It is truly exciting time at Big Switch where our non-stop software innovations in next-generation data center switching and monitoring continue to generate multiple “wows” with Global 2000 enterprise and service provider customers.  Instead of traditional value proposition of price-performance, port density and proprietary vendor lock-in features, we at Big Switch have focused on delivering operational simplicity, service agility, visibility and vendor-agnostic economics.  Our innovation velocity and value proposition have the underpinning of SDN-based control software and open networking switch hardware.  Concrete TCO benefits – CapEx (50% less) and OpEx (10X+ faster and simpler) – provide customers the needed business justification to atomically insert modern, Google-style network infrastructure in existing (brownfield) data centers.  Our latest releases of Big Monitoring Fabric and Big Cloud Fabric expand addressable use cases as well as support new Trident-II+ and Tomahawk based hardware platforms from Dell and Edgecore/Accton. Additionally, we are announcing tech preview of Big Cloud Fabric for Kubernes container environments.

Today, we support for deployment models with Big Switch fabric solutions – two for Big Cloud Fabric (OpenStack and VMware) and two for Big Monitoring Fabric (Out-of-band and Inline).  With today’s new releases and additional announcements, we are expanding product capabilities and use cases for all four deployment models.  This is shown in the diagram below. 

Big Cloud Fabric – Highlights

1.   NFV Everywhere Use Case: BCF has seen successful deployments in network function virtualization (NFV) use case.  Recently Verizon announced the use of BCF for telco NFV OpenStack cloud, along with Dell switches and Red Hat OpenStack Platform (RHOSP).  To meet telco customers’ requirements to roll-out NFV everywhere, BCF is adding new capabilities:

              a. Central Office Deployment: NEBS support

              b. Small PoP Deployments: VM-based BCF Controller

              c. Remote PoP Deployments: Stretched BCF Fabric with BCF Controller in regional DC/CO and a pair of leaf  switches in remote PoPs; support                           multiple remote PoPs within a metro region.

2.   Software-defined Storage Support: BCF has supported traditional disk/flash-based IP storage arrays, with lossless high-performance fabric, support for 10G/40G fabric as well as storage traffic prioritization.  BCF also supports software-defined storage – where commodity storage arrays can be pooled via software into a single logical disk – to provide petabyte size storage at very low cost.  While BCF can support any SDS solution, it is deployed with Red Hat Ceph. Please see additional details here.

 3.   Hyperconverged Infrastructure (HCI): BCF’s high-performance leaf-spine fabric is ideal for HCI as well. Big Cloud Fabric supports all HCI environments.  In particular, when running VMware Virtual SAN on Big Cloud Fabric, SDDC operators can get unique automation and visibility for Virtual SAN – thanks to BCF’s unique integration with VMware vSphere.  Once BCF is registered with vCenter and one-click fabric-wide multicast is enabled, Virtual SAN nodes can be added/deleted at will and clusters can be formed.  No configuration on BCF is required.  Even adding or removing new Virtual SAN hosts is detected automatically by BCF and leaf switch configuration is adjusted accordingly.   Please see video demo and our storage networking blog for additional details.

  4.   OpenStack Enhancements: BCF provides superior resiliency and scale for OpenStack clouds, due to it unified P+V architecture, where the BCF controller manages a single pSwitch + vSwitch network versus the 2X cost and complexity of overlay/underlay alternatives.  BCF now fully supports Red Hat OpenStack Platform (RHOSP) 8, and OpenStack Liberty release.  Furthermore, BCF has fully validated OpenStack solution with RHOSP 8 on Dell’s 300-node scale test bed, enabling data center and cloud organizations to deploy a resilient, scalable OpenStack cloud network. 

5.   Additional Enhancements*:  We are also announcing following enhancements which apply to any deployment scenario, including OpenStack and VMware.

                a. Inter-pod/inter-DC L2 VXLAN Extension: As customers deploy multiple BCF pods within and across data centers, we are leveraging HW                           VXLAN support in Broadcom Trident-II+ ASIC to extend L2 across BCF pods (within or between data centers).

                b. 2X Fabric Scale: Our current fabric scale of 32 leaf switches and 6 spine switches can support 600+ hosts and 12K+ VMs.  The 32-leaf scale                       has been limited by hardware, due to the 32x40G port density of spine switch.  With introduction of Dell S6100, a Broadcom Tomahawk switch that                 can support 64x40G, we can now double the BCF scale to support 1200+ hosts and 24K+ VMs in a single BCF pod.

