By Sterling Perrin, Principal Analyst, Heavy Reading
For the past several years, the optical industry has been driven by data center interconnect. We expect this trend to continue at OFC 2018 – with a host of announcements driving line rates up to 600 Gbit/s as well as boosting densities and reducing power consumption. These are good developments for the future of Webscale applications can cloud computing.
But -- more quietly – we are also expecting a separate trend to emerge this year – the return of packet-optical convergence, driven by the coming of 5G. There is good reason why packet-optical suppliers have been quiet lately - the market segment has struggled in recent years. As tallied by Heavy Reading, metro P-OTS revenue declined globally in 2017, following a flat performance in 2016. By contrast, the metro DCI segment has boomed during this time-period, increasing revenue at double-digit rates.
The divergent fortunes of P-OTS and DCI have led some industry observers to wonder if the P-OTS era is over? Is functional convergence (packet and optical) being replaced by functional disaggregation (as evidenced by optical white box)? Is traffic grooming and aggregation being replaced by point-to-point connected and optically switched wavelengths?
We think the answer is no, and that we are at the early stage of a packet-optical transport renaissance that will be driven by the specific requirements of 5G access and aggregation networks. Much is still being decided, but we know some fundamental things about 5G transport today:
- The 5G RAN will be fiber-led, creating a new role for fiber access technologies (including DWDM as well as PON variants) that did not exist in 4G or previous generations.
- The new eCPRI standard is creating a strong packet aggregation need in fronthaul and mid-haul RAN architectures. Technically, the “e” stands for “enhanced,” but the CPRI group expects the eCPRI transport network to be Ethernet-based, and many network operators want Ethernet fronthaul as well.
- 5G access networks don’t need 100G+ wavelength capacities. Rather, operators are designing 5G access networks with 10 Gbit/s access requirements at the cell tower in the near-term, with longer term plans to move to 25 Gbit/s Ethernet (one of several new Ethernet standards aimed at sub-100 Gbit/s applications).
Together these 5G access and aggregation requirements map well to the overall P-OTS value proposition: aggregating and grooming Ethernet-based traffic for efficient transport over higher speed optical links. We also note that the current and future class of purpose-built metro DCI gear doesn’t map well to the emerging 5G requirements, even at this early stage. 5G access and aggregation will require their own purpose-built optical developments, and market size will justify such dedicated investment.
We believe that smart optical suppliers will prepare for these two mega-trends – increasing 200 Gbit/s+ and functional disaggregation for DCI and increased Ethernet aggregation, packet-optical convergence, and sub-100 Gbit/s Ethernet feeds for 5G access and aggregation.
Posted: 13 March 2018 by
Sterling Perrin, Principal Analyst, Heavy Reading
| with 0 comments