The Roman emperor Julius Caesar said, “All Gaul is divided into three parts.” Were he surveying the global broadband industry today, he might say, “All residences are divided into three parts — fast broadband, slow broadband and no broadband.”
Chipmaker Lantiq uses numbers from the ITU and Akamai to show that 10% of the world’s residences have broadband speeds of 10 Mbps or higher, 30% have speeds of 1 to 10 Mbps and 60% have zilch — no broadband whatsoever. That’s no way to sell a movie or TV show when 60% of the world’s population can’t watch it over an Internet connection and another 30% can’t watch it in glorious 4K.
Lantiq says it has a solution that’s not nearly as expensive to deploy as FTTH or G.Fast and it is available now. It’s an FTTdp technology, like G.Fast, called Vinax dp. Telcos install a distribution point within 200 meters of the home, connect it to a fiber from the central office and to the existing copper phone wire to the residence and voila! 200 Mbps down and 100 Mbps up. It uses an ITU standard called G.993.2 Annex P, which has been available since 2011 so there’s no developing a new standard and subsequent certifications and testing. Effectively it is a form of short reach VDSL2 with, reduced power and an enhanced data rate, but most chip vendors have not bothered with it.
There are two markets that Vinax dp is aimed at, one being the “broadband have nots.” Lantiq says: “There are over 4 billion people who have no broadband at all. It is incredibly important to get these 4 billion connected. Vinax dp is an incredible attractive technology for that scenario. It is cheap, easy to install and very effective.” In this scenario, Lantiq’s multi-port solution will be used – 8 ports, each delivering 150 Mbps.
The other market is with Lantiq’s single-port solution. One box connects to one subscriber and delivers aggregated rates of 300 Mbps (200 down, 100 up). It’s easy-to-deploy yet is a premium service for customers. There are millions of consumers in developed countries that would like to have 200 Mbps available from someone in addition to the cableco.
Lantiq uses as one example India where there are 250 million households (compared to about 120 million or so each in the US and Europe) but only 13 million wireline broadband connections. There are 850 million mobile phone users in India but as those phones become more sophisticated, they cannot even download upgrades from their slow mobile networks or watch high quality video streams. Lantiq gave as an example of the slow cellular network people having to catch a bus to go to a distant location that has a WiFi hotspot to download a 700MB upgrade.
The giant Indian conglomerate ($67 billion in annual revenue) Reliance Industries has scoped out the situation and wants to create a “Digital India” where in 3 to 5 years about 120 million residences would have 100 Mbps down and 50 Mbps up. It budgeted $14 billion to build the national network. It has already spent between $8 billion and $10 billion building 800,000 kilometers (480,000 miles) of inter- and intra-city fiber network, set up an intercontinental undersea fiber network. Reliance has over 2,000 retail outlets. The chip vendor that wins that deal, is going to make profit on any chip design overnight.
Lantiq said it has talked with 23 telcos that have plans for FTTdp deployments as opposed to FTTH deployments. The 23 have to choose between the new and as yet not fully tested G.Fast broadband technology and an FTTdp technology based on the proven VDSL.
Vinax dp has several main attractions. The chips are smaller than other FTTdp chips, so Vinax dp boxes can be smaller and easier to install. They come in a weatherproof housing and operate in a wide temperature range (industrial temperature range of -40 to +85°C) so they can be installed anywhere — on utility poles or in basements.
The chips use very little electricity and so can be powered by a modem/gateway in the residence. The CPE end is customer-self-installed, either as a separate adapter to be used with an existing VDSL modem/gateway or embedded in a new one. Vinax dp uses industry standard and proven GPON in and VDSL2 technology out.
Perhaps most importantly, Vinax dp boxes are much less expensive than alternatives. It’s difficult to accurately estimate equipment costs for every circumstance, but Lantiq said Vinax dp costs are about one-tenth the cost of G.Fast, which admittedly has higher speeds and of course is trying to recoup the R&D costs of its development.
Not only are Vinax dp chips inexpensive, the accompanying VDSL products have become “dirt cheap” after years of being in production. The network architecture that Vinax dp uses makes it easy to upgrade later on to G.fast and its fiber-like speeds. Most modems that have been shipped in the last year or so are compatible with Vinax dp and support the 200/100 Mbps speeds. Older modems are compatible with Vinax dp but would provide slower speeds when connected to a Vinax dp distribution point.
Newer modems and gateways have profile 30 technology, which allows 200 Mbps down and 100 Mbps up. Older modems and gateways have profile 17 technology which provides a maximum of 100 Mbps down and 50 Mbps up. They use respectively 30 MHz and 17 MHz of spectrum on the wire. It seems like Vinax dp makes sense because of its low-costs and use of readily available and inexpensive VDSL components.
The questions that telcos in developed nations will have to ask are:
Assuming they have not decided to deploy FTTH, do they provide 200/100 Mbps now or do they dare wait for and spend the additional money for G.Fast?
Do they risk losing their top-paying subscribers to local cable companies? If they increase the speeds they offer, they may be able to charge only a small amount more, if any, but if they lose that subscriber, they lose $25 to $50 a month of very profitable revenue. The answer will be different in different countries.
As bandwidth hungry as UHD and cloud-based applications are, it’s hard to make the case for needing more than 200 Mbps for the foreseeable future. Netflix has increased its recommendations for UHD to 25 Mbps per stream. Not many homes are going to have multiple UHD TVs in the next few years although additional bandwidth is needed for all the other fixed and mobile devices in the home.
Netflix recommends 3 Mbps for SD quality video, 5 Mbps for HD quality and 25 Mbps for Ultra HD per stream, so if a home has a UHD TV, a couple of HD TVs and there are several tablets or smartphones in use, the recommended bandwidth could be upwards of 50 Mbps or more.
On the assumption that most every first world residence will soon have subscriptions to one or more OTT service and have one or more UHD sets plus several HD sets and mobile devices, the mass market for broadband is speeds of 10 Mbps to 100 Mbps and already averages in Europe are around 17 Mbps, a bit higher in the US.
Telcos have more choices today than a few years ago. Until vectoring came along last year, telcos had only two choices: all-fiber or stay still. Verizon went all-fiber in about 70% of its footprint as did some smaller telcos and local utilities. The other telcos have watched as cablecos have taken their profitable broadband subscribers.
The French regulatory agency, which has been pushing all-fiber network architecture, has recently acknowledged that 10% to 20% of residences cannot be connected directly to fiber. It now allows telcos to use FTTdp in areas where it is hard to deploy FTTH such as where there are ancient buildings and streets. Maybe other countries and all telcos should do the same. Several makers of telco gear have selected Lantiq’s Vinax dp chips for their future requirements. They don’t usually do that unless they have interested customers. Perhaps most importantly for telcos that compete against cablecos, Lantiq says, “It’s good enough to compete with DOCSIS 3.0.”