Author(s): Mark Hodapp, Posted on May 30th, 2012
Looking back at the Renaissance, imagine that you are Michelangelo’s parent. The 19-year-old returns from his sojourn in the Medici court. Your own banking efforts have faltered. With the ruling Medici expelled out of Florence due to the rise of Savonarola, Michelangelo is still eager to pursue sculpting and the study of anatomy. However, the question remains whether he should return to the safe world of the family quarry and farm? Or, would you as his parent advise him to continue pursuing art assuming the Medici-led renaissance will continue to create resource abundance to fund Michelangelo’s work? At that turbulent moment in time, without the knowledge of Michelangelo’s future works or the Medici’s return to Florence, a parent is not be faulted for guiding the youth to the ‘safe’ alternative.
Today, we are facing such similar decisions within our IT shops. The rise of resource abundance and virtualization has given us options to better utilize the available physical resources. We’ve come to realize that the operational management of virtual resources has added to our operations teams’ workloads because, they have to manage both the physical resources and virtual resources while continuing to operate with reduced budgets. This has led to the emergence of Cloud computing.
As we begin to explore the operational aspects of ‘Clouds’, we discover that Cloud stacks employ automation mechanisms to help the deployment and management of applications and virtual resources. We see that our current human-language-based run books and human-augmented processes have to be radically changed to do what Clouds can achieve; and we’re challenged to find knowledgeable people to lead the effort and manage the organizational change. We are leery of placing big bets on Cloud when new products enter the market or when small companies are acquired by larger companies. And there still remains more Cloud hype than fact in the marketplace.
There are several challenges to overcome as we envision our future IT departments ascending to higher echelons of artistry. It is important to note that Michelangelo wasn’t immediately world renowned, he too had to study under masters and seek studies of new subjects. IT departments need the same mentoring to understand what’s different about automation and how they scale up cloud workloads. Some past decisions, skills, and partnerships will need to be reconsidered.
One example for reconsideration is how human-centric operation processes change radically after automation. Prior to automation, humans could enter the intervention process to resolve unforeseen issues. For example, they could examine SNMP events or other systems management statistics. They were, by design, the central point for monitoring and trouble shooting. In contrast, automation is machine centric and machine augmented. Having technicians design and augment the automation code that runs on machines, as well as intervene and remedy problems will have to change. Interventions will now have to follow an automated course to resolution. As better versions are designed or implemented, the automation code, like application code, will have to be revised and redeployed. The revised versions require careful management and deployment to achieve iterative improvement.
Human supervision has to move from monitoring applications to overseeing both the applications and their automated management. Moving human-centric operations to the realm of automation allows administrators to manage operations at a higher control plane, such as measuring and verifying service levels and, delivering higher quality to the actual customers.
Even with times of uncertainty, we can see opportunity ahead. Seeking enlightened advice and skilled practitioners in the art is an appropriate step forward and can help us plan for our future success.