While economic uncertainty and recession are topics of the day, and an aura of gloom surrounds businesses, cloud technology is at an all-time high. Gartner predicts that cloud spending in 2023 will not stop and could reach $592 billion – 21 percent more than $490 billion in 2022.
Even as the cloud business continues to balloon, the ground reality for CXOs is sobering. The predicament for CXOs would lie in cloud cost optimisation and identifying measurable ROI on cloud spending.
Echoing this sentiment, TechTarget’s article `8 top CIO trends and priorities for 2023’ says that CXOs must do a balancing act of adding new services to gain a competitive advantage even as they continue trying to cut costs. It is becoming clearly evident that in 2023 cloud cost optimisation will remain a trending buzzword.
For CXOs, if the cloud empowers technology advancement but with measurable ROI, they have already begun delivering on their promises.
As per Gartner, the top reasons why budgets do not stay in check are:
- Ungoverned costs
- Unanticipated usage
- Long-term commitments
- No production sprawl
- Overestimation of the production environment
- Traditional design and implementation decisions
Despite the common notion that a cost management tool will solve all the cost problems in cloud. It is also important to consider the following areas that could save cost for the organizations:
- Governance
- Architecture refinement
Governance
For cloud computing, operational governance enhances security, enables smooth cloud operations, and also manages risks. This helps your business state policies as business processes and enforces these policies throughout your business.
The three key governance aspects are defining, implementing, and monitoring controls. The following are among the points to consider for effective governance:
Self-Service:
It enables end users to request cloud resources conveniently with adequate controls and visibility for the organization. This will also allow employees to access required cloud systems within the organization’s compliance and budget constraints.
Standardizing Resources:
It involves understanding organization-wide usage and defining standard ‘T-shirt’ sizes for most used resources such as virtual machines, databases, AKS clusters, etc. Where in this will also allow us to manage costs better by effectively applying the reservations without hampering productivity. It ensures that all the relevant resource-sizing options are available to the project teams; however, in exceptional cases, any other nonstandard resource size can also be considered to ensure business continuity.
Defining Resource-Tagging Strategy:
Tracking resources is essential to understanding cloud cost optimisation and architecture review. The correct tagging process for help in the cloud, identification of tag applications at the required levels, and automated deployment of tags are crucial to ensure that tag data is coherent.
Reservations:
Purchasing reservations for the pre-defined T-shirt sizes and applying them at the correct levels (tenant/subscription/RG) to ensure the practical and intended usage of reservations for cloud resources. Reservations are time bound, and breaking them may cause additional overhead. Hence, the reservation duration should be a key aspect during the purchase.
Orchestration:
Cloud orchestration optimizes and simplifies the process by removing repetitive steps and eliminating redundancy. Orchestration also helps automate IT processes, tasks, events, provisioning, etc.
Cloud Cost Reporting:
Cloud costing could be used for show-back or charge-back to individual BUs in the organization, visualization of cloud consumption costs collected via APIs, and tagging data through PowerBI or any other cost management tool to keep track of the ongoing usage and consumption trends, along with anomaly identification.
Hybrid Benefits/License Portability:
Ensuring that the currently procured on-premises license is usable on the cloud and identifying opportunities to use any unused/spare on-premises license. All cloud Hyperscalers provide a pay-as-you-go licensing model suitable for most resources with a short- to medium-term life. Still, in the long term, procurement of external on-premises licenses separately and applying them on the cloud has generally been a more cost-effective option. However, this depends entirely on Hyperscalers and existing OEM licensing policies, so a detailed analysis should be done to make such a change. Typical licenses considered for hybrid benefits are OS, databases, applications, etc.
Architecture Refinement
Architecture changes are disruptive and may impact the application functionality and end users. However, the following are among the opportunities that should be considered for better control of cloud architecture:
- Landing Zone: Periodic review of landing zone architecture will help identify orphan services and consolidate network and security services. This will also help visualize how cloud resources are utilized in the enterprise and create a long-/short-term strategy.
- Application Architecture: It refers to re-architecting or re-forming the application to a different platform in the same or new Hyperscaler. Such activities are strategic and also align with the long-term application roadmap. What more, a cost-benefit analysis should be done before such a migration.
- Compute Optimization: We need to benchmark computer-resource utilization and plan to scale down existing add configuration to a lower state to save on underutilized services costs. This must be designed well with application and business stakeholders, potentially leading to application performance issues.
- Network Optimization: Optimizing network design and network ingestion points to optimize network costs.
- Storage Optimization: Optimizing storage costs involves identifying external/internal storage types, configuration, access rules, and ingestion points.
To summarize, with governance and architecture refinement, there will be measurable returns on investment along with a visible and ongoing cost advantage.
Authored by – Sripad K B
