig. 7). If the PON system is designed for 32 subscriber nodes, then all 32 ONTs will be able to transmit their information about transmission requests only after four sequentially transmitted MBS slots, which constitutes a cycle. In a 64 ONT system, a cycle consists of eight MBS slots. Transmission of one frame at a speed of 155 Mbit/s lasts 0.15 ms. It will take 0.6 ms to transmit the entire cycle with 32 ONTs. In other words, with a frequency of 0.6 ms, the ONT sends service requests about the intent to transmit. The ONT sends a request when a transmission queue has been formed in its output buffer. Since the ONT will be able to transmit only after receiving permission in the PLOAM cell, in order to estimate the maximum time from the moment the queue is prepared in the buffer until the start of transmission, you should add a delay on the double RTT run to the cycle time of 0.6 ms (for a network with a radius 20 km RTT is 0.2 ms), resulting in 0.8 ms. Hardware delays on the OLT and ONT can be added to this value.


Rice. 7. MBS slot structure.

A minislot consists of 4 fields: a preamble (3 bytes), similar to the preamble in an ATM packet; two fields ABR/GFR and VBR, 8 and 16 bits long, corresponding to two types of bandwidth requests; CRC checksum fields (8 bits).

reliability and redundancy in APON

The weakness of APON access systems with a simple tree topology is the lack of redundancy. The worst-case scenario in this case would be that the fiber running from the OLT to the nearest splitter (feeder fiber) is damaged. The entire segment connected via this fiber loses connection - dozens of subscriber nodes, hundreds of subscribers are left without a network. Mean Time To Repair (MTTR) can vary widely from several days to several weeks depending on the operator. In this case of a single fiber failure, the disadvantage of the PON network compared to the SDH ring topology is most clearly demonstrated.
Therefore, already in the first recommendation of G.983.1 in Appendix IV, the issue of building secure APON systems was discussed. Due to the specifics of PON topology, this task is not as simple as in SDH ring topologies, since the reverse flow band in PON is common and is formed by many subscriber nodes. The G.983.1 recommendations suggested studying four different topologies. Only two of them were finally selected for development in the later recommendation G.983.5.
In Fig. 8-10 shows the main construction options backup systems PON. The first solution (Fig. 8) provides partial redundancy from the central node. To implement this solution, a 2xN splitter is required. The central node is equipped with two optical modules LT-1 and LT-2, in which two fibers are terminated. In normal mode, in the absence of damage to the fibers, the main channel is active, and duplex transmission is organized through it. Reserve channel - inactive - laser diode on LT-2 is turned off. The photodetector on the LT-2 can listen to the reverse flow. If the main channel fiber coming from the central node is damaged, the LT-2 transceiver system is automatically activated, and the multiplexing, switching and cross-connect module on the OLT switches to it, providing transport from the backbone interfaces. To increase reliability, it is advisable to take feeder fibers from different, physically separated optical cables.


Rice. 8. Protected PON topology. Partial redundancy from the central node.

Partial redundancy on the part of the subscriber node (Fig. 9) makes it possible to increase the reliability of the subscriber node. In this case, two optical modules LT-1 and LT-2 are required per subscriber node. Switching to the backup channel is similar to the previous option. When reserving subscriber nodes, it is not necessary to connect all subscriber nodes via the backup stream. The difference in cost of subscriber nodes with redundancy (two modules LT-1 and LT-2) and without it (one LT module) allows differentiated offering of services to different categories of subscribers.


Rice. 9. Protected PON topology. Partial redundancy on the part of the subscriber node.

In Fig. Figure 10 shows an option with full redundancy of the PON system. The system becomes resistant to both failure of OLT and ONT receiving and transmitting equipment, as well as damage to any section of the fiber-optic cable system. Information flows on the ONT are generated simultaneously by both nodes LT-1 and LT-2 and transmitted to two parallel reverse flows. At the OLT, only one version of the two copies of the signals is transmitted further down the backbone. Traffic duplication occurs in the same way in the forward stream. If the fiber or transceiver interfaces are damaged, switching to the backup stream will be very fast and will not lead to interruption of communication.


