With the rapid development of data center and 5G technology, faster processing, more bandwidth and increased density are required in data center network. In order to meet the rising demand and data volume, 400G optical modules came into being.
The demand for hyperscale data center networks has almost doubled in a year and a half to two years. In order to cope with the continuous growth of data volume, hyperscale data centers upgrade network systems on average every three years. According to this progress, 400GbE is expected to be deployed in 2020, and will enter the expansion deployment phase in about 2022.
Although there is no doubt that 400GbE will become mainstream soon, there is still controversy about which pluggable form factor that this technology is best delivered. We can all acknowledge that cost is a main concern for any technology transition. Thus, driving the optics volume up to drive the cost down is imperative for smooth transition.
In the current 400G Transceiver, 8 channels 53Gbps PAM4 (400G-SR8, FR8, LR8) or 4 channels 106Gbps PAM4 (400G-DR4, FR4, LR4) are mainly used on optical port to realize 400GbE transmission, while on electrical port, using 8 channels 53Gbps PAM4 (OSFP or QSFP-DD). Both OSFP and QSFP-DD can provide 8-lane electrical interface. In comparison, QSFP-DD is smaller, which is more suitable for data center applications. OSFP is slightly larger, and because it can provide more power, it is more suitable for telecommunications application.
400G is coming with 4 form factors: CDFP, CFP8, OSFP, QSFP-DD.
|Interface||Optical I/O||Transmission mode||Standards||Electrical I/O||Form factor|
|400G-SR16||16*26.5Gbps NRZ||100m MM||802.3bs||16*26.5Gbps NRZ||CDFP/CFP8|
|400G-FR8||8λ *53Gbps PAM4||2km SM||802.3bs||8*53Gbps PAM4||QSFP DD/OSFP|
|400G-LR16||8λ *53Gbps PAM4||10km SM 8||802.3bs||8*53Gbps PAM4||QSFP DD/OSFP|
|400G-SR8||8*53Gbps PAM4||100m SM||802.cm||8*53Gbps PAM4||QSFP DD/OSFP|
|400G-SR4.2||4*2λ *53Gbps PAM4||100m MM||802.3cm||8*53Gbps PAM4||QSFP DD/OSFP|
|400G-DR4||4λ *106Gbps PAM4||500m SM||802.3bs||8*53Gbps PAM4||QSFP DD/OSFP|
|400G-FR4||4λ *106Gbps PAM4||2km SM||100G/λ MSA||8*53Gbps PAM4||QSFP DD/OSFP|
|400G-LR4||4λ *106Gbps PAM4||10km SM||TBD||8*53Gbps PAM4||QSFP DD/OSFP|
|400G-ZR||DWDM+59.8Gbd DP-16QAM||>80km DWDM||OIF||8*53Gbps PAM4||QSFP DD|
- QSFP-DD can forward and backward compatibility with QSFP, QSFP28, AOC/DAC and so on.
- Using 8-lane electrical interface, operate up to 25Gb/s (NRZ modulation) or 50Gb/s (PAM4 modulation), supporting up to 200Gb/s or 400Gb/s aggregate bandwidth. QSFP-DD can enable up to 14.4Tb/s aggregate bandwidth in a single switch slot.
- Using SMT (Surface-Mount Technology) connector and 1xN cage, cage design optimizations and module case optimizations enable thermal support of at least 12W per module.
- Support media and transceiver types include passive Direct Attached Copper cables (DAC), multimode fiber (MMF) and single mode fiber (SMF).
QSFP-DD compared with QSFP (QSFP+ / QSFP28)
The bandwidth of QSFP-DD can be up to ten times of QSFP+ or four times of QSFP28.
QSFP has four electrical lanes with a rate of 10Gb/s (QSFP+) or 25Gb/s (QSFP28) per channel, providing solutions 40Gb/s or 100Gb/s aggregate.
QSFP-DD pluggable transceiver provide an eight-lane electrical interface, and each lane can support data rate up to 25Gb/s (NRZ modulation) or 50Gb/s (PAM4 modulation). Thus, it can support data rate up to 200Gb/s or 400Gb/s aggregate bandwidth.
QSFP-DD is backward compatible with QSFP+/QSFP28, ensures that users have access to technologies that are already optimized for the systems it is being designed in to.
QSFP-DD and QSFP+/QSFP28 have the same port density. Since each QSFP-DD port can support 8 lanes instead of 4 lanes, QSFP-DD doubles the number of ASIC ports it already supports (such as CAUI-4).
The mechanical interface of QSFP-DD on the host board is slightly deeper than QSFP+/QSFP28, and is used to accommodate an additional row of contacts.
Compared with QSFP+/QSFP28, QSFP-DD only increases the length a little to drive the bandwidth up to 10 times/4 times of the latter without changing the port density. It is also backward compatible, which can widespread benefit from network upgrade strategies, cost-reduction strategies, and development cost reduction strategies.
QSFP-DD compared with OSFP
OSFP (Octal Small Form Factor Pluggable) supports 8 electrical lanes of 50Gb/s for an aggregate data rate of 400Gb/s, which is not compatible with existing optical port. Cooling is achieved by integrating heat sink directly into the form factor so the size of OSFP is slightly larger than QSFP-DD. Therefore, OSFP need a larger area of PCB, which the structure of the PCB, hosting cage and 1RU front panels need to be rebuilt. The pins of its electrical port are different from QSFP-DD, with a row above and below.
OSFP can backward compatible with QSFP+/QSFP28, but need to use an additional OSFP to QSFP adapter to achieve.
QSFP-DD can currently only support up to 400Gb/s, but OSFP can support up to 800Gb/s.
QSFP-DD is mainly used for the large 400G network (and 200G for 100G to 400G transition) that is about to be deployed on a large scale, and OSFP is more likely to be prepared for the 800G network in the future. Therefore, combined with the current situation, QSFP-DD is more suitable as a 400G form-factors.
By comparing QSFP-DD with other, we believe QSFP-DD can meet the needs of market, low-cost and smooth transition. The rapid development of QSFP-DD will greatly benefit on the whole optical interconnect and communication industry. Let’s wait and see!
More product details about QSFP-DD please click HERE