As a result, enterprise IT teams can finally automate network provisioning and better integrate network resources into overall IT operations. Yet it can still be difficult to determine how to evaluate these new products and where to begin with implementation. The good news is that there are ways of using network virtualization in hybrid scenarios that don't require a complete rip-and-replace.
Why bother with network virtualization technology? Business drivers
Operational efficiency is the major driver for network virtualization. While storage and server resources can be virtualized and automatically provisioned, the network has lagged behind. In today's network, the provisioning process is complicated and slow.
In order to make even a simple change, networking teams currently go through a lengthy process that includes planning the change and then submitting it for review through a "change control" system. That system requires other people to consider and approve (or disapprove) of the proposal. Then the change is executed by admins or engineers, and finally it must be verified to see if it is working as expected. This process is woefully inefficient and is more likely to lead to human error. It can also take days or even weeks to complete a change.
Network virtualization, however, enables the automated provisioning of network services like virtual LAN (VLAN) creation, Ethernet port provisioning, load balancing and firewall policy maintenance. With automation comes a reduction in provisioning time, as well as a reduction in the potential for human error.
More network virtualization resources
Tackling network functions virtualization challenges
Can NVGRE and VXLAN solve your network challenges?
Network virtualization technology also has the potential to reduce network operational expenses and capital expenditure as well. With network virtualization, IT teams can make the most of their existing physical hardware because users can build multiple distinct environments on top of one physical infrastructure. Similar to the way service providers separate customer traffic across a common MPLS infrastructure, enterprises can use network virtualization to build virtual networks, each of which contains its own rich set of services, including routing and switching, plus Layer 4-7 services like firewalling and load balancing.
Going forward, network virtualization has the potential to reduce the cost of the underlying network hardware itself. The idea is that the underlay network can be simplified to be that of a high-bandwidth Ethernet fabric that is capable of ultrafast switching and routing, but doesn't require rich functionality beyond that. That richer functionality would be moved into the software layer at the virtualized network edge. In that case, the underlay network might not have to wear the badge of a premium network vendor. In fact, several startups are proposing alternatives, espousing a model of low-latency, high throughput, inexpensive white-box switches that focus on fast transport but leave the complex Layer 4-7 forwarding decisions to the virtualized network.
Your existing technology may be ready for network virtualization
For those convinced that network virtualization would benefit their organization, the next step is perhaps the hardest: How does an organization begin moving towards virtualizing its network?
It’s key to first understand that network virtualization is not an all-or-nothing proposition. While ultimately it has the potential to displace traditional technology, network virtualization can be introduced alongside an organization's existing infrastructure. There are a number of ways to implement this hybrid scenario.
Organizations can start by investigating their existing network virtualization capabilities. Firewall and load-balancing vendors often offer a virtualized form of their hardware appliance. However, it is a mistake to think of simple virtualized appliances as actual network virtualization. In a true network virtualization environment, Layer 4-7 functions can be spun up at will, isolated to a specific virtual network and be part of a service chain.
Beyond that, though, network teams should realize that their existing networks are able to forward overlay traffic, such as what runs in Virtual Extensible VLAN tunnels. In addition, they could query their networking vendors about their virtualization offerings that are specifically aimed at hybrid scenarios. For example, network switches already in production might be OpenFlow-capable, making it possible to run both SDN and traditional traffic on the same infrastructure at once.
In part two of this series, learn four key factors to consider for hybrid network virtualization implementation.
This was first published in October 2013