Rice. 10. Protected PON topology. Full reservation.

The first solution, in addition to providing only partial redundancy, requires a long reconfiguration time when the fiber is damaged. The main contribution to the delay comes from heating the laser on the OLT (LT-2) and performing the ranging procedure. It is practically difficult not to exceed 50 ms, one of the requirements formulated in recommendation G.983.5.
Conclusion. For the considered configurations proposed by ITU-T, almost only a fully redundant solution satisfies all requirements and seems to be the most attractive.

Petrenko I.I., Ubaydullaev R.R., Ph.D., Telecom Transport.

Or Gigabit PON began to be implemented relatively recently. Let's figure out what became the prerequisites for the emergence of GPON technology, what prospects it has, and also compare it with competing technologies - PON and GEPON.

2014 will mark the 45th anniversary of the first computer communication session conducted in the United States at a distance of about 640 km. This event is considered the beginning of the birth of the Internet. Truth preceding World Wide Web The ARPANET network at that time was available to a very narrow circle of people and organizations. Connecting to it for lucky “outsiders” who have computers became possible only in 1991. It was only the appearance of the NCSA Mosaic web browser in 1993 that provided the prerequisite for the explosive growth of the global Internet audience. So the history of the “mass Internet” as of 2013 is only 20 years old.

In the first decade of development global network Among the users who paid attention to such an indicator as “communication channel throughput (data transfer rate in bits)” or the associated “bandwidth” characteristic, there were “a few” people familiar with the theoretical foundations of radio engineering. And today everyone is talking about “Internet speeds”. And everyone wants to have “high-speed Internet” at their disposal.

Why high-speed? And where is the limit from which Internet access can be considered “high-speed”?

The mass user associates Internet speed, first of all, with the time intervals for downloading “heavy” video, music and graphic files, the number of which on the Internet is growing exponentially, and they themselves are “enlarging”. Corporate consumers of online services (and more recently also cloud services) need high speed response to requests in the business management systems they use.

This means that high-speed Internet is an urgent need, and not a whim (for both “users” and companies). The “border” from which high-speed Internet begins, today, according to experts, is at the level of 10 Mb/s.

"Optics" is replacing "copper"

The Worldwide Computer Network began to develop on the basis of existing telephone lines using xDSL technologies. The most “advanced” version of this “copper” family - ADSL2+ modem technology provides an incoming stream speed of 24 Mb/s (outgoing - 1.2 Mb/s). Currently, it is the undisputed leader in the number of connections in all countries of the world. However, “copper” communication lines, laid decades ago, are becoming obsolete both physically and morally and are gradually being replaced by FTTx optical networks, the use of which makes it possible to increase the speed of information exchange on the Internet by two orders of magnitude. And in the near future - even more.

Over the last five years, the process of replacing copper cable routes with optical ones has been increasing and, according to analysts, in another five years the “optics/copper” ratio in telecommunications will radically change in favor of “optics”.

The FTTx (Fiber to the x) architecture is a section of a fiber optic communication line connected on one side to an OLT (Optical Line Terminal) transceiver station installed at the operator, and on the other to the subscribers’ transceiver modules - ONT (Optical Network Terminal) or ONU (Optical Network Unit).

ONT is a personal use terminal (also called an optical modem) installed in an apartment. ONU - designed for installation in the distribution cabinet of an apartment building and has several ports for connecting computers, televisions, telephones located in neighboring apartments.

ONTs and ONUs convert optical signals received from the OLT into electrical signals (sent, for example, to computers, televisions, telephones), and also perform the reverse conversion of electrical signals received from user terminals into optical signals that are sent to the OLT.