 6.   L4-7 Technology Partnerships: Big Cloud Fabric supports built-in service insertion.  In the latest release, we have shown service insertion with F5 ADC for load balancing as a service (LBaaS) and Palo Alto Networks’ firewall for securing both north-south and east-west traffic.

BCF Tech Preview: Container Networking with Kubernets

While we continue to innovate BCF and BMF, we are also investing in emerging areas of container networking.  There are several key container networking requirements for enterprise and service providers to deploy them in production:

  • Support x86 based networking à virtual network (vSwitch) control is critical
  • Enable rapid connectivity across hosts/racks à physical network (pSwitch) control and automation is critical
  • Allow rapid troubleshooting of connectivity issues à no visibility gap across physical and virtual networks
  • Ensure security, load balancing à service insertion for firewall and load balancing services via physical and/or virtual appliances
  • Operate in a mix workload environment à support physical and VM workloads along with containers

Big Cloud Fabric P+V solution meets all of the above requirements, and hence is ideal for container networking.  As the below figure shows, Big Cloud Fabric has partnered with Docker and Mesosphere to support all three container-centric deployment models.  Our plug-ins for Docker and Kubernetes allow network automation and visibility across physical + virtual fabric, and allows a host to be a physical server, VM server or containerized server.  And there is tremendous synergy between Mesosphere DC/OS  (one logical computer) and Big Cloud Fabric (one logical switch), as was demonstrated during the DC/OS launch.

 Big Monitoring Fabric – Highlights

Big Monitoring Fabric – a next-generation network packet broker – has been at feature parity for some time.  Our investment in Big Mon is in regards to advanced DPDK-based services, sophisticated analytics and higher-density 40G/100G switch platforms. Popular customer use cases for Big Monitoring Fabric include: monitoring every rack, mobile/LTE monitoring and remote DC monitoring & DMZ/Extranet inline security.

1.   Advanced DPDK services: Today we support several services on the 40G Big Mon service node, namely de-duplication, packet slicing and regex matching/filtering. New services are announcing are Header Stripping and Netflow generation*. Header stripping service allows removal of user-defined packet header as well as well-known encapsulation protocols such as VXLAN, MPLS, ERSPAN and Cisco Fabric Path. Netflow generation service enables flow-level granularity for deeper security as well as network monitoring.

2.   High-density 40G/100G modular switch support*: Service providers and web-scale organizations need high-density and high-performance hardware for their monitoring and security needs.  With Dell S6100, Big Mon will meet this need, with up to 64x40G or 32x100G or mix-speed ports in a modular 4-slot, 2RU form factor.

3.   Integration with leading Tool Partners: Certified joint solution through Blue Coat's Encrypted Traffic Management (ETM) Program helps customers combat sophisticated attacks by addressing malware hiding in SSL traffic.  Also, Big Monitoring Fabric with the Riverbed SteelCentral NetExpress network performance management platform delivers all-in-one network monitoring solution combining flow and packet collection and analysis for the entire data center. Additionally, Riverbed NetExpress collect and visualize Netflow data that is generated by the Big Monitoring Fabric service node

Expanded Support for Open Networking Switches

Both Big Monitoring Fabric and Big Cloud Fabric will support latest Broadcom Trident-II+ and Tomahawk based Open Networking switches from Dell and Edgecore Networks, as shown in the table below.

Ready to modernize your data center networking step at a time? Start with a monitoring/security project or with a switching pod and experience the “wow factor” of Big Switch fabric solutions.  Experience it free on-line with Big Switch Labs or on-premise with attractively priced starter kits.  And if you are exploring cool use cases, perhaps enter our Community Edition Contest to win free open networking hardware..

 

Prashant Gandhi

VP, Products and Strategy


Additional Information