Optical Connections Magazine - Sept 2016_links

JOHN WILLIAMSON 400G SPEEDS

REVVING UP 400G

Sales of systems that support 200G and 400G transmission speeds are projected to grow at more

than 100% CAGR over the next five years. Here, we explain how and why.

JOHN WILLIAMSON

W hile the market for much lower speed optical communication systems continues to grow, industry and user interest in 400G technology is also quickly ramping up worldwide. NEED FOR SPEED The need for increased speed and capacity has a number of drivers. Historically we see the trac in the core network doubling every 18 months and this continues to be driven by: the knock-on eect of ever- faster broadband access; the proliferation of high definition video content in networks; the surging volumes of 4G (and soon 5G) mobile trac; the explosion in the uptake of cloud computing; the growing demand for massively high capacity Data Centre Interconnect (DCI); and the expectation that the future Internet of Things and machine-to-machine applications will put further major strains on network resources. According to new research from the Dell’Oro Group, shipments of systems supporting 200 and 400 Gbits/s speeds are projected to grow at a five-year compounded annual growth rate surpassing 100 percent. In practice, according to Uwe Fischer, CTO of SDN-enabled transport networking solutions company Coriant, 400G on the line side “is available already today.” He adds that on the client side,“ everyone is waiting for the standard to be finalised.” Dr Paul Brooks of Viavi’s optical transport

business unit, and one of the members of the IEEE working group developing the 400G standard, expects that work to be complete by the end of 2017, with shipments of systems gathering speed from 2018 onwards. Given that 100G solutions are just now hitting the big time, it might seem a little premature to be boosting 400G. Not so, says Brooks, recalling the lead times for other optical networking standards. For example, work on 100G itself began before 2004 and was only standardised in 2010. NOT A WALK IN THE PARK But talking about 400G is manifestly easier than actually deploying and running 400G. The technology presents a number of non-trivial physical challenges. “As you go up in bandwidth, density becomes an issue,” says Dr Adam Carter, CCO at optical

components, modules and subsystem specialist Oclaro. “And you have to be able to shrink the components and the various optical subassemblies and the ICs in the transceivers and modules. Usually “generation one” technology doesn’t give you a dense solution.” In parallel, heat may be an issue. “As you go to smaller packages the power envelop has to reduce, otherwise you can’t dissipate the heat,” points out Carter. “At the system level, dissipating heat adds time, cost and complexity.” Fischer notes that the move to ever higher bandwidths inevitably involves trade-os and compromises between

FORTY NEINERS

While the need for many slower speed optical communication solutions will continue, experts make something of an exception for 40G systems. Here the long term prospects for investment and new deployments look somewhat less than stellar. Fischer says that on the long haul line side, 100G has more or less replaced 40G, and in the metro network the transition is mainly from 10G directly to 100G. “In long-haul, 40G was an intermediate system, a kind of technological stepping stone,” he says. Brooks reckons 40G was a bit of a strange animal. “It had a very sexy price point because it leveraged cost-ecient 10G technology,” he comments. “It had a very successful window of opportunity, but that’s going now and the move is towards 100G. It’s going to be 100G because the price premium for 100G is closing over 40G, and the 25G technology used in current 100G clients aligns with I/O bandwidth of high performance servers utilising 25G.”

18

| ISSUE 7 | Q3 2016

www.opticalconnectionsnews.com

Made with FlippingBook - Online catalogs