If splitters (passive signal splitters coming from OLT) are introduced into a section of an optical line and ONTs are connected to their outputs, then such a transition from a single-fiber FTTx structure to a tree structure will lead to the formation of a passive optical network - PON(Passive Optical Network).

The work of PON is to organize multiple access through a single optical fiber through time division multiplexing access (TDMA) and frequency division of reception and transmission paths ( Wavelength-Division Multiplexing- WDM). WDM multiplexers operating as part of OLT and ONT separate forward (incoming) and reverse (outgoing) signals, broadcast at different wavelengths (forward - 1.49 microns, reverse - 1.31 microns). To these streams can be added a cable television signal transmitted at a wavelength of 1.55 microns.

The first seeds of PON technology appeared about 15 years ago, and since then, the International Telecommunication Union (ITU) has released five standards for data transmission over optical fiber. Active equipment manufactured in accordance with the requirements of these standards provides speeds from 155 Mb/s to 2488 Mb/s. The features of these standards will be discussed below, but for now we emphasize that the advantages common to all types of PON technologies are the ability to easily expand the subscriber base, its maintenance and modernization, as well as low (compared to “copper” technologies) operating costs.

GPON: the driving force behind the standard

The first standard of the PON family - APON (ATM PON) was approved by the MES at the end of 1998, and the following year American and Japanese telecom operators began building passive optical lines. Data transmission according to this standard is carried out on the basis of the ATM protocol, which describes a switching and multiplexing method based on data transmission in the form of fixed-size cells (ATM cells). Data transfer speed - 155 Mb/s.

The introduction of new technologies into APON, in particular, dynamic bandwidth assignment depending on applications, support for SDH, FE, GE, SDI PAL, El, E/FE and telephony protocols, provided additional functionality in the areas of voice broadcasting, various video content and television broadcasting (the first appearance of a third wavelength in PON). This led to the approval of the “daughter” standard of APON - BPON (Broadband PON). At the same time, the data transfer speed increased to 622 Mb/s.

The next “link in the chain” APON - BPON was the GPON standard ( Gigabit capable Passive Optical Network), the implementation of which ensures the network operates in both symmetric and asymmetric modes. The second mode is most often used, in which the data transfer speed in the forward stream reaches 2.488 Gb/s, and in the reverse stream - 1.244 Gb/s (usually these numbers are rounded and spoken of 2.5 Gb/s and 1.25 Gb/s).

Typically, a home PC is connected to an optical modem (ONT) of a GPON network either via twisted pair cable or wireless connection (Wi-Fi). The ONT also has ports for connecting a TV and VoIP phone.

The basic protocol in GPON technology has become GFP (Generic Framing Protocol), although TDMA, SDH, Ethernet, and ATM recommendations are also used.

In parallel with the improvement of PON technologies in the world, the development of optical Ethernet networks took place and the achievements of this communication “branch” in the field of high-speed data transmission were used in the EPON (Ethernet PON) standard, which was developed based on the MPCP (Multi-Point Control Protocol) protocol, managing multiple nodes. And its improved version - GEPON(Gigabit EPON) in its characteristics and capabilities today is second only to the undisputed leader of PON technologies - GPON.

What “catches your eye” in the above mini-review of technologies used in passive optical networks? - The fact that the differences in their functionality are mainly due to which data transfer protocols form the basis of the standards.

GPON and GEPON: simple arithmetic

If numerical indicators (or even descriptions) are known that express any characteristics of the objects that need to be compared, then such a comparison is quite simple to make by placing the corresponding numbers in a row or column. And it will immediately be clear “who is better than whom.” Let's make this comparison between GPON and GEPON.

So, the forward transmission speed of GPON is 2.5 Gb/s, and that of GEPON is 1.25 Gb/s.

The maximum number of subscriber nodes per fiber for GPON is 64, and for GEPON - 16, which leads to a lower cost of port per subscriber in the operator's optical terminal manufactured according to the GPON standard, and significantly lower power consumption by station equipment than when using operator equipment GEPON standard.

