RLS Hyper-Rail: Scaling Optical Networks for the AI Era

1 00:00:00,000 --> 00:00:02,940 AI is redefining optical infrastructure. 2 00:00:02,940 --> 00:00:06,120 Clusters are scaling from thousands to millions of 3 00:00:06,120 --> 00:00:09,870 GPUs — no longer confined to a single building. 4 00:00:09,870 --> 00:00:12,330 They're spreading across geographies, 5 00:00:12,330 --> 00:00:15,060 wherever power and space are available. 6 00:00:15,060 --> 00:00:16,770 Our customers are facing 7 00:00:16,770 --> 00:00:19,920 real challenges — and Ciena can help solve them. 8 00:00:19,920 --> 00:00:23,100 The next generation of AI networks will 9 00:00:23,100 --> 00:00:26,835 require tens of petabits per second of capacity, 10 00:00:26,835 --> 00:00:29,085 across hundreds of fiber pairs. 11 00:00:29,085 --> 00:00:30,540 But they can't get there by 12 00:00:30,540 --> 00:00:32,565 lighting up one wavelength at a time. 13 00:00:32,565 --> 00:00:35,100 That model was never built for this scale. 14 00:00:35,100 --> 00:00:37,410 To unlock what comes next, 15 00:00:37,410 --> 00:00:39,570 we had to rethink the photonic layer 16 00:00:39,570 --> 00:00:42,180 itself — create a solution for 17 00:00:42,180 --> 00:00:45,870 fast deployment of multiple fully filled fiber pairs 18 00:00:45,870 --> 00:00:48,795 – or rails – deployed over the same route. 19 00:00:48,795 --> 00:00:52,320 And that's what Ciena's RLS Hyper-Rail delivers. 20 00:00:52,320 --> 00:00:57,600 With RLS Hyper-Rail, we move from one rail per module to 21 00:00:57,600 --> 00:01:03,075 four — enabling up to 128 rails in a single rack. 22 00:01:03,075 --> 00:01:07,035 That's up to 32x more density and 23 00:01:07,035 --> 00:01:09,420 up to 75% power savings 24 00:01:09,420 --> 00:01:12,600 — and that's only part of the equation. 25 00:01:12,600 --> 00:01:17,070 Because at hyperscale, it's not just about capacity. 26 00:01:17,070 --> 00:01:20,115 It's also about how fast you can deploy it. 27 00:01:20,115 --> 00:01:22,590 Our customers need to turn up hundreds of 28 00:01:22,590 --> 00:01:26,220 fiber pairs — quickly and reliably. 29 00:01:26,220 --> 00:01:28,830 RLS Hyper-Rail is built with 30 00:01:28,830 --> 00:01:30,495 advanced instrumentation 31 00:01:30,495 --> 00:01:33,360 enabling automated deployment at scale. 32 00:01:33,360 --> 00:01:38,770 Ciena RLS Hyper-Rail — built to power AI at scale.