Next-Gen Data Center Switching Fabric

 

BSN Labs

Try Big Cloud Fabric Hands-on in the Online Lab Now!

Big Cloud Fabric™ is the industry’s first data center fabric built using open networking (white-box or brite-box) switches and SDN controller technology.  Embracing hyperscale data center design principles, Big Cloud Fabric solution enables rapid innovation, ease of provisioning and management, while reducing overall costs, making it ideal for current and next generation data centers.

Big Cloud Fabric is designed from the ground up to satisfy the requirements of physical, virtual, cloud and/or containerized workloads. Some of the typical OpenStack or VMware data center workloads include NFV, VDI, High Performance Computing, Big Data and Software Defined Storage (SDS).  Big Cloud Fabric can be easily deployed in existing (brownfield) data centers, as a pod fabric, and supports full communication between modern Big Cloud Fabric pod and legacy network via core router.

Available In Two Editions

  • Physical Fabric Edition: a leaf and spine physical Clos fabric controlled by the Big Cloud Fabric SDN controller
  • Physical + Virtual Fabric Edition: a leaf, spine and virtual switch fabric controlled by the Big Cloud Fabric SDN controller

With Big Cloud Fabric, data center infrastructure teams can automate their VMware or OpenStack networks with 16x faster configuration and upgrades, 4x faster application deployment and issue resolution and troubleshooting.

The ACG Research report provides a detailed analysis: The Economic Value of Open SDN Fabrics.

Physical + Virtual Fabric Architecture

The Big Cloud Fabric architecture maps readily to a traditional redundant chassis switch architecture:

  • Supervisor Card(s) --> Controller Cluster
  • Chassis Backplane --> Spine Switches
  • Line Card(s) --> Leaf Switches, virtual Leaf Switches

Big Cloud Fabric's logical chassis architecture -- or a one logical pod switch -- enables data center users to construct high-performance, resilient and scale-out (pay-as-you-scale) pod fabrics.  SDN controller (aka chassis supervisor) is a critical and necessary component without which chassis disaggregation is not possible.  It enables zero-touch fabric operation, single-point of fabric management and centralized fabric analytics -- thus significantly speeding up application deployment and reducing operational cost.  Moreover, with HW/SW disaggregation, Big Cloud Fabric supports a variety of high-performance (10G/40G), high-density switch hardware enabling full vendor choice and tremendous capex cost optimization.

The Big Cloud Fabric Physical + Virtual Edition consists of the following components:

  1. SDN-enabled Big Cloud Fabric Controllers (HA Pair)
  2. SDN-enabled Switch Light OS
  3. SDN-enabled Switch Light VX virtual switch for KVM hypervisor
  4. Open Networking (white-box or brite-box) Ethernet Switches

Big Cloud Fabric Physical + Virtual Edition - Ideal OpenStack Networking Solution

The Physical + Virtual Edition enables full OpenStack network resiliency at scale, with no hairpinning of traffic, efficient use of compute resources and purpose built design for high availability. Additionally, it provides:

  • Single Physical + Virtual network for L2/L3 Neutron networking compared to costly and complex two-network alternative of overlay/underlay
  • Seamless Integration with Red Hat OpenStack Platform (RHOSP) and Mirantis OpenStack
  • Most scalable OpenStack SDN solution – validated for a 300 node scale benchmark
  • Integration with L4-7 tenant-level services including load balancing as a service and firewall service insertion for both north-south and east-west traffic

The solution supports multiple OpenStack use cases, including:

Physical + Virtual Edition Operational and TCO Benefits

The Big Cloud Fabric Physical + Virtual Fabric Edition brings modern hyperscale-style benefits to enterprise and service provider data centers.  

  • Simplicity: zero-touch P+V fabric and single-point of management compared to manage-twice overhead of managing two overlay/underlay networks
  • Agility: network automation for OpenStack, REST APIs for optional NetOps/DevOps automation
  • Visibility and Troubleshooting: Built-in P+V fabric analytics with historical replay, vSwitch--leaf--spine--leaf--vSwitch fabric tracing for rapid troubleshooting, end-point visibility at VM-level, tenant self-service troubleshooting in Horizon GUI.  Never have to deal with underlay/overlay blind spots
  • Economics: Dramatic reduction in OpEx and CapEx compared to pay-twice overlay/underlay alternative
  • Vendor choice: choice of open networking (white-box and brite-box) switches based on state-of-the-art Broadcom Trident-II, Trident-II+ and Tomahawk networking ASICs.
  • Innovation Velocity: Faster time to service enablement as well as rapid (hitless) software upgrade enables consumption of new features