Bandwidth utilization using GPON technology is no less than 93%, and using GEPON technology is no more than 60%. This difference is due to the fact that active GPON equipment uses GEM (GTC Encapsulation Method) frame fragmentation technology, which increases the efficiency of bandwidth use. GEPON technology does not have such a tool.

That's all the “simple arithmetic” that explains the popularity of GPON.

GPON: house wiring cables

The GPON network consists of trunk and distribution lines. The length of GPON backbone routes currently reaches 20 km (in the coming years, the developers of GPON technology promise to increase the maximum length of backbone optical fiber to 60 km). The main sections are being laid (more about laying fiber optic cable) using traditional methods of aerial or underground laying of optical cables with a protective sheath, which ensures the durability of the cable line in conditions of high humidity and temperature changes.

For GPON distribution infrastructure, created, for example, within an apartment building, drop and riser cables are used. A feature of “storey” drop cables, intended for branching an optical line from an overhead distribution cable, is the possibility of “flexible” routing with small bending radii provided by their design.

Riser cables used for vertical interfloor wiring contain 6-12 optical fibers, which are easily laid in cassettes, and their welding requires significantly less time than with welding of optical fibers other types of cables.

GPON: the speed of evolution is accelerating

The advantages of the GPON standard compared to other types of PON technologies have been undeniable since its approval in 2003. However, by 2010, in Russia there were only 80 thousand broadband users based on GPON, according to J’son & Partners Consulting. The main barrier to greater growth, as is almost always the case with a product entering the market, was the high price of active optical equipment. In the last few years, prices for station transceivers and subscriber optical modems have decreased significantly, due to which by the beginning of 2017 (according to analysts of the same company), the number of Russian GPON users will approach 6 million, that is, it will increase almost 75 times over the next seven years!

Or Gigabit PON began to be implemented relatively recently. Let's figure out what became the prerequisites for the emergence of GPON technology, what prospects it has, and also compare it with competing technologies - PON and GEPON.

2014 will mark the 45th anniversary of the first computer communication session conducted in the United States at a distance of about 640 km. This event is considered the beginning of the birth of the Internet. True, the ARPANET network, which preceded the World Wide Web, at that time was accessible to a very narrow circle of people and organizations. Connecting to it for lucky “outsiders” who have computers became possible only in 1991. It was only the appearance of the NCSA Mosaic web browser in 1993 that provided the prerequisite for the explosive growth of the global Internet audience. So the history of the “mass Internet” as of 2013 is only 20 years old.

In the first decade of the development of the global network, among the users who paid attention to such an indicator as “communication channel capacity (data transfer rate in bits)” or the associated characteristic “bandwidth”, there were “a few” people familiar with the theoretical foundations of radio engineering . And today everyone is talking about “Internet speeds”. And everyone wants to have “high-speed Internet” at their disposal.

Why high-speed? And where is the limit from which Internet access can be considered “high-speed”?

The mass user associates Internet speed, first of all, with the time intervals for downloading “heavy” video, music and graphic files, the number of which on the Internet is growing exponentially, and they themselves are becoming “larger”. Corporate consumers of online services (and more recently also cloud services) need high speed response to requests in the business management systems they use.

This means that high-speed Internet is an urgent need, and not a whim (for both “users” and companies). The “border” from which high-speed Internet begins, today, according to experts, is at the level of 10 Mb/s.

"Optics" is replacing "copper"

The Worldwide Computer Network began to develop on the basis of existing telephone lines using xDSL technologies. The most “advanced” version of this “copper” family - ADSL2+ modem technology provides an incoming stream speed of 24 Mb/s (outgoing - 1.2 Mb/s). Currently, it is the undisputed leader in the number of connections in all countries of the world. However, “copper” communication lines, laid decades ago, are becoming obsolete both physically and morally and are gradually being replaced by FTTx optical networks, the use of which makes it possible to increase the speed of information exchange on the Internet by two orders of magnitude. And in the near future - even more.

