Various announcements and recent activities by the HomeGrid Forum over the unified home wiring G.hn standard look like a last desperate throw of the dice as far as North America is concerned. But they really highlight a change in focus to Asia where the technology has decent prospects of success, along with an admission that MoCA has won the home wiring battle hands down in North America. We think we all, really, already knew that.
Europe and other regional markets such as Latin America lie somewhere in between, with some hopes for G.hn in MDUs (Multiple Dwelling Units), where integration of coaxial, power wiring and telephony wiring in a single chip is more likely to appeal.
More to the point G.hn has two tricks up its sleeve beyond PHY integration, which are mitigation of interference between neighboring powerline systems and support for the emerging plastic fiber standard as another wiring option. These both hold greater interest for operators deploying inside MDUs, of which there are many in the key Asian markets of China and the Far East especially.
There are many MDUs elsewhere, but in North America these are already well wired with coax, while in Europe there has been growing momentum behind both MoCA and Home Plug, the leading standards for coax and power cables respectively, outside G.hn itself. To a large extent the ‘no new wires principle’ holds, making MoCA very popular in the US where there is a lot of coax already in buildings, while in Europe Telcos such as BT, France Telecom and Deutsche Telekom are going for Home Plug since they have accepted that telephony wiring is usually inadequate. Meanwhile there are signs that MSOs and to an extent DTH operators in Europe are leaning towards MoCA, if with less enthusiasm than their US counterparts, because they at least have coax in their subscriber homes running from the access point to the set top box. Liberty Global for example specified MoCA as its fixed wiring option for its Horizon hybrid platform, while emphasizing that Wi-Fi was its preferred home networking medium.
The G.hn standard itself has virtually been complete for three years but it has taken a long time getting chip makers on board, firstly with Marvell and most recently with Sigma Designs gaining certification from the Home Grid Forum for its CG5200 G.hn chip. The Forum now rather misleadingly claims there are G.hn chips from multiple vendors, which is stretching the spirit of definition a bit when there are just two. There were at least supposed to be three, with Lantiq having originally backed G.hn alongside Marvell and Sigma, but it has been stalling and concentrating its efforts on other areas, such as VDSL2 with vectoring. The Home Grid Forum is hoping that the promise of pickings in Asia will revive Lantiq’s dormant interest.
The other main recent development has been the final subsumation of the HomePNA Alliance into the HomeGrid Forum, which at least reduces the fragmentation among home wiring standards. HomePNA was originally set up to define standards for data communication over home telephone wiring and did gain some traction among IPTV operators, eventually embracing coax as well in competition with MoCA. But with declining interest in transmitting video over telephony wiring because of poor quality and inability to carry HD, coupled with failure to make any headway against MoCA over coax, HomePNA had already come to focus increasingly on G.hn and it made sense to merge quietly with the larger HomeGrid Forum.
The problem with G.hn has been that while it promises cost savings for vendors of routers, set tops, gateways and other devices that need to support multiple PHYs by having a single chip to support each, in practice equipment is segmented into different sectors, given for example the huge market for MoCA on its own in North America. In that case G.hn if anything imposes an unnecessary cost rather than a saving. It was panned mercilessly by leading players in the MoCA camp, notably by silicon vendor Entropic and its technology chief Anton Monk, who also highlighted G.hn’s lack of compatibility with most of the systems already out there.
The likes of Entropic have been highly successful in keeping G.hn out of North America and to an extent Europe, although of course aided by the fact few existing deployments have combined more than one wiring system. This firstly confined G.hn to either new buildings or existing sites that as yet have no suitable wiring candidates, but then to morph from being the universal PHY to one that is best suited to just one. This is where G.hn’s two new tricks alluded to earlier come into play.
The first one is focused on power line transmission, where the Home Grid Forum has observed the struggles the Home Plug AV alliance was having sustaining high bit rates consistently with good QoS over the wide range of home power distribution networks, which vary in terms of quality and also phase. However internal arguments leading to delays in developing G.hn chipsets gave the Home Plug Alliance time to catch up and itself address many of these issues. Home Plug came up with AV2 incorporating some of the same technology that G.hn had originally touted as competitive advantages for data over power cables, including MIMO (Multiple Input Multiple Out[put) for multiple transmit and receive antennae combined with beam forming to improve signal propagation.
G.hn has argued this would enable whole home coverage with consistently high bit rates at every outlet, but now Home Plug chip vendors Qualcomm and Broadcom are delivering these benefits actually before G.hn. Broadcom has just announced its new family of Home Plug AV2 devices designed for residential gateways, set tops, PLC-Wi-Fi Extenders and powerline adapters, claiming they have up to five times higher throughput and up to twice the coverage of currently available Home Plug AV packages. Qualcomm for its part pre-announced such capabilities at the Las Vegas based Consumer Electronics Show in January 2013.
But the Home Grid Forum believes that these performance advantages will be confined largely to single dwellings and that there is an opportunity to make power line operation more robust and reliable within MDUs, where interference between adjacent home networks is an additional challenge. This has already given rise to problems in some instances where there is significant powerline usage within a MDU, with different tenants’ systems in close proximity and often connected to the same wiring. Then signals can be detectable on adjacent networks, causing interference and performance degradation, and in turn screen freezing or failure as well as interruptions to VoIP calls.
