Stacking the odds in favour of PON
The Full Service Access Network (FSAN) group, a forum that aims to drive the development of gigabit passive optical network (GPON) technology, has laid the ground work for the new generation of equipment, called NGPON2, which promises not only increased bandwidth, but also a much higher level of flexibility for service providers.
Even though the ITU hasn’t yet finished the set of technical standards required to implement this technology, it hasn’t stopped vendors trying to march on their competitors. Last October, Alcatel-Lucent said it intended to launch a commercial NG-PON2 product in the first quarter of 2015 as the access network speed war moves up another level.
And none too soon. Business users and mobile backhaul are pushing the limits of the bandwidth available under GPON, the most widely deployed standard for passive optical networks. Although it is still suitable for residential consumers, and will be for some time, the writing is clearly on the wall.
Not only can the NGPON2 standard operate at up to 10 Gb/s symmetrical bit rate per wavelength, it supports multiple wavelengths that can be ‘stacked’ onto the same physical plant to multiply capacity even further. Four wavelengths per system are described in the standard, to give a total system capacity of 40 Gb/s, and 8 or even 16 have been discussed.
With this kind of capacity, the optical distribution network becomes suitable for any combination of residential, business, and backhaul services. Its flexibility will allow service providers to monetise and control their networks more efficiently, because they can move subscribers as quickly as needed. In addition, the standard can be run on top of, or in conjunction with, older technologies, such as GPON, and XGPON1.
But technological factors are not the only consideration. Before service providers will invest in a new technology, they must have confidence that they will be able to deploy it in sufficient volume and that it will have sufficient longevity for the investment to be worthwhile.
The missing generation
Robert Conger, director of global carrier strategy and solutions for Adtran’s carrier networks division, said: ‘If you look across the different PON technologies, GPON is what has been mass-deployed today and it has been around for quite a while now.’ But there were some problems, he continued: ‘The first FSAN/ITU standard for 10-gigabit PON was XGPON1 which is defined as 2.5 gigabit downstream and 1.25 gigabit upstream – so four times the downstream and two times the upstream bandwidth [of GPON]. That standard has basically been in effect since at least 2010, but it was never really deployed in any significant numbers.’
The reason operators chose not to deploy XGPON1 in any significant numbers was mainly because the demand for 10 Gb/s bandwidth was not high enough at that time, according to Ana Pesovic, senior marketing manager for the Wireline Division in Alcatel Lucent: ‘I would not say that the opportunity window for XGPON1 is closing, I would say it has closed already,’ she said. ‘We have had XGPON1 since 2010. We have the technology; we did some trials; but we did not see any traction in the market at that time, because it was just too early and people really did not need the bandwidth at that time.’
Other technologies, such as EPON, have had some success in Asia, but this technology also never found much traction in the US or European markets.
Conger said: ‘10 Gigabit EPON did come to market and it has been deployed in small numbers over the last couple of years, primarily in markets in Asia. It would have been very comparable to XGPON1, and what would have been XGPON2, but it never came to market. The teams at FSAN that put together XGPON and GPON standards started working on NGPON2, and most operators are going to wait and jump straight to that.’
NGPON2 products are based on what’s called TWDM-PON technology, which stands for time- and wavelength-division multiplexing. Conger said: ‘When we talk about GPON and XGPON – all these technologies they are considered TDM [time division multiplexing] PON technologies. The combination of that with DWDM [dense wavelength division multiplexing] is where you get the name TWDM-PON.’
As NGPON2 uses different wavelengths from any previous technology generations it means that TWDM-PON can be deployed on any combination of GPON, XGPON, and NGPON2, all on the same fibre. Conger highlighted the benefits of this flexibility: ‘You could conceivably have eight TDM-PONs on a single fibre and also add in a bunch of overlay wavelengths. So there is a lot of flexibility there, and that was the intent. It borrows a lot of technology from the DWDM technology that is in place for the point-to-point technology today.’
The NGPON2 standard is still being worked on, but the MAC or physical layer that governs, among other things, the wavelengths that will be used is expected to be finalised very soon, explained Conger. ‘There is still some work to be done as far as coordinating wavelengths over multiple systems, and then there is the whole control-channel coordination between multiple systems so there are still some further specifications that need to be finished. I wouldn’t say that the standard is 100 per cent complete, but it is definitely complete enough that vendors can begin development.’
One of the main considerations in the new standard, and indeed any technology that needs to be mass deployed, is of course cost. This is one of the main advantages of NGPON2, along with the possibility of the symmetrical 10/10 Gb/s bit rate.
Pesovic said: ‘When FSAN was looking at these options, of course they were looking at the cost; because the point is not to make the best technology that nobody can afford, but to provide a good balance between the cost and the benefits. There were actually a lot of technologies that FSAN were looking at, some of them very exotic, but there were three main contenders: XGPON1 – which is also known as 10-Gigabit GPON – WDM-PON, and TWDM-PON.