Over the last five years, the process of replacing copper cable routes with optical ones has been increasing and, according to analysts, in another five years the “optics/copper” ratio in telecommunications will radically change in favor of “optics”.

The FTTx (Fiber to the x) architecture is a section of a fiber optic communication line connected on one side to an OLT (Optical Line Terminal) transceiver station installed at the operator, and on the other to the subscribers’ transceiver modules - ONT (Optical Network Terminal) or ONU (Optical Network Unit).

ONT is a personal use terminal (also called an optical modem) installed in an apartment. ONU - designed for installation in the distribution cabinet of an apartment building and has several ports for connecting computers, televisions, telephones located in neighboring apartments.

ONTs and ONUs convert optical signals received from the OLT into electrical signals (sent, for example, to computers, televisions, telephones), and also perform the reverse conversion of electrical signals received from user terminals into optical signals that are sent to the OLT.

If splitters (passive signal splitters coming from OLT) are introduced into a section of an optical line and ONTs are connected to their outputs, then such a transition from a single-fiber FTTx structure to a tree structure will lead to the formation of a passive optical network - PON(Passive Optical Network).

The work of PON is to organize multiple access through a single optical fiber through time division multiplexing access (TDMA) and frequency division of reception and transmission paths ( Wavelength-Division Multiplexing- WDM). WDM multiplexers operating as part of OLT and ONT separate forward (incoming) and reverse (outgoing) signals, broadcast at different wavelengths (forward - 1.49 microns, reverse - 1.31 microns). To these streams can be added a cable television signal transmitted at a wavelength of 1.55 microns.

The first seeds of PON technology appeared about 15 years ago, and since then, the International Telecommunication Union (ITU) has released five standards for data transmission over optical fiber. Active equipment manufactured in accordance with the requirements of these standards provides speeds from 155 Mb/s to 2488 Mb/s. The features of these standards will be discussed below, but for now we emphasize that the advantages common to all types of PON technologies are the ability to easily expand the subscriber base, its maintenance and modernization, as well as low (compared to “copper” technologies) operating costs.

GPON: the driving force behind the standard

The first standard of the PON family - APON (ATM PON) was approved by the MES at the end of 1998, and the following year American and Japanese telecom operators began building passive optical lines. Data transmission according to this standard is carried out on the basis of the ATM protocol, which describes a switching and multiplexing method based on data transmission in the form of fixed-size cells (ATM cells). Data transfer speed - 155 Mb/s.

The introduction of new technologies into APON, in particular, dynamic bandwidth assignment depending on applications, support for SDH, FE, GE, SDI PAL, El, E/FE and telephony protocols, provided additional functionality in the areas of voice broadcasting, various video content and television broadcasting (the first appearance of a third wavelength in PON). This led to the approval of the “daughter” standard of APON - BPON (Broadband PON). At the same time, the data transfer speed increased to 622 Mb/s.

The next “link in the chain” APON - BPON was the GPON standard ( Gigabit capable Passive Optical Network), the implementation of which ensures the network operates in both symmetric and asymmetric modes. The second mode is most often used, in which the data transfer speed in the forward stream reaches 2.488 Gb/s, and in the reverse stream - 1.244 Gb/s (usually these numbers are rounded and spoken of 2.5 Gb/s and 1.25 Gb/s).

Typically, a home PC is connected to an optical modem (ONT) of a GPON network either via twisted pair cable or wireless connection (Wi-Fi). The ONT also has ports for connecting a TV and VoIP phone.

The basic protocol in GPON technology has become GFP (Generic Framing Protocol), although TDMA, SDH, Ethernet, and ATM recommendations are also used.

