All-Optical Switching Supports Full Mesh Backbone Networks to Achieve One-Hop Transmission Between DCs

Cloud-based service provisioning becomes an irreversible trend. In future networks, services, IT, and networks can all be implemented and deployed using cloud technologies to reduce costs and improve efficiency. Generally, cloudified network resources are centrally deployed in DCs.

It is predicted that DC traffic will maintain an annual growth rate of about 30%, and the inter-DC (east-west) traffic will increase to four times the existing volume in the next five years (a CAGR of 32%). Traditional backbone networks often use a ring or chain topology. The traffic on each node is mainly pass-through traffic, and the capacity of the network is consumed fast. If a ring network has 10 or more nodes, even when the traffic on each node is 1 to 2 Tbit/s, the traffic on the ring network will be 10 to 20 Tbit/s. Therefore, ring networks require frequent capacity expansion, unable to meet the need for large-capacity transmission and grooming.

Meanwhile, service cloudification places higher requirements on latency. For example, financial services require reducing every possible millisecond in latency, intra-city disaster recovery requires the latency to be within 1 ms, live streaming services within 20 ms, and cloud desktop services within 30 ms. Latency occurs on not only devices, but also fibers. Ring networks have many nodes and long routes, resulting in long latency. The direct connections between backbone nodes significantly reduce latency (by 50% in average).

The challenges of the cloud era can be addressed by changing conventional ring or chain backbone networks to mesh networks, which have many advantages: The one-hop connections minimize latency, bandwidth can be expanded on a per-hop basis, DCs have more alternative routes to each other in case of congestion or failure, making the network more robust and resilient.

However, the mesh connections require optical grooming in more degrees and ultra-large switching capacity on core nodes.

Huawei OXC all-optical switching solution has large-capacity switching and multi-degree optical-layer grooming capabilities, which can better meet requirements on full mesh connections on backbone networks.

32-degree lines: Each OXC supports 32-degree all-optical switching that can achieve full mesh interconnection between DCs. To add a degree, you only need to add optical line boards, ensuring stable network architecture and smooth expansion.

Large capacity: The OXC provides an 320T to 640T all-optical switching capacity and interworks with large-capacity OTNs to meet high grooming requirements on backbone networks and reduce network construction costs.

All-optical switching: The leading liquid crystal on silicon (LCOS) technology of Huawei OXC implements all-optical switching architecture. For optical wavelength grooming, device footprint and inter-device fiber connections can be reduced by 90%. For service grooming, conventional method of manually connecting fibers has been replaced by centralized grooming and intelligent management.

Optical monitoring technologies: The OXC provides intelligent optical monitoring functions that enable visibility into wavelength switching and transmission paths, resources, and performance data. Optical monitoring technologies help address the reliability and O&M challenges of all-optical switching.

Huawei OXC all-optical switching+large-capacity OTN solution is an optimal transmission solution for DC interconnection. It helps achieve non-blocking and full-mesh interconnection and one-hop transmission, and can smoothly evolve with the increase of DC capacities and degrees to protect investments. This solution also guarantees high bandwidth and low latency between DCs and from users to DCs, achieving optimal user experience.