Pesovic concluded: ‘This meant comparing the bandwidth available for each technology; the convergence of services; unbundling possibilities, which is very important especially in Europe; and also co-sharing and co-investment.’
One of the reasons costs for NGPON2 are manageable is that PON can be initially deployed using a single wavelength pair, which can then be increased as the needs of the business grow. Pesovic said: ‘With this one wavelength which you would start with you can define the same bandwidth that you have in XGPON1, which is 10 gigabit up, 2.5 gigabit down or 10/10 so you can do even better and then in the future you can just add an additional wavelength at a very small cost.’
This provides the performance of XGPON1 at a similar cost but with a much higher potential for growth in the future, as more wavelengths can be added. These can be defined separately so one could be used to serve business customers that require high bit rate symmetrical services, while a smaller bandwidth could be used to serve residential customers.
From Alcatel-Lucent’s perspective, the deployment of TWDM-PON is very straightforward, as Pesovic explained: ‘TWDM, for us, will be implemented in already existing access nodes, which is what makes our implementation special because, as far as we know, we are the only company that can do that on the existing fibre platform.’
Recently Huawei announced the successful trial of 100G passive optical network (100G-PON) through its optical access innovation lab. The new 100G-PON technology also uses a hybrid time-division and wavelength-division architecture but at a higher bit rate on each downstream wavelength. Thus the system supports a 4x 25 Gb/s downstream rate and fx 10 Gb/s upstream rate. Hauwei says its 100G-PON technology is compatible with GPON, 10GPON and 40G TWDM-PON, allowing carriers to upgrade their existing networks.
Although only in the prototype stage, this technology could be a route for a future generation of PON technology as consumers’ demands for data continue to grow. However, some people have reservations that this technology is in a similar position to XGPON1, in that the demand for bandwidth is just not there at this time. It remains to be seen whether data rates grow fast enough for this technology to become viable, or if it will be swallowed by the next big thing a few years down the line.
Robert Conger explained: ‘One of the biggest challenges with all of these technologies is to get the cost reduced to the point where it can be mass market adopted and not be a very niche technology. To do that, you need volume; and so coming up with different really high-rate solutions that may show well in a lab but are unlikely to be seen in the field, doesn’t really help the market.’
Conger continued: ‘NGPON2 is a technology where most of the vendors are saying they are going to put their volume behind it. Now that may be largely business and backhaul services in the initial days, with residential to follow at a later point, that is really what most operators are planning.’
These comments were echoed by Pesovic from Alcatel-Lucent: ‘This technology is not standardised, this is more of a lab trial or a prototype. We are looking to deploy a standardised technology that can be deployed as soon as possible to answer the operator’s needs, but we also have Bell Labs, which is our research organisation, and they are constantly busy with new technologies. ’
Conger explained that NGPON2 looks as if it will be able to cover all the use cases that operators currently see, at least within the PON market. ‘There is definitely a place for the 100-gigabit technology and 40 gigabit and other things, but they can be deployed over point-to-point or other scenarios. They don’t need a point-to-multipoint system to deliver those services. Today it’s more about getting the volume to reduce costs.’
Conger added: ‘We have still got a long way to go with NGPON2. It is the cost that is ultimately going to drive adoption versus, say, point-to-point. If you build an ONT [optical network terminal] and it is $5000, then you are just not going to get many people to adopt it, no matter what the rate is. They can always go for a point-to-point technology if they want to do that, and do it more effectively.
‘Right now, we need a 10-gig PON technology in a really bad way and a lot of operators are saying “well, do I make a jump to [10G] EPON, at least that exists today?” but I think now we have addressed those concerns. NGPON2 is close enough; we are at the point now that we can demonstrate some of the technology; and I think that there is more assurance that it is not going to go the way of XGPON1. So that is what we need in the market now: a successful NGPON2.’
Conger highlighted that GPON has seen around a 10-year lifespan, and he expects that NGPON2 will last a similar amount of time before it is surpassed by new technology. However, no one would be surprised if the demand for data rates is driven even higher than they currently are. Smart phones for instance were developed in the late 1990s, but it was only in 2007 when the iPhone was released that there was a proliferation of smart devices that consumed large amounts of data.
With operators around the world launching gigabit services (see feature p14), it will not be long before the European and other markets follow suit. In the opinion of Kurt Raaflaub, global product marketing, carrier networks division, Adtran, these gigabit services will drive the development of new applications that inevitably will consume more bandwidth than those that came before.
Raaflaub said: ‘We have to recognise with the proliferation of gigabit services mainly in the US right now, that is going to fuel application innovation. There are a lot of people with great ideas on how you can cram an application into a gigabit pipe. That is going to lead to applications that you or I have not even heard of yet, but 18 months from now someone will come up with something.’