In parallel with the improvement of PON technologies in the world, the development of optical Ethernet networks took place and the achievements of this communication “branch” in the field of high-speed data transmission were used in the EPON (Ethernet PON) standard, which was developed based on the MPCP (Multi-Point Control Protocol) protocol, managing multiple nodes. And its improved version - GEPON(Gigabit EPON) in its characteristics and capabilities today is second only to the undisputed leader of PON technologies - GPON.

What “catches your eye” in the above mini-review of technologies used in passive optical networks? - The fact that the differences in their functionality are mainly due to which data transfer protocols form the basis of the standards.

GPON and GEPON: simple arithmetic

If numerical indicators (or even descriptions) are known that express any characteristics of the objects that need to be compared, then such a comparison is quite simple to make by placing the corresponding numbers in a row or column. And it will immediately be clear “who is better than whom.” Let's make this comparison between GPON and GEPON.

So, the forward transmission speed of GPON is 2.5 Gb/s, and that of GEPON is 1.25 Gb/s.

The maximum number of subscriber nodes per fiber for GPON is 64, and for GEPON - 16, which leads to a lower cost of port per subscriber in the operator's optical terminal manufactured according to the GPON standard, and significantly lower power consumption by station equipment than when using operator equipment GEPON standard.

Bandwidth utilization using GPON technology is no less than 93%, and using GEPON technology is no more than 60%. This difference is due to the fact that active GPON equipment uses GEM (GTC Encapsulation Method) frame fragmentation technology, which increases the efficiency of bandwidth use. GEPON technology does not have such a tool.

That's all the “simple arithmetic” that explains the popularity of GPON.

GPON: house wiring cables

The GPON network consists of trunk and distribution lines. The length of GPON backbone routes currently reaches 20 km (in the coming years, the developers of GPON technology promise to increase the maximum length of backbone optical fiber to 60 km). The main sections are being laid (more about laying fiber optic cable) using traditional methods of aerial or underground laying of optical cables with a protective sheath, which ensures the durability of the cable line in conditions of high humidity and temperature changes.

For GPON distribution infrastructure, created, for example, within an apartment building, drop and riser cables are used. A feature of “storey” drop cables, intended for branching an optical line from an overhead distribution cable, is the possibility of “flexible” routing with small bending radii provided by their design.

Riser cables used for vertical interfloor wiring contain 6-12 optical fibers, which are easily laid in cassettes, and their welding requires significantly less time than with welding of optical fibers other types of cables.

GPON: the speed of evolution is accelerating

The advantages of the GPON standard compared to other types of PON technologies have been undeniable since its approval in 2003. However, by 2010, in Russia there were only 80 thousand broadband users based on GPON, according to J’son & Partners Consulting. The main barrier to greater growth, as is almost always the case with a product entering the market, was the high price of active optical equipment. In the last few years, prices for station transceivers and subscriber optical modems have decreased significantly, due to which by the beginning of 2017 (according to analysts of the same company), the number of Russian GPON users will approach 6 million, that is, it will increase almost 75 times over the next seven years!

PON - what is it?

PON- this is, in fact, a technology for subscriber multiple access via a single fiber using time division multiplexing (TDM) and frequency division of reception/transmission paths (WDM). PON- from abbr. Passive Optical Network, which translates as passive optical network.

What is the operating principle of a PON network?

All network subscribers PON connected to provider equipment via 1 fiber. Transmission and reception occur at different wavelengths. To ensure that subscriber signals do not mix in the fiber, each individual subscriber device is always allocated a certain time slice during which it can transmit a signal.

What are the advantages of PON over FTTx?

PON technology has the following advantages:

  • Active equipment is used minimally;
  • Cable infrastructure is minimized;
  • Maintenance costs are considered low;
  • There is the possibility of integration with cable TV;
  • Excellent scalability;
  • Subscriber ports have a high density.

What information transfer rate is supported by PON technology?

Proposed GEPON technology actually works at a speed of 1.25 G, but at the same time, 0.25 G is redundant data used to encode the channel. It turns out that the real speed will be 1G.

What kind of equipment is needed to create a PON network?

