Data center users have increasingly moved away from traditional 3-tier network architectures towards flatter, scalable designs that support virtualized environments and higher East-West traffic. Many users gravitated towards fabric-based switching because it can offer more reliable performance, greater agility and scalability, and enhanced integration with orchestration systems. As data centers adopt software-defined technologies, cloud architectures and open hardware, the choice of network fabric becomes critical to realizing scale, flexibility, efficiency, and total cost of ownership (TCO) goals. But not all fabrics are created equal.
Brocade offers two options for data center networking, VCS Fabric and IP Fabric (deployed on VDX and SLX switches). VCS’ primary advantage is it’s simplified management, while IP Fabric enables greater scalability. Unfortunately, a major downside of both these fabrics is the uncertainty around the fate of the VDX switch platform/Network OS, which was recently sold to Extreme Networks.
The good news is that Brocade users seeking a scalable fabric with simplified management are taking a closer look at Big Switch’s Big Cloud Fabric (BCF).
Big Cloud Fabric™, Big Switch’s next-generation data center networking fabric, was designed specifically to support modern data centers, providing open SDN networking for faster service enablement and lower OPEX. Brocade’s Network OS Data Center fabrics provide some benefits to cloud and SDDC environments, but they each have deficiencies that prevent them from serving as an ideal networking fabric.
Here’s a quick comparison of how BCF and Brocade data center fabrics handle provisioning, troubleshooting, upgrading and visibility.
Comparison of Standard Operations (plus scale and hardware options)
In each instance, Big Cloud Fabric offers greater operational advantages versus Brocade DC fabrics. BCF also provides the deep integration with VMware, OpenStack, and container orchestrators that today’s data centers need in order to automate network provisioning and rapidly troubleshoot issues.
Brocade’s Network OS Data Center Fabrics
Brocade’s Network OS data center fabrics include VCS Fabric and IP Fabric. They both deploy on the same proprietary hardware — the VDX switch platform — but each fabric forms a dramatically different solution with its own pros and cons.
VCS Fabric: While VCS Fabric is an L2/L3, topology-agnostic switching fabric for the data center, it has significant downsides related to provisioning, upgrade process, ease of management, scale and vendor specific hardware.
Provisioning challenge: Individual switches can only be added to the fabric if they share the same NOS version prior to insertion. Loading NOS on a device must be done via FTP or a USB. The upgrade process is painful in that the initial image must be extracted to the firmware directory of the USB (15-18 minutes), then download and reboot takes an additional 20 minutes.
Manual operations via command line interface (CLI): While a centralized CLI does offer some management conveniences, each switch must still be individually called up in the centralized CLI in order to perform standard operational tasks, such as troubleshooting.
Bolted-on network management: VCS Fabric supports greater management functionality and a graphical user interface through Brocade Network Advisor (BNA). Unfortunately, this is not an integrated product and comes at an additional cost. Without BNA, fabric management is performed solely through a CLI. A CLI’s character-based interface may be sufficient for certain tasks, but a well designed GUI can provide fabric-wide visibility, analytics, and workflow templates, significantly lowering the time it takes execute standard network operations, such as deployment, troubleshooting, and upgrades — which is especially important as networks expand and IT groups are expected to do more with existing resources.
Scale: VCS Fabric can scale only up to 48 switches in a Fabric. The Fabric topology is built using Brocade’s proprietary FSPF routing protocol that has it’s own limitations in terms of fabric scale.
Hardware: The future of the VDX switch platform is also worth considering for several reasons:
VCS Fabrics can only be deployed on Brocade’s VDX switches, many of which have been end of life’d. You cannot deploy VCS on other switches that Brocade offers such as the ICX, SLX or MLX.
One of the value propositions that the VDX provides for is the ability to support native fibre channel. What is often misunderstood is that this FC implementation is very limited in scale and is good for no more than one hop through the fabric (in and out). If more than one hop is required - and this is most often the case - then the addition of a true FC switch is required.
