On the surface, the 100G discussion seems like gasoline for the optical bubble theorists' fires. But Heavy Reading sees reason to be optimistic about the prospects for a big leap forward in optical transmission rates: Our surveys of network operators around the world show that they are clamoring for more speed, and specifically for 100G.
In a Heavy Reading operator survey conducted in late 2006 – the results of which were published in Long-Haul DWDM: Market & Technology Outlook (Vol. 5, No. 2, February 2007) – 45 percent of the network operators we canvassed reported that 100G technology was currently under investigation within their organization, and only 11 percent of the survey group expressed no interest in 100G.
In a follow-up survey conducted this summer, 37 percent of 107 network operator respondents said their company planned a wide-scale rollout of 100G technology in the 2010-2012 time frame, with an additional 16 percent planning to roll out 100G even earlier. Results of this survey are revealed in Heavy Reading's latest report, The Future of Optical Transport Networks: 40G & the Road to 100G.
The good news in these numbers is that the need for 100G is not a matter of vendor push, but of customer pull. This point alone makes optical networking in 2007 drastically different from optical networking circa 1999. In fact – due in large part to a telecom recession that stifled R&D – optical equipment makers are behind the curve in rolling out new transmission speeds, both 40G for today and 100G for tomorrow.
The picture for 100G is not all rosy, of course. Given the dearth of attention, investment, and research in the technology over the past few years, the industry has its fair share of challenges to surmount to get to 100G. Among them:
- Components and electronics technologies are not ready for prime time. We are seeing some early 100G technology trials from vendors including Alcatel-Lucent (NYSE: ALU), NEC Corp. (Tokyo: 6701), and Nokia Networks . These trials are important steps, but the vendors – and service provider participants in them, such as AT&T Inc. (NYSE: T) – readily admit that the trial systems are too expensive for commercial deployments. Volume is part of the issue here, but entirely new techniques may be required to bring 100G to market commercially, regardless of what works in the lab.
- 100G standards are in their infancy. This is a big one. The key standards bodies for 100G are the International Telecommunication Union (ITU) and the Institute of Electrical and Electronics Engineers Inc. (IEEE) . More specifically, the standards groups are the ITU-T Study Group 15, which is responsible for optical and other transport network infrastructures, and the IEEE 802.3 Higher Speed Study Group (HSSG), which defines new Ethernet interfaces beyond 10G. The IEEE and the ITU work together closely. The IEEE is defining the physical interface for 100G Ethernet (as well as for the newly added 40G Ethernet). The ITU is defining the standards for wide-area transport of 100G Ethernet over WDM. If standards efforts remain on track, the industry is looking at standardized 100G Ethernet in the spring of 2010 – at the earliest. Standardization is critical for Tier 1 operators around the world and for the components and chip makers that need stable specs from which to build.
- 40G is a monkey wrench in the works. With 10G networks starting to creak and 100G still in the labs, 40G is emerging as a here-and-now solution for operator capacity exhaust. But we believe that 40G will be just an interim solution on the road to 100G. To some extent, vendor 40G development will divert R&D investment from 100G, slowing the migration. In addition, our operator surveys indicate that, with both 40G and 100G on the table, service providers are confused about the best migration path. 40G transmission is a bit like what ISDN was to broadband: It's here, but it's too little, too late.
— Sterling Perrin, Senior Analyst, Heavy Reading