OLT(from the abbr. Optical Line Terminal) - an L2 switch equipped with Uplink ports (to connect to the L3 switch), then - Downlink ports (to create a network PON). For example, OLT BDCOM P3310 has 2 optical, 2 copper and 2 “combo” 1G Uplink ports and finally, 4 optical 1G Downlink ports.

ONU(from the abbr. Optical Network Unit) is an excellent VLAN switch of a compact size. Standard ONU equipped with 1 optical 1G port (Uplink) and one 1G or 4 0.1G copper ports (Downlink). There are models ONU with 8, 16 and 24 ports and a model with a CATV receiver.

Splitter (Splitter) is a device that operates in a splitter mode in the “provider-client” direction and in a mixing mode in the opposite direction.

SFP OLT module- is a special transceiver for PON networks. An important difference from standard SFP modules is greater power and channel coding.

How is a PON network created?

PON network, as a rule, is a tree topology or a “bus” topology. Final ONU subscriber devices connect to the port OLT-and through splitters(to 1st port OLT-it is possible to connect no more than 64 ONU). It follows that to create a core network PON 1 is required for 64 subscribers OLT, then 1 module SFP OLT, 64 ONU and finally, a few splitters(the number of the latter depends on the type of topology).

What distance does the PON network support?

SFP OLT modules support operation over a distance of 120 km (point-to-point network type), but since traditionally the network PON has a tree structure (point-many-points), then the maximum operating distance PON, due to branching on fiber splitters, will be about 20 km.

How many subscribers can be connected to the PON network?

During network construction PON- it’s good practice to use one ONU- by one subscriber. In this case, the number of subscribers will be 256 per one OLT. If desired, to ONU it is possible to connect a Switch. Then the number of subscribers is limited only by the size of the MAC address table itself OLT, plus - ONU. Below are the sizes of MAC tables for OLT and individual ONU: OLT P3310-8192, ONU P1004B-1024, ONU P1501B-64, ONU P1504B-2048.

What is the difference between AC, 2-AC, DC and 2-DC OLTs?

The letters DC mean that to work OLT-a power source of 36-72V constant voltage is required. Similar OLT-s are necessary when a problem arises in organizing a 220 V electrical power supply. As an alternative, remote power supply through low-current communication lines is used.

The letters AC mean that OLT powered by a traditional 220 V electrical network. The mark in the form of the number “2” indicates the number of power sources: this OLT-and there is a backup power source, which turns on instantly after an emergency failure of the first one.

What types of splitters are there?

Sami splitters conditionally can be divided by the number of pins and manufacturing technology. As for the number of output streams, there are splitters: x2, x3, x4, x6, x8, x12, x16, x24, x32, x64, x128. Regarding manufacturing technology, splitters are divided into welded and planar. More splitters are divided by connector type: regular (SC/UPC) and special for CATV (SC/APC).

What is the difference between welded and planar splitters?

Splitters, which we called welded, are not equal-armed, namely: they do not divide the signal between the outputs evenly (for example, 5/95, 10/90 ... 45/55, 50/50). Splitters planar - will always be equal-armed and have more predictable attenuation at each output, because they have efficient manufacturing technology. In addition, planar splitters- broadband, and welded ones - have only 3 transparency windows (1310, 1490 and 1550 nm).

In what situations are welded splitters used?

In a situation where, for example, it is necessary to divide the signal in 2 directions, where, for example, the distance to one end point is 2 km, and to the other - 8 km. In this case, it is quite possible to use a 20/80 welded splitter. Splitters welded ones are also used to create a “bus” topology.

What would be better: welding splitters or maybe using SC connectors?

In such a situation, we get a double-edged sword. One side - welding gives an attenuation 10 times less (0.05 dB) than an SC connection (0.5 dB). The other side is that SC connectors will provide the ability to quickly search for faults in the network by connecting measuring instruments. You can find a compromise: weld Uplink using splitters, and connect Downlink using SC connectors. Here, let everyone decide for themselves.

