+86 166 2042 3014 WhatsApp: +86 166 2042 3014 [email protected] Mon–Fri, 09:00–18:00 CST
Get Quote

Goodbye to Expensive DWDM Transport: 800G ZR/ZR+ Coherent Modules Enable Zero-Relay Cross-City Interconnect

As cross-region AI compute coordination becomes national strategy, enterprises face an awkward gap: single-city compute centers are maxed out, while cross-city interconnect remains absurdly expensive.

Traditional cross-city optical transport means steep dark-fiber leases plus complex DWDM transmission equipment in both end sites. This transport-heavy architecture takes weeks to build, burns power and multiplies operational complexity. In 2026, the arrival of 800G ZR/ZR+ coherent pluggables is rewriting that economics.

The technology shift: IPoDWDM powered by 3nm DSP and silicon photonics

800G long-haul used to be carrier-backbone territory. Now, with volume production of ultra-low-power 3nm DSPs and silicon photonics, 800G coherent technology has moved down into data centers and enterprises, enabling lean IP over DWDM networking: plug an 800G ZR module straight into the switch and replace the router plus OTN/DWDM transport stack - true plug-and-play.

ZR or ZR+: 80km metro vs 800km cross-region

800G ZR is built for metro interconnect: sub-24W power and 80km single-span, relay-free reach - the default choice for same-city disaster recovery and hot compute migration between two data centers.

800G ZR+ adds probabilistic constellation shaping (PCS). While keeping the full 800G line rate, it stretches relay-free reach to 800km-1000km, delivering direct cross-province fiber paths and eliminating mid-span relay sites and their cost.

Cost and power: the compounding effect

Beyond box savings, removing relay sites cuts power, colocation and maintenance linearly with distance. For multi-city AI fabrics, total TCO of ZR/ZR+ pluggable architectures typically beats traditional DWDM transport by a wide margin over a 5-year horizon.

OTSLink 800G coherent portfolio

OTSLink offers 800G ZR/ZR+ OSFP and QSFP-DD coherent modules with per-unit pre-shipment testing across mainstream switch platforms, plus link-budget design support, helping you build a zero-relay AI compute backbone.