[wp_tech_share]

We just published our latest Carrier Economics Report – covering the second half of 2017 and detailing capital expenditure (capex), revenue, capital intensity, and subscriber trends by region for the top wireless and wireline telecom carriers. With publication of the report, I wanted to share some key takeaways:

  • After two consecutive years of declining capex trends, we believe there are reasons to be optimistic about the future. We have adjusted our overall three-year capex expectations upward to reflect a more optimistic investment view than we had originally envisioned in both the US and Chinese markets (Figure 1). Worldwide capex growth is now expected to increase at a compound annual growth rate (CAGR) of one percent in constant currency terms between 2017 and 2020.
  • After three consecutive years of declining capex and improving capital intensity trends in the US telecom market, we maintain the view that conditions are stabilizing and that both capex and capital intensity will continue to trend upward. Aggregate US telecom investments are expected to grow at a high-single digit rate in 2018 and advance at a CAGR of two percent between 2017 and 2020. Multiple factors underpin the renewed optimism for capex in the US, including: (1) Sprint is once again investing, (2) larger data plans are propelling capacity investments, (3) FirstNet investments are set to commence in 2018, (4) corporate tax cuts are boosting AT&T’s 2018 capex, and (5) revenue trends are stabilizing.
  • Constrained operator revenue growth is expected to be one of the primary inhibitors of further capex acceleration with 5G. We anticipate that currency adjusted carrier revenues will remain flat between 2017 and 2020. The forecast assumes that operators will struggle to identify new revenue streams, so as to offset slower smartphone revenue growth.
  • We remain optimistic about the long-term possibilities with IoT. At the same time, we believe the upside will remain limited over the next couple of years. Our baseline estimates assume that carrier IoT revenues will grow ~1.75.x between 2017 and 2020, accounting for about two percent of total mobile operator revenues by 2020. Preliminary IoT connection pricing trends for 2017 are cause for concern, with downside risks to the IoT carrier revenue forecast, should price trends prevail (Figure 2).

For more details, please see https://www.delloro.com/products-and-services/carrier-economics-2

[wp_tech_share]

I just finished up three days of meetings at OFC 2018. I wanted to share a few of the key highlights from the show relating to Optical Transport.

First, I think the components market is gearing up nicely to enable optical system vendors to deliver better coherent line cards. What do I mean by better? With the help of a new generation of coherent DSPs and 64 Gbaud optics (transmit, receive, and drivers), 200 Gbps WDM wavelengths will be able to reach farther before a costly signal regeneration needs to take place, and fiber constrained buyers can opt to turn on higher speed wavelengths such as 400 Gbps, 500 Gbps, and 600 Gbps. However, for now, 600 Gbps span length will likely be under 200 km depending on conditions. This is really about producing better performing line cards capable of expanding the 200 Gbps and higher speeds into a broader market.

One of the critical items to enable higher wavelength speeds is a great coherent DSP, and there were a few announced at the show this year. Companies with new DSP announcements were Acacia, NEL, and Nokia. All three seemed to have a similar readiness timeline—available soon and in optical system vendor line cards ready for sale by the end of 2018. Among these three, Nokia’s new PSE3 seemed to be pushing the technology envelope (benefit of Bell Labs). Beyond the fancy name—probabilistic constellation shaping—what the new DSP does is improve optical performance by optimizing the modulation constellation; not by changing it from 64QAM, but by selecting the best points on the 64QAM constellation to use (I’m over simplifying what the PSE3 does, but I accepted the technology is beyond my grasp).

Also, among the three DSPs, Acacia and NEL’s development is really important to the optical ecosystem. The reason being that outside of the system vendors that design and make in-house DSPs, everyone else will be relying on these two vendors to allow them to compete against vendors such as Ciena and Huawei, companies that already have working 400 Gbps single wavelength solutions. This is really important in the very competitive and fast moving market for disaggregated systems. New 600 Gbps wavelength capable systems relying on these two DSP companies were announced by ADVA, Cisco, Coriant, Fujitsu, and PacketLight.

Second, I learned a few new things that I found interesting. As some of you may know, fast encryption has become an important requirement. In the past optical system vendors would add encryption on to their FPGAs, but it took time to license and get the IP (intellectual property) rights. What if there was an easier way? I spoke with Microsemi and learned that they have a new OTN chip called DIGI-G5 that not only performs 1.2 Tbps of OTN switching but also wire speed encryption, all in one ASIC. DIGI-G5 will also support 25 Gbps and 50 Gbps Ethernet as well as 400 Gbps Ethernet.

I also learned that adding optical DWDM optics onto switches and routers is back. It was just a little over 10 years ago that Cisco first added 40 Gbps DWDM optics to their large core router, and over time there have been a number of attempts to grow sales into this type of hardware convergence, but it never amounted to much. Well, the idea isn’t new, but there is a slight twist. Rather than add DWDM optics to large, core routers and switch, system houses are looking to add them to smaller access or edge routers and switches. This is what we saw from Ciena’s new 8180 and Juniper’s new ACX6360. Both platforms will have packet aggregation/switching and 200Gbps or higher DWDM optics. I think the opportunity for this type of platform could be better than the false starts in the past. The reason is that as you move closer to end-users, Ethernet traffic and switch/routing become more pervasive, and with Fiber Deep and 5G, the capacity requirements will also undoubtedly increase.

There were a lot of interesting things announced at OFC. Unfortunately, I’m not able to write about all of them here. Other announcements that were exciting include:

  • Infinera announced plans for two new optics engines: ICE5 for late 2018 and ICE6 for 2020.
  • Jabil will have a factory direct model for CFP2-ACO type 2 modules.
  • Huawei’s in-house CFP-DCO is now 200 Gbps capable.
  • NeoPhotonics developed a COSA (coherent optical subassembly) for 400 Gbps ZR that is meant for an OSFP pluggable.