EoW November 2008

english technology news

The drive for future proof cabling standards

Martin Rossbach, director of product management and new market development at Nexans Cabling Solutions, looks at a current industry dilemma. Cabling and Ethernet standards have rarely been controversial. When the industry in general has had its fair share of standards wars, there have been fewer debates around the progression from Cat3 to Cat5 and Cat 5e, or from Ethernet to Fast Ethernet to Gigabit Ethernet. Now the industry is being tested because it is caught between two trends: the desire for more – 40G and 100G Ethernet – and less, as developments in green computing drive innovation in Energy Efficient Ethernet for the first time. What this means is that whether in the data centre or the network, purchasers of infrastructure are making important, far-reaching decisions that they may regret if they get them wrong. Everywhere we look, demand is mushrooming. An example: global subscriber access traffic shows growth that, if we assume 50 million IPTV subscribers in 2011, IPTV will consume more bandwidth than internet and phone traffic combined. Put all three together and by 2011 that’s traffic of almost 300 million Terabytes every year, five times the volume in 2007. At the server level, demand will continue to grow. Intel and Broadcom show figures of x86 server units at fewer than 10 million in 2008, almost all of which use Gigabit Ethernet. By 2015 those numbers will have doubled, and the overwhelming networking technology is 10G Ethernet, with the rest 100G. That timescale is well within the lifespan of today’s cabling, and that means that anyone installing cabling today has to be mindful of the way the data centre will evolve. The demands of 10G Ethernet dictate Cat6A cabling as a practical minimum standard, but in the world of 10G Ethernet, that cabling will need to be put in before we settle on the likely dominant technology. While there’s an overwhelming desire to see the 10G Base-T standard ratified, it’s not a success. The complexity of the project has created what is probably the most power-consuming chipset ever created. In an era when the wasteful power consumption of IT is being questioned – IEEE calculates that the total energy consumption of network equipment is 13 terawatt hours per year – 10GBase-T is not a successful project at the moment. Out of these technological challenges comes innovation, and there’s much better progress for the working group on Energy Efficient Ethernet (EEE). What would have been seen as a pointless exercise a few years ago is now an exciting area of innovation: by stepping down the power consumption of the chip set during its idle time – most of the time it is in operation – it can reduce the waste of energy. In simple terms, if the chipset is idle 90% of the time, the ability to step down power consumption to 10% or even 0% will save 90% of the energy supplied. EEE has the aim of defining a mechanism to reduce consumption for 100Base-TX, 1000Base-T, 10GBase-T, 10GBase-KR and 10GBase-KX4 among others, and has widespread industry support.

As we solve one problem another comes racing towards us. Away from the data centre and copper cabling, we constantly have to upgrade our own expectations. As service providers begin to create services to supply IPTV and video on demand to the home, it stretches the boundaries of what can be supplied over IP networks. For the home user, 30 minutes of TV uses the same network capacity as 30 days of Internet surfing. When HDTV and VoD are the main business driver for service providers, the largest cable TV supplier in the US, Comcast, predicts that per user traffic demand, currently 3.5 million, will explode to 19 million in the HDTV era. The demand for 40G Ethernet is urgent, and yet the predictions for the uptake of service show that 40G Ethernet may not be enough. To satisfy likely user demand in the near future, service providers will need to provide 100G Ethernet on their networks. The lesson from this is that whatever the network, we are at the beginning of an explosion in demand that will tax our infrastructure, and our ability to ratify standards and commercialise them, to the limit. For anyone responsible for specifying and installing infrastructure today, short-termism isn’t the answer. Nexans has secured an €8.9 million contract from the Istanbul Transportation Authority (IETT) to provide 988km of specialised rail cables for an extension of the city’s Metro and Light Rail systems. Nexans will supply to Gulermak-Dogus JV, the overall project contractor, 72km of low voltage and 262km of medium voltage power cables which will be used for power distribution and DC systems for a 5.3km Light Rail Transport (LRT) line and a 15.6km metro line expected to carry 67,000 passengers an hour. Nexans, involved in an earlier expansion of the Istanbul Metro, will also provide lighting cables for use in 20 new stations. All these cables will have a Halogen Free Flame Retardant (HFFR), sheath, meaning that they are preventing the propagation of fire, while providing low toxicity, low corrosivity and low smoke density, thus reinforcing the safety of people and equipment on board. The total installation is scheduled for completion by the end of 2009. All the cables will be manufactured by Nexans plants in Germany. Commuter rail cable contract for Nexans Nexans Cabling Solutions – Belgium Email : martin.rossbach@nexans.com Website : www.nexans.be

Nexans – France Fax : +33 15669 8484 Email : nexans.web@nexans.com Website : www.nexans.com

46

EuroWire – November 2008