AWS Replaces Fat-Tree Data Center Networks with Random Graph Theory, Cutting Routers by 69%
AWS disclosed that Resilient Network Graphs, a flat network architecture based on quasi-random graph theory, is now the default for most new data center builds. The design replaces fat-tree hierarchies with direct ToR-to-ToR mesh connections using passive optical ShuffleBoxes, cutting routers by 69%, boosting throughput by 33%, and reducing network power consumption by 40%. By Steef-Jan Wiggers
AWS has likely reached a scale where traditional data center network architectures are becoming a significant bottleneck for efficiency and cost, necessitating a foundational redesign.
This move by AWS, a leading cloud provider, indicates a significant architectural shift in data center design, impacting cost structures, performance, and potentially the future of cloud computing infrastructure.
Data center networks are moving away from traditional fat-tree hierarchical designs towards flatter, more efficient quasi-random graph theory-based architectures with direct ToR-to-ToR connections.
- · AWS
- · Cloud consumers
- · Optical networking vendors
- · Hyperscalers
- · Traditional network hardware vendors
- · Companies with legacy data center infrastructure
AWS will benefit from significant reductions in operational costs and improved performance in its data centers.
This architectural innovation could influence other hyperscalers and large enterprises to adopt similar network designs, driving a broader industry shift.
Increased efficiency and lower costs could lead to more aggressive pricing in cloud services, further accelerating cloud adoption globally and potentially fueling more compute-intensive applications like AI.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at InfoQ