There's more than one type of cloud—and they don't run by themselves
My colleague Jeff James wrote about the feeling of doom and gloom hanging over IT pros regarding cloud computing, in contrast to the blue skies forecast for developers. I’ve thought about this a bit, and I’ve talked to different people about the topic—and I’ve come to the conclusion that the net change for IT pros will be very small. Why the change in perception? Maturity.
Specifically, maturity of the cloud computing discussion. When this new computing model began to intrude into our day-to-day consciousness, “cloud” was the only adjective in front of “computing,” and it described what we now know as Software as a Service (Saas), deployed in the public cloud: off-premises applications, running on shared resources that are low-cost and flexible but that you have less control over.
As companies began to seriously investigate the possibility of using the public cloud to host their data and applications, hard questions emerged for service providers: You have a “one size fits all” SLA? You don’t have adequate auditing and compliance controls? We can’t make an unattended visit to your data centers where our data is stored? We might not even know which data center it’s stored in? How do we get our virtual machines back out of the cloud? Many of the answers to these questions simply didn’t meet security or compliance requirements.
Enter the private cloud computing deployment model. The private cloud uses the same service models (IaaS, PaaS, SaaS) as the public cloud, but in the vast majority of instances, private clouds will be hosted on premises. (Private clouds can be hosted off premises, but the off-premises configuration—in particular, reliance on an external service provider to store your applications and data—has many of the same risks associated with the public cloud deployment model.) An on-premises private cloud can provide a company the benefits of a public cloud, though on a much smaller scale, while eliminating the risks associated with public clouds. As IT professionals and management become more familiar with cloud computing, the on-premises private cloud—not the public cloud—is stepping into the spotlight as the destination for most existing applications. And finally, a hybrid model (where, for example, workloads might usually remain on premises but burst into public cloud capacity for peak usage) will be the end state for most IT environments within the next decade. This demonstrates once again how the deployment pendulum swings back and forth in our business, before it ultimately settles into a reasonable median.
A new on-premises computing model comes at a cost, however. A huge difference between public and private clouds is that whereas public cloud service providers are well established (all it takes is a credit card to start using one), a company must build its own private cloud before it can start taking advantage of the cloud computing model. This means that instead of the “no capital expenditure” attraction of the public cloud, a private cloud is a major IT investment that will take time and money to complete.
And this is one of the major places IT pros can look for their career advancement.
On-premises private cloud means servers in a company’s data centers—and those servers need to be supported by people. Ultimately there’ll be fewer servers because virtualization and cloud computing make more efficient use of computing resources, but companies won’t be boarding up all their data centers anytime soon.
It’s not just the servers, though. What will really change in the corporate data center during the cloud computing era is how all the resources will be managed. Virtualization and virtual server provisioning are at the bottom layer of a cloud infrastructure; they’re the major components of infrastructure as a service (IaaS). But to have a real self-service private cloud, you must have layers of management, such as orchestration, workflow, and user interface. Some of these management layers will be available off the shelf, but all must be configured, and some must be custom-built because every company’s workflow is a little different. Almost nobody knows how to do this sort of work right now, so there will be a wide range of opportunities if you’re aggressive in taking advantage of the situation.
Another aspect of the whole “cloud computing killing IT pro jobs” scare is that it very much depends on your job role. One one hand, if a large portion of your job tasks can be commoditized or is a popular target for public cloud migration (such as email), then you should be looking to broaden your skills. On the other hand, if you’re a database administrator, logical databases must still be taken care of—regardless of whether they’re in your data center or someone else’s.
Finally, the impact will vary depending on the type of company you work for. Financial and health-related companies, with their strict regulatory and compliance environments, will be highly focused on the private cloud. Media companies, which don’t have the same kind of restrictions, will try to use resources in the public cloud as much as possible to keep the overhead of their content delivery low.
I believe the net impact of cloud computing will be far less than the doom and gloom originally predicted. But the impact of this new model will vary considerably depending on your job role and your company’s business. You must objectively look at where you and your current skills fit into this jigsaw puzzle—and take appropriate action.
Follow Sean on Twitter: @shorinsean
Subscribe to Sean's Windows IT Pro articles: RSS