Open Distributed Core Architecture

Scaling the Data Center Core

As data centers scale to support thousands of servers, IT managers are seeking better ways to network those servers while also reducing costs and power consumption. Moreover, especially in large-scale compute-cluster environments inter-node communication bandwidth is increasingly becoming the main bottleneck. For these environments, where east-west data center traffic is both significant and networking requirements in terms of bandwidth, QoS, and latency stringent, new data center networking architectures are required. The answer is an open distributed core architecture

An open distributed core architecture is a new way to design high-performance data center core networks made possible with the new Z-Series and Z9000 distributed core switching system. Based on Clos network design principles and using the Z9000 as the primary building block, data center operators can construct highly-scalable, non-blocking, leaf-spine fabrics capable of scaling  up to 160 Tbps. And since the Z9000 is only 2 rack units and draws less than 800 watts, these fabrics are extremely cost-effective compared with large chassis based implementations.

And best of all, our fabric solutions are 100% open, leveraging standards-based control-plane protocols allowing mix-and-match capabilities within your core. Distributed core solutions can be deployed at Layer 3 using OSPF, BGP and ECMP or at Layer 2 using TRILL allowing for maximum flexibility and interoperability. For ease of management, our open distributed core solutions are fully supported within our Open Automation framework and future Open Fabric Manager providing simplified provisioning and management, standards-based scripting  environments and multi-vendor virtualization capabilities.

Find out what an open distributed core architecture can do for you.