Cost Optimization in Multi-Cloud Era

Couple of years back when I discussed cloud with my customers, they used to ask me “Why cloud?”. Surprisingly in a short period of time the conversation has steered towards "How to onboard new cloud providers in a cost-effective manner?". So multi-cloud has become the norm and organizations have also realized that controlling costs has become much more complex. Some IT organizations are not equipped to handle management of Infrastructure OpEx at such a large scale. They are comfortable budgeting for the on-prem CapEx spend because if they get it wrong, they will just run out of capacity. No harm no foul. However in the cloud there is virtually unlimited infrastructure, so the month end billing report comes as a shock when they overshoot their budget. 

So, there are two main challenges that the CIOs face in the multi cloud era

1) how to reduce the cost of onboarding a new cloud hyperscaler 

2) how to standardize, control and optimize the cloud consumption costs

Reducing the cost to onboard new cloud hyperscaler

            Every time a new cloud provider is added to the enterprise landscape a significant effort goes into the landing zone setup and the cost of integrating it into the rest of the enterprise securely.  We already see leading public cloud providers like AWS taking proactive steps to streamline some of the landing zone setup by launching AWS Landing Zone and Account Vending Machine but there is a need for a comprehensive cloud foundation architecture. I recommend starting with a Open Cloud Foundation Framework which includes abstract blueprints for account structure, network architecture, application and data integration, and security controls. Also establish an Integrated Cloud Operations Framework so that the end users and the Site Reliability Engineers (SRE) have a seamless user experience. There is an inherent challenge of data sprawl and need for cloud data governance, which I will cover in a different post. 

Control and optimize the cloud costs

            The biggest challenge with cloud cost forecasting and optimization is that the SKUs are not standard across the cloud vendors and there is no way to compare costs. Moreover as the organizations start using the PaaS services the pricing model becomes more diverse and difficult to collate. There are other organizational challenges like the organization unit which consumes the cloud resources may be different from the one which pays the bill. So there is a need for a clear definition of cloud financial governance roles and responsibilities which spans across the enterprise. There are cost saving options linked to long term commitments like Reserved Instances or Committed Use Discounts which will significantly bring down costs, but the challenge is to identify the right volume to commit without losing the benefit of flexibility or innovation.

A mature Cloud Center of Excellence (CCoE) is the answer to these challenges, which includes people, processes and technology to rigorously manage the cloud financials. The cloud financial governance process should be proactive and continuous in choosing the right services. The governance process starts with a mechanism to budget and compare costs across cloud providers. Then come the guardrails to ensure the development team is using the best practices while consuming the cloud. The CCoE can incorporate the guardrails in the cloud platform reference architectures leveraging platforms like Tanzu or OpenShift which can abstract all the infrastructure services and make the application portable across cloud. The financial governance team then configures the integrated billing cockpit to monitor the billing trends and anomalies. There are constantly new innovation in cloud which make it faster, better and cheaper way to leverage the cloud. The CCoE should constantly look for optimal ways to leverage these new services and update the catalog, best practices and guardrails.

Summary

            The economic challenges in the multi-cloud era amplifies the importance of a Cloud Center of Excellence and an Open Cloud Foundation architecture. For brownfield implementations plan a staged transition starting with an assessment of the current cloud governance model and migrate to the target integrated cloud operating model and a Continuous Cost Governance framework to standardize, optimize and control costs.

About the Author

Mohan Udyavar, CTO Automotive Vertical, Wipro Limited

Mohan has 23 years of IT experience. Currently, he is responsible for providing consulting services to the Automotive clients in the areas of digital transformation, cloud adoption, application portfolio rationalization, platform architectures and architecture resilience. He is an AWS Certified Solution Architect Professional.