The launch of 5G by South Korean operators serves as a first benchmark for other operators around the world
Webscale and China appear strong at this year's OFC
Twenty years has passed since my first time attending the optical industry’s big event, OFC, so the latest one was bound to have some contrasts.
A look back: changes in the vendor landscape
Back in 1999, my focus was Asia, where WDM had appeared in Japan but was just beginning to emerge in most other countries. Most networks were full of PDH and other asynchronous protocols, and would remain so for many years – despite the hype at the time. The lag was due in part to strong demand for E1 circuits in the late 2000s, as mobile broadband traffic rose in backhaul networks (but before carrier Ethernet was well established).
Network protocols were more complex in 1999, and varied more among countries. Procurement also tended to differentiate local vs. foreign. Not surprisingly local vendors often prevailed, and market share patterns varied by country. Samsung was a key transport vendor to KT, for instance. These patterns have changed quite a bit. Operators have tried to move towards global protocols and simple procurement processes in order to benefit from a global supply chain. That’s the case especially in Japan and Korea. In China, local suppliers still have some inherent benefits that foreign vendors never will have, so that market remains ruled by the Huawei-ZTE-Fiberhome trio. (Nokia and Ericsson have big RAN wins in China, but their optical presence is more limited.)
Consolidation has swallowed up many of the key US and European optical vendors from 1999: Alcatel, Lucent, and Siemens (networks) (all now part of Nokia); Nortel (Ciena); and Marconi (Ericsson). Japanese suppliers Fujitsu and NEC are still active and important at home, but less so elsewhere. China-based Huawei and ZTE were strong in 1999 mainly due to home market revenues, but grew soon after into competitive global vendors. US-based Ciena has grown much bigger. Other suppliers with pockets of strength in 2019 include Infinera/Coriant, ECI, Adva, Cisco, and Fiberhome. But Huawei and Ciena stand out.
Huawei and Ciena as early pioneers
While Huawei and Ciena stand out among optical’s key suppliers now, in 1999 they were considered relative upstarts. Ciena was the new(ish) US-based startup pioneering a cool new technology (WDM), but many Asian operators weren’t sure it was worth the investment, and if it was manageable over the longer run. Ciena’s technology chops were not in doubt, though, as it had landed big early contracts with demanding Tier 1 carriers in Japan and the US. Huawei, by contrast, faced a much different battle. There had never been a competitive network equipment vendor from China. Getting carriers to jump right into the optical deep end was not an easy task.
Needless to say, Huawei is in a different position today. Public statements from Huawei suggest it has been #1 globally for many years, in multiple optical product areas. MTN Consulting does not estimate optical share, but does compile quarterly share on the overall network equipment (and software/services) market. In 2018, we estimate Huawei’s share of telco spending on network infrastructure at 22.4%, about the same as 2017. That is 22.4% of all telco spending on network infrastructure: tech hardware, software and services, globally. (Total in 2018: $198.4 billion). Number 2 Nokia had just over half Huawei’s share in 2018, at 11.9%.
OFC 2019 reveals a new world optical order
OFC’s 2019 event in San Diego clearly showed a new world order in optics – even relative to five years ago, my last OFC. There was less regional presence from Japan (including in optical components), more from webscale network operators like Alphabet, Facebook, and Microsoft, and more from a range of Chinese companies. Mostly from Huawei, as the globe’s optical systems giant, but also from a number of components suppliers based in China, like webscale operator Alibaba.
OFC had less telco presence than hoped. AT&T made a strong showing as did a few other operators (including Verizon and DT), but their R&D budgets are shrinking. That is not a promising trend for those that hope innovation can continue to come from the telco side, not just the vendors.
On the plus side, webscale operators are doing some amazing things with their networks to cope with growth, and sharing their experiences fairly openly at events like OFC. The most interesting presentation was from a Google transport engineer, Eric Breverman. The focus was the company’s “Optical Zero Touch Networking”. The goal of ZTN, Breverman explained, is essentially to “keep people from actually touching the network”. Humans make mistakes, and they are too costly to keep hiring at the same pace as traffic. Google says it now supports intent-driven networking on 50% of “Google’s Production Optical Network” (with the unfortunate acronym of GPON). Just a handful of engineers running an entire global network becomes a reality with software.
2019’s vendor landscape
Over the last five years, large portions of the optical systems market have become dominated by a single vendor, Huawei. Many large European and Asian operators have deployed Huawei gear across their networks, in the RAN, core and fiber access plant; some have long-term services contracts; some have made long-term technology commitments to Huawei (such as joint R&D). As telcos consider optimal solutions in the transport/optical domain, some may find themselves over-leveraged to a single vendor.
It’s a welcome sign, then, that more of the world’s optical innovation is coming from the software and webscale sectors, and more of it is open source. And that disaggregated networks are becoming more of a reality. This was a hot topic at OFC 2019.
Competition and disaggregation
In the optical networks space, disaggregation generally refers to open line systems, where systems and transponders are decoupled. That allows for faster upgrades of transponders, and helps to avoid vendor lock-in. Many Tier 1 operators are now pushing this “open optical network” approach, along with broader efforts to promote a white box ecosystem where hardware can be decoupled (or disaggregated) from software.
This ecosystem is getting increasingly mature as 5G arrives. AT&T in fact has already committed to deploy white box routers in their mobile RAN, at up to 60,000 towers over the next few years. But fiber networks are harder, according to Scott Mountford from AT&T, both due to the long-life of installed cabling, the variety of attached equipment, and the fact that “operational systems need development” in this area. Operators will benefit from disaggregated networks as they become reality, but in the meantime face a highly concentrated vendor landscape.
Matt Walker is chief analyst at MTN Consulting
Image credit: iStock Photo