Editor's note: Brocade VCS Fabric is the winner of the November SearchNetworking Network Innovation Award. The...
network fabric technology enables engineers to configure and manage multiple switches as one body, and introduces a flat network design for multipathing and automation in the data center LAN.
All of the talk about network fabrics, multipathing and flat data center networks often overlooks a key challenge: Most network managers won't toss aside their traditional LANs to start fresh with a new architecture. Brocade was mindful of this issue when it launched its Virtual Cluster Switching (VCS) Fabric two years ago.
With a VCS fabric, if you have multiple links between any two switches in the fabric, the formation of that logical trunk is automated, as is the load balancing technique.
vice president of data center networking, Brocade
Like most network fabric vendors, Brocade wants engineers to rearchitect their data center LANs with VCS so every switch in the network is automatically linked and managed as one body. This architecture reduces the need for manual configuration and creates a flat network with direct communication between all nodes to better support virtual machine (VM) migration and converged networking.
But, Brocade allows customers to introduce fabric technology in stages -- or just a few switches at a time. In fact, VCS will soon be able to interact with Spanning Tree at the edge of the fabric, while using Transparent Interconnection of Lots of Links (TRILL) at the core in order to integrate fabric with existing data center architecture.
Brocade also added modular switches to VCS this fall to provide scalability and density for those looking to push into a larger deployment in a high-performance environment.
This month, Brocade was recognized with the SearchNetworking Network Innovation Award for its VCS Fabric. In this Q&A, Jason Nolet, Brocade' vice president of data center networking, explains how VCS changes the data center network.
What are the basic components of a Brocade VCS fabric?
Jason Nolet: A VCS fabric is made up of a combination of fixed configuration and modular switches. These are conventional Layer 2/Layer 3 switches that can run both at top-of-rack or at the middle-of-row, end-of-row aggregation layer. We started [two years ago] with the fixed configuration switches that supported 10 Gigabit Ethernet connectivity from the server into the network and 1 Gb connectivity that continues to be a large part of the install base. In the last couple of months we've enhanced that portfolio with some new additions, notably a pair of modular chassis that deliver VCS functionality that is more scalable and dense.
How does the VCS fabric enable network automation?
Nolet: We've delivered a series of automation capabilities at multiple layers of the network. Today when you add a link or switch to an existing conventional data center LAN, [it requires] a fair bit of manual configuration. To add a link, you have to get it up and running and make sure it can participate in all of the protocols and services that are available in that data center LAN. With VCS fabric, we have removed as much of that manual configuration as possible. With a VCS fabric, if you have multiple links between any two switches in the fabric, the formation of that logical trunk is automated, as is the load balancing technique.
Read more about Brocade technologies
New Brocade VDX chassis adds scale to data center network fabric
Brocade-Vyatta acquisition: Enhanced virtualization networking
Brocade: Some OpenFlow switches can't operate at line rate speeds
We [also] built in discovery of virtual machines attached to the network, [which includes] discovery of the port profiles or the networking requirements that each VM needs from the network. That includes quality of service or security. We then take that discovered information and propagate it throughout the fabric automatically, so that if the VM is moved from one server to another within the fabric, the receiving switch is able to apply those properties and network behaviors without manual configuration or coordination between the VM administrator and the network administrator.
Automated load balancing is a key part of Brocade VCS. Please explain the load balancing features.
Nolet: There are three different ways in which we multipath within the fabric. In Layer 1, when you have two or more links between any two switches, we are able to load balance across those links in an automatic fashion. That uses a patented technology that we lifted out of our Fibre Channel product line, which allows you to achieve load balancing on up to eight links between those two switches. At Layer 2, we use TRILL. At Layer 3, we allow multiple routing instances to appear to the rest of the fabric as a single Layer 3 instance, [acting] as a gateway if you want to route traffic out of the fabric and into the rest of the network.
Does the VCS fabric work across a multivendor environment, and can it be added to an existing, traditional data center architecture?
Nolet: When a switch is added to the fabric, it learns most of what it needs to know from the other existing switches in the fabric. It learns about the topology, about its peers and its configuration from the other members of the fabric. With respect to our fabric participating in an existing data center LAN architecture, VCS appears to the rest of the data center network as a single logical Layer 2 switch. Any traffic that is destined for either side of the fabric is passed through as if it were any Layer 2 switch. In fact, you're going to see us enhance the edge of our fabric to fully participate in Spanning Tree [while utilizing TRILL in the core] because there are many legacy-installed architectures that leverage [Spanning Tree].
If the goal of fabrics is to introduce network automation and multipathing, is there really a need for software-defined networking (SDN)? Do the two technologies intersect?
Nolet: When we think about SDN, we think about the ability to influence network behavior from outside the network by using OpenFlow protocols. OpenFlow is largely supplementary to existing forwarding and routing techniques in a network, and it's supplementary to VCS. What many are going to do with OpenFlow is say, "I've got some unique network behavior that I would like to instantiate using OpenFlow because my networking vendor doesn't provide it natively within the switch." But they're not looking for OpenFlow to displace all of their routing and porting techniques. Outsourcing the entire control plane would be a pretty big bite to take. Our intent is to implement OpenFlow, both within the VCS fabric and other data center platforms, in that same supplementary fashion.
For example, on our MLX platform, which is a high-end router for the data center, we released OpenFlow in hybrid form. [We've added] the ability to make OpenFlow forwarding policies supplementary to the forwarding policies that exist within the routing domain.
Another variant of SDN is network virtualization, the ability to apply logical networks or overlay networks on top of an existing physical network infrastructure. One of the benefits of that is to give the customer more freedom in terms of scalability and going beyond VLAN ID and MAC address table size limitations. In introducing logical networks, you're actually increasing the overall administrative overhead that the customer has to deal with. The beauty of a VCS fabric is that through automation and simplicity, you can reduce your administrative burden and your operational overhead in the physical network infrastructure, and invest more time in that logical overlay.
Read what Nolet had to say about data center fabrics and the future of FcoE.