With 100G in wide-scale deployment today and the promise of mainstream 400G deployment seemingly ubiquitous, Cloud Data Centers are eager to take advantage of any and every opportunity to bridge the throughput gap and keep pace with the data deluge. 200G (4 x 50G) optical modules answer this immediate need head-on.
At the broader market level, while 100G technology is already mature and component integration is well established, 200G end-to-end interoperable chipsets have just recently hit the market. Looking to the past as our guide, in the short term, 200G modules are expected to emulate cost structures akin to 100G modules when they entered the market a few years ago and follow a similar downward cost curve as component integration is further standardized and volume shipments accelerate. In due course, 200G modules are expected to achieve a cost structure that’s comparable to today’s 100G modules.
While 100G CWDM is a mature and well-understood technology and will continue to ramp in the coming year, many of the big Cloud Data Center OEMs are turning their sights to 200G, to meet the pressures of enabling faster connections at scale volumes.
Google started deployments in 2x200GbE transceivers in 2018 and we expect that demand for these products will peak in 2022, as Googles starts to transition to 2x400GbE modules.
Amazon
The forecast for 400GbE includes 4x100GbE DR4 modules selected by Amazon. These DR4 modules will be deployed in a breakout configuration with DR1 modules on the opposite side of the link. Effectively, each fiber will be carrying 100GbE traffic, aggregated into a DR4 module on one side. Deployments of true 400GbE transceivers will be limited in 2019-2021 to upper levels of switching in mega-datacenters and core routers. Implementation of high-radix configurations in leaf and spine networks using 400GbE connectivity will be challenging until switching ASICs reach 51Tbps capacity, probably by 2022-2023.
Facebook is staying with 100GbE for now and plans to use 200GbE next.
More than 2.6 billion people now use its services.
Facebook publicly stated their intent to stay with 100GbE optics for now and use 200GbE or 400GbE transceivers in the next upgrade cycle in 2021-2022. Facebook’s new F16 data center network architecture， will require 3-4 times more optical connections compared to their previous design (F4). The first implementation of F16 topology will rely on 100GbE CWDM4 transceivers, boosting demand for these modules in 2020-2022.
Facebook is already the largest consumer of 100GbE CWDM4 modules. They use a sub-spec version of CWDM4 transceivers with 500-meter reach instead of 2km, also known as CWDM4-OCP (for Open Compute Project). The latest forecast database includes sub-spec CWDM4 modules as a separate category. Segmenting the sub-spec products also helped us to refine the market data collected for 2018, resulting in higher than previously reported sales.
Once these issues are resolved, the demand for CWDM4 is expected to skyrocket in the second half of 2019 and make a real difference to the market in 2020-2022. Sales of sub-spec CWDM4 modules are projected to peak in 2022, as Facebook starts the transition to 200GbE connectivity.
Conclusion
Though intermediate between 100G, and 400G – the customer demand for 200G, is shaping this market to be sizable, with deployments expected by the second half of next year. The good news for module vendors is that there are multiple component vendors, such as MACOM, who have 200G compatible components on the market today.
Comparing the cost for 100G versus 200G, we have to look specifically at the cost of components themselves. While 100G is already at the point of integration, 200G end-to-end operable chipsets have just hit the market. 200G will therefore emulate a similar price point as 100G did when it entered the market a few years ago, following a similar cost curve as integration furthers.
Gigalight is committed to leading the evolution of Data Center interconnects from 100G to 200G and 400G. Gigalight 200G products such as 200G QSFP-DD SR8 NRZ 100m, 200G QSFP-DD PSM8 NRZ 2km, 200G QSFP-DD PSM8 NRZ 10km, 200G QSFP56 SR4 PAM4 100m, 200G QSFP56 FR4 PAM4 2km, 200G QSFP56 LR4 PAM4 10km and so on. Among them, the Gigalight 200G QSFP-DD PSM8 NRZ 10km optical transceiver (GDM-SPO201-LR8C) is an eight-channel, hot-pluggable, parallel, fiber-optic QSFP Double Density module designed for 2×100-Gigabit Ethernet PSM4 and InfiniBand DDR/EDR applications. It is a high-performance module for data communication and interconnects.
The 200G QSFP-DD PSM8 (dual PSM4) module integrates eight data lanes in each direction. Each lane can operate at 25.78Gbps up to 10km over G.652 SMF. It is designed to operate over single-mode fiber systems using a nominal wavelength of 1310nm. The electrical interface uses a 76-contact edge type connector. The optical interface uses a 24-fiber MTP/MPO connector. This module incorporates Gigalight proven circuit and optical technology to provide reliable long life, high performance, and consistent service.
Source: LightCounting