Over the last couple of months, the enterprise computing space seems to have started a shift from private cloud paradigm into the hybrid cloud model. That makes sense, because the hybrid cloud allows companies to forgo the capital and the operating expenses associated with private clouds, in exchange for a pay-as-you-go model where you can just sit back and consume the business applications.
The hybrid cloud model is all about economies of scale. When managed service providers are operating the clouds for enterprises, they want the associated workflows to be automated. That facilitates an industrial process which allows the business applications to be provisioned and consumed immediately.
As far as IPv6 is concerned, it has a direct link to the hybrid cloud delivery model. Whenever a new application workload is being deployed in the hybrid cloud, the preparation of the virtual footprint requires release parameters — IP address, name and other network settings — that match with the specific VLANs used by the enterprise end-user for whom the cloud application is being set up. In practice, this makes manual assignments and technologies such as DHCP impractical, because neither has been designed for the elasticity of the hybrid cloud.
With this in mind, I think there are three reasons why wide-scale adaptation of the hybrid cloud will positively impact the IPv6 adaptation.
First, as enterprises migrate to hybrid clouds, they cease to be the bottle-neck for IPv6 adaptation. Insofar, the enterprises have been reluctant to move to IPv6 due to lack of business case. But once they hand over the keys to their business infrastructure and applications to managed service providers running the hybrid clouds, they no longer look after the network layer themselves.
Second, as managed service providers automate the release parameter provisioning process, the complex syntax of IPv6 addresses and vast size of the IPv6 space becomes irrelevant. When IP address allocation is carried out as part of an automated workflow, IPv6 is no more difficult to cope with than IPv4. The system doesn't care about the complexity of a syntax, so long as it is programmed correctly.
Third, the easier it becomes to consume applications, the more they are used. The more they are used, the cheaper they become, making them again that much easier to consume. And since all these cloud application workloads require their own IP address, the scalability and the economies of scale associated with the hybrid cloud business model will ultimately depend on whether or not there is sufficient address space to go around.
So the way things are looking to play out right now, I would put my money on the hybrid cloud as far as IPv6 adaptation in the enterprise is concerned. Assuming it takes off briskly, who knows how far along we will be in IPv6 adaptation by 2015.
|Cybersquatting||Policy & Regulation|
|DNS Security||Registry Services|
|IP Addressing||White Space|
Minds + Machines