This year more than last, in cloudy neighborhoods everywhere, SDN costumes are sure to be the most popular. Everyone wants to be SDN this year.
And for good reason… Who doesn’t want a more open cloud, a more dynamic cloud network, or a way to turn up networks as easily as they can spin up compute in support of their IT applications?
SDN has already shown massive potential for doing just that. At its root, SDN represents a major shift in operational mindset in which networks seamlessly follow applications to bridge the worlds of IP and IT in the cloud. That mindset is accompanied by a set of technologies that enable application-friendly abstraction of network capabilities and the policy-based automation of network instantiation.
Cloud service providers want to quickly deliver secure slices of rich networking capability to thousands of tenants across dozens of datacenters. In the same exact spirit, enterprise CIOs need to deliver those secure slices across departments with a myriad of applications and user communities within and outside their organizations.
They want to do it instantly, without surrendering control and visibility. They want to do it cost-effectively, by making the best and most use of their assets. That should include both their virtual and physical assets, across multiple data-center locations, preferably encompassing a variety of hypervisors. Said another way, it means reaching beyond network virtualization to cloud networking, in hybrid environments that are the reality of large-enterprise end-users.
Dreaming of the Hybrid Environment
In a true hybrid cloud scenario, an enterprise with its own data centers would leverage SDN to virtualize and automate its network. IT administrators would set up policies for each of their entities, departments, and user groups to grant appropriate permissions. Once devised, those policies could be readily templated and applied endless times as required. The IT admins would maintain full visibility and control of workload placement and movement within and across their data centers and would be free to operate in a heterogeneous environment with a mix of hypervisors, network equipment, and cloud management platforms.
That’s the value of hybrid environments and hybrid clouds. Heterogeneity of environments is key in private cloud deployments, where multi-hypervisor scenarios are increasingly desired as enterprises go beyond virtualization and toward the cloud.
Choice expands the mix of environments within the data center, improves economics, and increases flexibility of options for cloudbursting and disaster recovery. While doing so, enterprises ensure compliance by maintaining full control of assets placed outside their private data centers, and full choice of assets they elect to keep within their own walls.
True to the SDN mission, it’s all about rapid application delivery in an open environment, whereby SDN programmability and network automation would help unleash any application on any cloud, every time.
Sounds like a treat… So where’s the trick?
Well, the trick is that a real hybrid environment has not yet emerged.
A true hybrid cloud simply can’t force homogeneity of hypervisor or cloud-management environments. That remains a major restriction of early virtualization offerings and flies directly in the face of the premise of open clouds.
Real hybrid scenarios should span physical and virtual assets, in a multi-hypervisor, multi-site way. While many enterprises will continue to elect to build their own private clouds, the lack of a true hybrid hypervisor option limits their possibilities and stunts opportunity for cloud service providers.
Well, if open is good, and choice drives creativity, why would we accept this restriction? The answer is that we shouldn’t.
So as the weather turns cold and the leaves fall, be sure not to set your networks back an hour along with your clocks.