Optical budget - what is it?

This phrase is understood as the difference between laser power per OLT-e and receiving sensitivity on ONU.

Is it possible to branch a PON network into 128 ONUs? (when the optical budget allows it).

No. Even if the signal power allows the network to be forked again, OLT still has a limit on the number of connected ONU on a physical level. Connect more than 64 ONU it's possible, but OLT will still register only 64 of them.

How to transmit over a PON -- CATV network?

On the side OLT-you need to install a CATV transmitter and then a CATV amplifier, which operate at a wavelength of 1550nm. On the side OLT-and it is necessary to use a CWDM flask at 1550nm. to inject the CATV signal into the fiber. All others splitters Must have SC/APC connectors. On the subscriber side it can be installed ONU with CATV receiver, or separate CATV receiver.

Is it possible to use SFP CWDM 1490nm. module instead of SFP PON?

Impossible. Even though CWDM itself is 1490nm, the module uses the same wavelength as SFP PON, these modules have different channel encoding algorithms.

Using PON technology, what Internet speed can be provided to subscribers?

If subscribers PON tree (64 subscribers) will simultaneously download a large amount of information from the Internet, then each subscriber will have a channel of 16 Mbit/s. And if you also take into account that not all subscribers use the Internet at the same time, and those that do do not consume the channel resource to the maximum, then a subscriber can even have up to 50 Mbit/s, sometimes even higher.

Why is it not acceptable for the ONU signal to be less than -26 dBm?

The whole point is, if on one/several ONU- the signal level will be very weak (< -26 дБм), то появляется большая вероятность возникновения ошибок в пакетной передаче с таких ONU. In the above case OLT wastes time to give ONU possibility to send the package again. These repeated requests reduce the efficiency of network throughput.

What should be taken into account when calculating the PON tree?

To build correctly PON tree, you need to take into account the optical loss arising from passive equipment. In theory, PON will cover an area with a radius of 20 km. Almost everything depends on the loss budget on a certain branch of the tree. For a correct calculation, it is better to be guided by the most unprofitable indicators of attenuation, power and radiation sensitivity of transmitters.

Are the copper ports in the ONU burning out?

PON the tree is created on optical fiber and, accordingly, is not affected by thunderstorms or interference. The problem only arises when one ONU- several users are connected via copper, and ONU placed on a pole. Such problems are solved by turning on a buffer switch, which is used in cases of interference. the thunderstorm takes the blow.

How can you determine the optical performance of a line?

To determine line attenuation, you can use special reflectometers for PON(they are significantly more expensive than regular ones), or optical testers. When the network is already built, the simplest solution for checking signal levels is to use special commands from the command interface OLT-A.

Can 2 ONUs communicate directly with each other?

Can not. Each transaction involving the exchange of information between ONU happens through OLT.

Will ONUs from other brands work with the BDCOM P3310 OLT?

No. ONU With OLT- this is something single and is a switching system. It is recommended to use equipment from one manufacturer (in some cases, cross-brand compatibility is of course possible).

OLT BDCOM P3310B supports - DHCP snooping (Option 82)?

Certainly. But for Option 82 to work effectively ONU Of course they should support this feature. Currently Option 82 only supports ONU P1504B model.

Is the PON network protected from flooding?

The TDM and TDMA technologies used guarantee network protection against flooding and broadcasting.

How can a PON network be disabled?

Apart from the obvious methods (cutting the cable), wood PON may stop functioning when constant radiation at 1310 nm appears in it. This, extremely rarely, occurs due to a breakdown ONU or due to the fault of attackers connecting a 1310nm media converter to the splitter.

Is it possible to assign a VLAN to each of the ONU ports individually?

Is it worth using splitters for CWDM, DWDM systems?

Certainly. Similar construction schemes are possible. Here you need to use planar splitters because they are wideband.

Publications on the topic