Previously, I had posted a blog about the results of PwC research about the changing design criteria for data center networking. Follow-on blogs detailed the insights gleaned from that research. Namely, security has risen to the top position in IT priorities for data center networking, and automation is firmly entrenched as the number two.
Based on these findings, it’s clear that IT is changing. But the question that remains is: how will all of these changes happen?
The move to cloud is imminent
We don’t need extensive research to know that the future for most enterprise companies is going to include at least some cloud workloads. The growth of well-known cloud titans globally, along with the emergence of a strong second-tier of more regional cloud providers, is a strong indicator of the demand for enterprise cloud.
But the PwC research does underscore the rate at which this transition is happening.
PwC’s findings suggest that virtually all mid-and large-sized enterprises expect to move some workloads to the cloud in the next 1-3 years. The broad consensus means we are beyond the evaluation of cloud as a viable IT model, and the timing means that companies have already started planning. Change is imminent.
All types of workloads
The other striking finding in the PwC research is that the move to cloud will not be restricted to just some types of workloads. IT leaders indicate that they expect workloads of all types to move to the cloud.
The above survey results suggest that workloads spanning customer service, marketing, business applications and beyond will all move to the cloud. In some cases, these transitions are easy to forecast. As companies consume more SaaS offerings, it seems obvious that cloud will be an integral part of the solution.
But companies conveying that they see at least 70% of their workloads migrating to the cloud over the next three years is surprisingly aggressive. While our industry tends to underestimate how long these sea change transitions take, it does indicate an unusually strong consensus. Minimally, this ought to be enough to make even the diehard skeptics stand up and re-evaluate their positions.
Not everything will be public
It’s worth pointing out that while virtually all types of workloads will move to the cloud, that migration doesn’t mean that public cloud will get everything. In their three year future, PwC respondents predict that close to 40% of total workloads will be in a private cloud.
That said, across all types of workloads, IT leaders projected basically flat three-year growth of private cloud workloads. This means that the three year cloud growth will come primarily from broader adoption of public cloud, an insight that is consistent with what we see happening with public cloud provider revenues.
Cloud and dual-vendor strategies
Most large enterprises have already taken advantage of dual-vendor strategies to gain economic leverage over their suppliers. As anyone in procurement will tell you, the greatest negotiating mechanism that a company has is the threat of competition.
As the industry speeds towards cloud, it’s important to recognize that the advantages of a dual-vendor strategy will translate from equipment to cloud providers. More succinctly, leveraging more than one public cloud provider for your enterprise work loads will create pricing leverage over suppliers. While we refer to cloud as a utility, it does not mean that the suppliers should not be subject to the same management strategies that have proven effective across all other parts of IT.
If the future will span on-premises, private cloud and one or more public clouds, then how enterprises design their data centers will need to change. Minimally, there needs to be a strategy for cross-domain management, as tools and processes will need to be applied in varying contexts. More likely, there will need to be careful consideration paid to how workloads are managed and orchestrated dynamically.
How do you monitor a data center that spans multiple domains? Do your architectural choices allow for uniform operations in different contexts? Do your device choices support in-cloud versions so that you can connect to different resource pools regardless of where they reside?
This will certainly drive steeper requirements into the underlying infrastructure as it has to support multicloud architectures. It could be, for instance, that an inflexible gateway choice limits the potential architectural solutions that are available. Or perhaps limited support for programmatic interfaces will prove to be an operational bottleneck preventing adoption of the most common and effective cloud tools.
Security and automation in a multicloud world
In addition to the basic operation of infrastructure, if IT is expected to make strides in both security and automation, those plans will need to include multicloud as the context for change.
For security, this means pulling threat intelligence from devices that reside across different domain boundaries. Architects need to account for data collection, central processing and multi-domain policy distribution and enforcement, all in real-time.
On the automation side, it means settling on a common set of telemetry, automation and NetDevOps tools that can extend across both physical and virtual devices regardless of where they reside. It also means stricter adherence to common interfaces so that the automation scaffolding doesn’t have to be unique to each environment.
As companies continue their march toward cloud, it is critical that they consider where the likely end-game will take them. Failing to prepare for the next step might leave enterprises stranded at a time when corporate velocity is the leading indicator of success. If the future is for the fast, the worst thing IT can do is go down a dead-end path.