Webinar: The 800G Migration Reality: Scaling DWDM Networks Without Redesign
Date & Time - 3 Jun 2026,
Duration - 60 min
Webinar Overview
AI has broken the old rules of capacity planning. Traffic growth used to be predictable enough to forecast three to five years out, now it's bursty, irregular, and compressed into one to two year cycles.
In this session, PacketLight's Erik Garland breaks down why 800G is less a speed upgrade and more an architectural and economic shift, and walks through practical, non-disruptive migration paths for brownfield DWDM networks.
Speaker
Erik Garland, Director of Sales,
PacketLight Networks
Erik Garland is Director of Sales at PacketLight Networks, driving adoption of next-generation DWDM and OTN transport solutions. With over a decade of experience in the fiber optic and wireline industry, he has worked with enterprises, carriers, and hyperscalers to design and scale high-performance optical transport networks.
Based in the San Francisco Bay Area, Erik focuses on helping organizations evolve their networks with open, modular, and scalable architectures. His work centers on enabling efficient bandwidth growth, reducing complexity, and delivering high-performance transport solutions for data center, AI, and mission-critical infrastructure.
Who will benefit from this webinar?
- Network architects and planning teams scaling toward 800G without a full network redesign.
- CTOs and executives managing capex, avoiding forklift upgrades, and planning for AI-driven traffic growth.
- Engineering and operations teams looking for practical migration paths for existing DWDM/OTN networks.
- System integrators and technology partners advising on optical transport upgrades for AI and DCI growth.
Key Takeaways
- Why 800G's value isn't raw capacity - The real win is needing half the number of wavelengths, lower power per terabit, and higher rack density, compared with 400G.
- Why forklift upgrades are off the table - Carriers and cloud providers "can't justify forklift upgrades across long-haul, metro, and DCI infrastructures."
- Three migration paths for brownfield networks - L-band expansion, flex-grid ROADM migration, a hybrid spectrum split that transports(?)legacy and 800G traffic separately.
- Why planning timelines are shorter- AI's bursty traffic between data centers and inference clusters has compressed 3-5 year capacity plans into 1-2 year cycles.
- Supply chain is now part of network design - component shortages and hyperscaler demands are extending lead times for coherent DSPs, pluggables, and ROADMs, making vendor diversification a planning requirement, not an afterthought.
- Coherent pluggable modules vs. IPoDWDM - a dedicated optical layer scales independently of router refresh cycles, while IPoDWDM ties your optical lifecycle to your IP lifecycle.
FAQs
1. Is the webinar recording available?
Yes. Watch the full on-demand recording above, anytime.
2. Does upgrading to 800G mean ripping out my existing network?
No. PacketLight recommends three brownfield migration paths, L-band expansion, flex-grid ROADM migration, or a hybrid spectrum split, that add 800G capacity without a full redesign.
3. Does 800G actually double my fiber capacity?
Not directly. Each 800G wavelength uses a wider channel, so total capacity is similar to 400G. The benefit is needing roughly half the wavelengths to carry the same traffic, with lower power consumption and higher rack density.
4. Why has AI shortened network capacity planning cycles?
AI workloads create bursty, unpredictable east-west traffic between data centers, inference clusters, and storage systems, compressing planning horizons from 3-5 years down to 1-2 years.
5. Should I use coherent pluggables or IPoDWDM?
A dedicated optical layer with coherent pluggables (ZR/ZR+) lets you scale IP and optical infrastructure independently. IPoDWDM can simplify architecture but ties your optical upgrade cycle to your router refresh cycle.
6. Who should watch this webinar?
Network architects, engineering and operations teams managing brownfield DWDM/OTN networks, and executives or system integrators planning AI-driven capacity growth.