To refresh your memory, in part 1 of this blog, we discussed how the limited visibility of multi-cloud costs is a major contributor to this challenge, together with the first three steps in the journey to avoiding bill shock: quick wins, smart control and architecture optimization.
For CSPs, who transmit petabytes of data on a regular basis, the proactive governance approach to monitoring cloud costs is not enough. As cloud expenditures rise, they also need to revisit their commercial agreements or pricing models.
At the same time, the enormous magnitude of their data operations, combined with their direct access to nearly every business in the planet, makes CSPs attractive to do business with, putting them in a favorable position to negotiate prices with the leading cloud vendors.
With these factors in mind, let’s proceed to step 4 of our journey:
Matching pricing models to anticipated consumption: Leveraging Amdocs SmartOps FinOps management tools, customers should analyze historical data and run ‘what if’ scenarios on a continual basis to predict the future behavior of any application, as well as related costs. Sometimes, calculations and best practices will suggest that a typical, long-term commercial agreement (~3 years) with defined data consumption may not be the most beneficial model. This is where a more flexible approach might come into play. For example, dramatic cost reductions can be achieved by replacing on-demand cloud instances with reserved ones. In other words, it means purchasing instances up-front and committing to specific volumes for a defined period of time. Such a strategy has proved especially popular when combined with longer-period discount programs offered by most cloud vendors. Yet the ability to make such decisions requires the active engagement of product teams, who must examine usage patterns, identify seasonality differentials and calculate risks – all while striving to meet their predictions.
Indeed, we believe that adopting a healthy cloud “discipline” and cost optimization culture is an essential component of the journey towards becoming completely digital. One highly effective way we have achieved this at Amdocs is by combining the methodologies above with fun learning activities.
This brings us to the final step:
Step 5 – turning product teams into FinOps experts through gamification: By playing management games, which complement our FinOps management tools, teams win prizes and awards if they meet predefined levels of consumption beyond their initial commitment or expectations. The games focus on activities that were previously scattered across the different operations teams, but with the move to cloud and the adoption of DevOps, shifted to product teams. Examples include shutting down unused testing environments while not in use, detaching unnecessary cloud resources, right-sizing architecture for non-production environments and so on. If teams and individuals exceed their designated resources, they lose points and ultimately, miss out on the perks awarded to their more successful peers.
The results speak for themselves. We found that these games increased our product and development teams’ motivation to learn about cloud tech and applications, while engaging them to help achieve the organization’s business goals.
Stay tuned for more articles on the challenges and best practices of CSP cloud operations.