(Originally posted in the In-Q-Tel Quarterly)
The cloud-enabled enterprise fundamentally changes how personnel interact with IT. Users are more effective and efficient when they are granted on-demand access to resources, but these changes also alter the technical skill-sets that IT organizations need to effectively support, maintain, and advance their offerings to end users. Often, these changes are not always immediately obvious. Automation may be the linchpin of cloud computing, but the IT staff’s ability to effectively implement and manage a cloud-enabled enterprise is critical to the IT organization’s success and relevance. Compounding the difficulties, all of the existing legacy IT systems rarely just “go away” overnight, and many workloads, such as large databases, either don’t cleanly map to cloud-provided infrastructure, or would be cost-prohibitive when deployed there. The co-existence of legacy infrastructure, traditional IT operations, and cloud-enabled ecosystems create a complicated dance that seasoned IT leadership and technical implementers alike must learn to effectively navigate.
In the past five or so years, and as enterprise IT organizations have considered adopting cloud technologies, I’ve seen dozens of IT organizations fall into the trap of believing that increased automation will enable them to reduce staff. In my experience, however, staff reductions rarely happen. IT organizations that approach cloud-enabled IT as a mechanism to reduce staffing are often surprised to find that these changes do not actually reduce complexity in the environment, but instead merely shift complexity from the operations to the applications team. For instance, deploying an existing application to Amazon Web Services (AWS) will not make it highly available. Instead of IT administrators using on-premises software tools with reliable access—and high speed, low-latency network and storage interconnects—these administrators must now master concepts such as regions, availability zones, and the use of elastic load balancers. Also, applications often need to be modified or completely re-designed to increase fault tolerance levels. The result is that deployments are still relatively complex, but they often require different skillsets than a traditional IT administrator is likely to have.
A dramatic shift in complexity is one of the reasons why retraining is important for existing IT organizations. Governance is another common focus area that experiences significant capability gains as a result of cloud-enabled infrastructure. Automation ensures that every provisioned resource successfully completes each and every lifecycle management step 100% of the time. This revelation will be new to both IT operations and end users. I’ve also frequently seen components of the IT governance mechanism totally break down due to end user revolt—largely because particularly onerous processes could be skipped by the administrators as they manually provisioned resources.
Cloud-based compute resources will dramatically change the computing landscape in nearly any organization I’ve dealt with. For example, one IT Director worked to automate his entire provisioning and lifecycle management process, which resulted in freeing up close to three FTE’s (Full Time Equivalent) worth of team time. Automating their processes and offering end users on-demand access to resources helped their internal customers, but it also generated substantial time savings for that team. The IT director also recognized what many miss: the cloud offerings may shift complexity in the stack, but ultimately all of those fancy cloud instances are really just Windows and Linux systems. Instances that still require traditional care and feeding from IT. Tasks such as Active Directory administration, patch management, vulnerability assessment, and configuration management don’t go away.
Another common learned-lesson that I have witnessed is that with shifting complexity comes dependence on new skills in the availability and monitoring realms. Lacking access to physical hardware, storage, and network infrastructure does not remove them as potential problem areas. As a result, I have seen organizations too slowly realize that applications need to be more tolerant of failures than they were under previous operating models. Making applications more resilient requires different skills that traditional IT teams need to learn and engrain in order to effectively grow into a cloud-enabled world. Additionally, when developers and quality assurance teams have real-time access to needed resources, they also tend to speed up their releases, placing an increased demand on the workforce components responsible for tasks such as release engineering, release planning, and possibly even marketing, etc.
I’ve encountered few customers that have environments well suited for a complete migration to the public cloud. While a modern-day IT organization needs to prepare for the inevitability of running workloads in the public or community clouds, they must also prepare for the continued offering of private cloud services and legacy infrastructures. Analyst firms such as Gartner suggest that the appropriate path forward for IT orgs is to become a broker/provider of services. The subtext of that statement is that IT teams must remain in full control over who can deploy what, and where. IT organizations must control which apps can be deployed to a cloud, and which clouds are acceptable based on security, cost, capability, etc. Future IT teams should be presenting users with a choice of applications or services based on that user’s role, and the IT team gets to worry about the most appropriate deployment environment. When this future materializes, these are all new skills IT departments will need to master. Today, analyzing cloud deployment choices and recommending the approaches that should be made available are areas that typically fall outside the skillsets of many IT administrators. Unfortunately, these are precisely the skills that are needed, but I’ve witnessed many IT organizations overlook them.
The Way Ahead
While IT staff can save significant time when the entirety of provisioning and lifecycle management is automated, there are still many needs elsewhere in the IT organization. The successful approaches I’ve seen IT organizations use all involve refocusing staff to value-added tasks. When IT administrators are able to spend time on interesting problems rather than performing near-constant and routine provisioning and maintenance, they are often more involved, fulfilled, and frequently produce innovative solutions that save organizations money. Changing skillsets and requirements will also have a likely affect on existing contracts for organizations with heavily outsourced staffing.
Governance is another important area where changes in the status quo can lead to additional benefits. For example, manually provisioned and managed environments that also have manual centralized governance processes and procedures typically have significant variance in what is actually deployed vs. what the process says should have been deployed: i.e. processes are rarely followed as closely as necessary. No matter how good the management systems, without automation and assignment, problems like Virtual Machine “sprawl” quickly become rampant. I’ve also seen scenarios where end users revolt because they were finally subjected to policies that had been in place for a while, but were routinely skipped by administrators manually provisioning systems. Implementing automation means being prepared to retool some of the more onerous policies as needed, but even with retooled processes, automated provisioning and management provides for a higher assurance level than is possible with manual processes.
Automation in IT environments is nothing new. However, today’s IT organizations can no longer solely rely on the traditional operational way of doing things. Effective leadership of IT staff is critical to the organization’s ability to successfully transition from a traditional provider of in-house IT to an agile broker/provider of resources and services. Understanding the cloud impacts much more than just technology is a great place to start. This doesn’t mean that organizations that are currently implementing cloud-enabling solutions need to jam on the brakes, just realize that the cloud is not a magic cure-all for staffing issues. Organizations need to evaluate the potential impact of shifting complexity to other teams, and generally plan for disruption. Just as you would with any large-scale enterprise technology implementation, ensuring that IT staff has the appropriate skills necessary to successfully implement and maintain the desired end state will go a long way to ensuring your success.