XaaS — Everything as a Service: IT Operations
The new value of IT will now be given by its ability to select, assemble and operate technology provided by third parties, it assumes, of course, a commitment to operational continuity, something essential for the organization to conduct its -now digitalized- businesses.
From all IT areas affected by the adoption of a XaaS model, no doubt IT operations will have the greatest challenges of unlearning to learn again. Used to manage the operational details, from the voltage level needed for the data center, the updating the platforms, to monitoring alert consoles on resource consumption on servers and mainframe computers, through the promotion of changes between environments, they will now have to focus their activities -always with the same objective of maintaining a high productive capacity and its continuity- but developing a more analytical model for management, with more dynamic but also a preventive management of the operation.
Change and Operations Management areas, and other support units to the solutions development process, must adapt their practices to this model in which the speed of change achieved is much greater, promoting a trend towards self-service within the development teams, but taking advantage of an environment where there are many tools that help to govern these processes. In sum, going from “police” type task to become a dynamic promoter of good practices and planning.
Undoubtedly, resource allocation mechanisms are something that IT will have to modify. From a scarcity model, -whether for the very high costs of accessing on-premise licenses and infrastructure or for the long term of traditional implementation-, we will move on to a model of “infinite” technological resources. The adoption process must be rational, we cannot “lose our mind”, implementing technology that may have very low entry costs (and barriers), but that in the long run, after expanding service to massive customers, can mean operational risks or financial unpayable costs.
IT operations areas must leverage the capacity of the cloud to respond real demand. Traditional provisioning models with manual and hierarchical approval flow from finance areas, infrastructure managers, change management areas, and other units clearly anachronistic. On the contrary, IT will have to adopt more dynamic processes, real-time scheduling, and dispatching, more similar to those top companies in manufacturing, with its sophisticated and automated MRP systems.
This search for efficiency in the allocation of resources should lead us to control the computing capacity needed to perform each task. For example, for batch processes, while a smaller server can cost less per minute of processing, the allocation of a higher-capacity server may reduce the process times to run that batch, so the final cost will be less with the higher capacity equipment, which can be returned after use, with the added benefit that we have served our user in a shorter time.
The technology upgrade rate is increased with cloud computing models because we do not have to deal with the entire financial justification process of technology upgrade projects because the cost of using the new versions of the services is usually considered as part of the recurring payments we already made to the provider.
This new speed will also require automating most possible processes related to configuration management and promotion between environments. While changes create risks, ease of access to resources (and being able to return them at will later), should allow us to define new mechanisms to reduce their impact. For example, using progressive deployment techniques that minimize the possibility of service disruption, such as “Blue/Green”, “Canary releases” or “A/B testing” deployments.
This same dynamism of acquisition -and disposal- can also be used, for example, in development and testing environments, returning resources that are not used outside office hours and “buying” them again at the beginning of the working day, or during complex tests that requires a temporary growth in capacity. In this sense, the production monitoring of the applications and their platforms will not only aim to ensure operational continuity but also help to decide on the just-in-time allocation of resources to the productive capacity of the different lines of business and distribution channels.