IP Fabric: Brocade’s IP Fabric is an L3 leaf-spine fabric that has a scale advantage over VCS Fabric, but because the fabric must be managed in the traditional networking approach, it suffers the management downside of the fabric having to be provisioned and configured manually on a switch by switch basis. If something goes wrong or if the fabric requires upgrading, significant management resources need to be allocated to servicing these tasks. Brocade WorkFlow Composer can be used to cut through some of this management complexity, but its functionality is currently limited. From a management perspective, IP Fabric has limited advantage over traditional “switch-by-switch” networking.
Taken together, with their best features combined, VCS Fabric/IP Fabric is an adequate networking solution that is widely deployed— VCS Fabric offers management benefits, while IP Fabric offers scale. Unfortunately, they each have technical pros and cons summarized below:
What is the Future of Brocade?
The recent acquisition by Extreme Networks has led to uncertainty around the future of Brocade data center fabrics. VDX itself is based on Brocade’s own proprietary ASIC architecture. Many of the switches in this platform have been discontinued, and the direction of the platform is unclear. Even though IP Fabric can also be deployed on the SLX platform, the future status of SLX is also in question. With significant talent drain, feature development and product roadmaps for Brocade fabric solutions are expected to stagnate.
For Brocade fabric users looking for a better option, Big Cloud Fabric offers superior resiliency, scalability, simplified management, and deep integration with SDDC and Cloud environments.
Big Cloud Fabric — Next-generation Data Center Switching
Big Cloud Fabric is a leaf-spine topology fabric leveraging SDN and open networking (britebox/whitebox) switches. It combines the main advantages from both Brocade Network OS Data Center Fabrics, plus offers automation, zero-touch operations, visibility, and innovation velocity achievable with BCF’s controller-based SDN architecture.
A key benefit of software-defined networking is the increased visibility possible with the adoption of an abstracted and hierarchical control plane. Big Cloud Fabric’s (BCF) controller provides a single point of management and visibility across a fabric of networking switches. Its logical chassis, “one big switch” architecture enables the creation of resilient, high-performance, and scale-out data center fabrics. The BCF Controller (aka chassis supervisor) enables zero-touch fabric operation, single-point of fabric management and centralized fabric analytics, significantly speeding up application deployment and fabric upgrade, and reducing operational costs.
Big Cloud Fabric as Disaggregated Chassis
BCF’s L2 and distributed L3 for scalable routing means it’s an ideal fabric for private cloud and SDDC environments. It also supports multiple deployment scenarios, including data center fabric, physical underlay, and unified physical/virtual fabric (using its virtual switches).
Another advantage of BCF is its richly featured, high-performance GUI. Physical and logical topology views, heatmaps, Fabric Analytics, operations workflows, fabric test path, and VMware, OpenStack, and container-related views accelerate operations.
Big Cloud Fabric GUI
The Big Cloud Fabric controller directly integrates into VMware vSphere™ to simplify and automate application deployments on its physical SDN fabric — across both virtual and physical networks. BCF’s GUI and Fabric Analytics provide deep visibility of VMware ESXi hosts, virtual networking with vSwitch/vDS and VMs for network administrators, and enables VM-to-VM path testing.
vCenter Integration in BCF GUI
Fabric Analytics provides preconfigured dashboards that cover both logical and physical components and topologies. Physical dashboards include Switch Interface, Configuration Changes, and Errors. Logical dashboards include Tenant, Segment, Endpoint, Protocols, and VMWare vCenter. Together these dashboards enable admins to rapidly discover and troubleshoot issues.
Fabric Analytics in BCF GUI
BCF also supports seamless integration with OpenStack through the Neutron plugin for L2/L3 networking and L4-7 service insertion. BCF’s UIs and Horizon dashboard enhancements enable seamless management of both physical and virtual infrastructure.
Big Cloud Fabric Highlights
According ACG Research’s 2017 report, Creating Agility & Efficiency at Scale, Big Cloud Fabric running on open networking hardware provided significantly greater benefits in NFV deployments compared to a tightly bundled platform.
Big Cloud Fabric has proven advantages delivering service velocity, operational simplicity and overall lower TCO. If you’re currently a Brocade user looking for an upgrade, it may be time to make the switch to Big Switch.