The data centre is heading into an era of “fabrics.” When most IT professionals hear about “fabrics” they either think of proprietary technologies, like QFabric, or perhaps of Fibre channel which is a storage technology. Cisco, Juniper, and Brocade have really taken to the term “fabric.” While there are certainly some new developments here – TRILL (transparent interconnection of lots of links) and DCB (data centre bridging) among them – what a fabric really means is a flatter Ethernet network, with greater link resiliency. A fabric network makes building a private cloud easier because it’s built for east-west traffic flows that are caused by virtual machines moving from physical server to physical server in search of more/less resources, depending on the automation and orchestration configuration of your private cloud.

So instead of thinking about “fabric,” really you can picture a flatter network – two tiers instead of three – that don’t rely on STP, where routing is technically being done a Layer 2, and routing can be based on location-independent L2 MAC addresses.

I recently attended a vendor-agnostic session called “Ethernet Fabrics 101” put on by Brocade in Toronto. It was really refreshing to see a vendor present basic facts about new technology with barely any vendor spin on it. Kudos to Brocade for that, and I’m looking forward to attending their next session.

Anyway, Brocade broke its definition for an Ethernet fabric into five characteristics:

• Flatter Architecture
• Distributed Intelligence
• Scalability
• Efficient
• Simplified Management

I’m going to spend the next few blog entries exploring those things, as well as TRILL, DCB, and where data centre networks are headed. 2012 is a pivotal year, but I suspect things will really take off in 2013 and 2014 as prices for 40GbE comes down, and even 100GbE for really demanding data centres (Apple, Microsoft, Google, Facebook, Twitter, and any cloud IaaS providers).

Virtual local area networks (VLANs) were invented over 20 years ago, and so far VLANs have given IT a significant capability in segmenting networks and traffic, letting networks run faster and more efficiently. However, the advent of cloud computing, both public and private, along with cloud bursting, has broken the traditional network and made network management far more complex. What’s needed in the enterprise now is a platform-agnostic VLAN that can span across geographic and logical limits to rein in the infrastructure your workloads are running on. Here I’m going to focus on emerging network virtualization solutions that promise to give you, the network professional, a unified view of your network and servers.

Where Has My Server Gone?

VLANs were first introduced to the networking world in 1988. They have a come a long way since then, with major revisions as recently as 2007. VLANs are not a perfect solution though, as you cannot extend them beyond the private, corporate network. They are limited to physically connected networks, making them unsuitable for today’s shifting IT environment.

Cloud computing has broken the traditional network boundaries, and so a new solution is needed. VMware and Cisco have a proprietary solution that uses Cisco’s Nexus 1000V virtual switch integrated with vSphere, but this solution doesn’t integrate with hypervisors from other vendors.

Instead, network virtualization gives you the ability to create VLANs across geographically and logically disparate networks. Each virtual network (vNetwork) can be isolated and managed separately from other vNetworks, giving IT the power to create vNetworks based on applications that extend from on-premise infrastructure to public cloud infrastructure.

Any node, in any location, can be registered as part of a vNetwork, whether it’s a physical server, virtual server, on-premise, or in a public Infrastructure-as-a-Service (IaaS) cloud. Each node receives virtual MAC and IP addresses unique to the vNetwork, allowing data to flow solely through the vNetwork.

This is different from a VLAN, which traditionally required configuring ports on your physical switch to make the connected device part of the VLAN. vNetwork switches are pure software implementations of network switches, providing more flexibility over the configuration of the network, with the ability to bring nodes from public IaaS clouds (e.g. Amazon EC2) and servers rented from hosting providers (e.g. Rackspace) all into a single vNetwork.

vNetworks Sound Great on Paper

If you have a network at all resembling the examples shown on vCider’s website (a startup in the network virtualization space), vNetworks will provide at least two key benefits to network management:

1. Flexibility. Organizations are looking to optimize costs and operational efficiency by moving workloads to the least expensive platform and to improve service availability by building application delivery networks and high availability configurations. vNetwork software makes this possible.
2. Platform agnosticism. Existing solutions from Cisco and VMware are proprietary and only work in Cisco networks supporting VMware environments. However, a new breed of vNetwork solutions from vCider, Nicira, and Big Switch Networks will provide you with the ability to manage virtual machines running on multiple hypervisors or cloud servers from multiple providers, all on a network that could be built on HP, Dell, Cisco, and Aruba networking gear.

This is Just the Beginning

At the moment, these solutions are working at layer 3 (the network layer) of the OSI stack. That is a good and natural place to be, given the path selection and internetworking capabilities embedded at that layer. Ideally, virtual switches would also operate at layer 4, providing great control over TCP and UDP connections, fostering end-to-end connectivity as well.

In the meantime, vCider, Nicira, and Big Switch are all very young companies with, for the most part, unproven solutions. Only vCider’s solution has launched and is commercially available, though both vCider and Big Switch showcased their products at GigaOM Structure 2011.

Initially, these products will support the most well-known cloud IaaS and networking vendors – Amazon, Rackspace, Cisco – and then move to other vendors with less market share. If you didn’t jump on the Amazon and Rackspace bandwagon, instead waiting for a cloud IaaS provider that would better suit your needs, you’ll have to wait a bit longer for a vNetwork solution.

There are really two key takeaways network professionals need to take from this:

1. Network virtualization will solve problems to do with cloud bursting, workload optimization, and connecting applications running on disparate networks. This will make it easier for businesses to manage those applications, regardless of where the instance is running.
2. Paired with cloud orchestration and automation software, vNetworks put us one step closer to being able to dynamically, and automatically, shift workloads from one infrastructure to another while being able to maintain a persistent network configuration.

The bottom line is that virtual networks, or vNetworks, are coming on the scene to solve a problem: managing public cloud servers as a disparate network. When you can wrap your public cloud servers, rented dedicated servers, and on-premise servers into a single vNetwork because they all belong to the same application or department, the promise of workload optimization and dynamic reallocation gets one step closer to reality.