The HomeGrid Forum has addressed this phenomenon known as Neighboring Networks (NN) interference by developing its Neighboring Domain Interference Mitigation (NDIM) technology, aiming to automatically mitigate interference between adjacent powerline networks that use G.hn. This includes an impressive range of measures that together do seem to solve most of the problems specifically related to NN.
The HomeGrid Forum is naturally fond of pointing out the perceived shortcomings of NN, mitigation techniques employed within the latest Home Plug AV2. One of these is line filtering, in which powerline filters are manually located on the wire between apartments in a MDU to block the data signals even though the mains power flows unimpeded. The argument is that while this can in theory be quite effective, it relies on expert placement of filters which depends on the region and type of device, being expensive to deploy. The forum also quite correctly points out that line filtering does nothing to prevent interference via electromagnetic induction between different wires running close to each other, which they often do in dense MDUs.
Under Home Plug AV2, power reduction is employed to address this inter-network interference via electromagnetic induction, by reducing the output of each node. However this has the side effect of reducing signal to noise ratio along the wires, which increases the potential for interference from extraneous electrical noise from nearby appliances. Again it requires careful management with manual intervention.
However Home Plug’s power reduction served as a starting point for G.hn’s more advanced adaptive power control within its NDIM, designed to cater for varying levels of low interference by adjusting the power output independently across each node and ensuring that signal to noise ratios are maintained. This is then enhanced with a technique called orthogonal preambles, which means deliberately ignoring relatively low levels of interference from neighboring networks that will not actually cause a problem ‘ in other words where mitigation would be more trouble than it is worth and degrade rather than improve performance.
It is important however to align neighboring powerline networks for Media Access Control (MAC) so that all nodes are coordinated in their cycle time determining when they access the network to transmit and receive signals. This makes it possible quickly to identify when a newly added node is causing interference so that the whole system can respond to protect all the other nodes and ensure that services continue uninterrupted without a blip.
There are also some other tools within the ITU-T G.hn standards suite as a whole, which has been designed to be upgradable so that algorithms can be evolved and expanded through software changes. By enabling neighboring networks to be effectively insulated from each other in terms of data signals, even though they are part of the same mains delivery infrastructure, these tools also help allay concerns over privacy and security through signal leakage.
G.hn’s other relatively new differentiator, almost exclusive of the powerline angle, is support for the emerging Plastic Optical Fiber (POF) designed primarily for in building use as a natural extension to FTTH. The argument is that just as fiber to the curb creates a desire to go all the way to the home to boost performance further, so FTTH in turn will generate demand for FITB (Fiber In the Building). This will be more likely the case for MDUs, given that there is talk of dispensing altogether with fixed wired options in smaller dwellings in favor of total Wi-Fi. So this plays into the Home Grid Forum’s MDU strategy for G.hn, but targeting a complementary market among new builds primarily. The case is then quite compelling, since gigabit speeds can be guaranteed without interference and at costs comparable or not much higher than coax, when starting from a clean slate.
POF was first mooted over two decades ago as an alternative to coax and copper for enterprise LANs (Local Area Network) but failed then partly through lack of interest from manufacturers, which meant connector costs did not come down. This time it is precisely the much lower cost of connectors that promises to make POF competitive for Greenfield MDU developments, with the cable itself being little cheaper than its glass counterpart.
Another factor enabling POF to come of age has been the development of higher performance so called per-fluorinated polymers, which for the first time allowed a comparable proportion of the fiber core, about 96%, to be used for light transmission as in glass fiber. However plastic still suffers from much higher light attenuation than glass, which means it will still be confined to in-building applications at distances under 100 meters. By contrast glass fiber comes in to its own over longer distances for telecommunications when the higher connector cost is less of a factor.
Greatest interest in POF has so far come from Europe, where there are several research projects, such as POF-ALL and POF-PLUS, as well as standards efforts. The first major G.hn over POF trial was also held in Europe late 2012 in the Netherlands, Germany and Spain, involving Marvell as G.hn silicon provider, Telefonica, FTTH equipment maker Genexis, ISP in Germany Teleconnect, fiber chip coupling specialist DieMount and Eindhoven Institute of Technology.
The trial successfully demonstrated delivery of multiple services over FITB, effectively FTTH as far as inhabitants of MDU apartments are concerned, and has helped breathe new life into G.hn. But it is far soon to declare that it is right back on the map at a time when it is clear that the biggest silicon vendors are backing Home Plug and MoCA as a pair and integrating them into single PHY chips, in Broadcom’s case adding DOCSIS as well for cable operators. At the same time failure to embrace Wi-Fi within the G.hn strategy also needs to be addressed, given that is also covered by the P1905 standard, which bridges it with Home Plug and MoCA as well as Ethernet.
The fate of G.hn lies mainly though with its ability to persuade designers of new buildings, primarily in Asia, that it has a significant enough advantage in powerline performance and reliability, or alternatively to back POF.