The is an authored article by Swaroop Rao, SVP & Head Digital Engineering Practice, that appeared in IndiaTechOnline on 19th August 2021. Follow this link to view the original article.
Companies in India are increasingly adopting the cloud and moving their applications to a single or hybrid cloud platform.
According to an IDC report, the public cloud services market in India touched US$ 1.6 billion for the first half of 2020 and is estimated to reach US$ 7.4 billion by 2024 with a CAGR of 22.2 percent. While this would contribute to the overall efficiency of the business, unplanned costs and other complicated bills for cloud may create issues. In the absence of any clear best practices to optimize cloud-related costs, it becomes a challenge for companies to budget and forecast for cloud operations. Monitoring and controlling the expenditure has to be an ongoing task rather than a one-time process, considering the dynamic aspect of cloud.
To optimize the cost for cloud operations, it is important to assess the spend at the very start to check for anything that falls above or below the allocated budget. Advanced analytics can help in visualizing the cost spend and optimizing the processes to increase operational efficiency. Automation can help organizations to implement cost reduction recommendations across cloud operations. Another important part of cost optimization is a robust reporting mechanism that provides a clear and easy understanding of the operations.
In order to identify the right opportunities and arrive at the right recommendations to optimize the cost related to cloud operations, here are few steps that can be followed.
Accurate assessment and understanding of cloud-related billing is the first and most important step. There needs to be complete visibility into the cost including details of the expenditure, billing of services and resulting business outcome. It has been observed that around 20 percent of the total cloud spend is on underutilized resources. This loss can be significantly reduced by assessing which projects cost the most.
Companies often maintain a large number of stock keeping units (SKUs) but in the absence of correct information about who is opting for what services and why, it becomes cumbersome to deduce the total cost of ownership (TCO) for the cloud applications and services. In such a situation, it is imperative to assess and determine how to set up the budgets, billing reports, comprehensive dashboards customized for detailed cost views and identify wasteful payouts.
A cost visibility model can be established to gain an accurate view of the costs. Once the pain points are identified, a strategy can be formulated along with the best practices to follow.
Post assessment, it is necessary to implement the best practices to optimize the cloud operations cost. Unpredictable costs can be reduced through customized dashboards, reorganized resource hierarchy and billing setups along with optimizing the costs related to network and storage. Implementing automation - including machine start/stop schedule, optimizing virtual machines (VMs), use of pre-emptible VMs for big data workloads and clean-up of compute instances - can help in reducing the development and quality assurance (QA) costs by as much as 50 percent.
A practice should be created to identify the usefulness of an object to the business. Storage classes can be used for best-fit use cases, setting up lifecycle policies, deduplicating the datasets and object versioning – contributing to the overall goal of cost optimization.
Along with this, it is important to optimize data warehousing costs. This can be done using tools like Google BigQuery, that can run quick queries and provide real-time insights with streaming data. Advanced and predictive analytics can further help with built-in machine learning (ML) capabilities.
Another critical spend area is the network that every cloud deployment requires to move the data. The network usage has a direct impact on the performance, availability, security and cost of the deployment. Hence, a thorough review of the networking spend strategy should be done to understand the usage and traffic details of the cloud environment.
Cost optimization is not a one-time activity, but an ongoing process, with iterative tactics to achieve the desired result. It requires deep collaboration across application and operations teams to balance the costs with performance, availability and security.
A recent Nutanix Enterprise Cloud Report showed that 97 percent of respondents in India felt hybrid cloud was the right architecture for their organizations, 10 percent higher than the global average. Given the estimated steep increase in cloud adoption in near future, it is imperative for organizations to take a serious look at how to maintain the cost at the optimum levels. Conducting regular reviews and operationalization once cloud implementation is done, forms the key to the whole cost optimization strategy.