Two reports reveal that most organizations have made little progress in optimizing their cloud infrastructure resource consumption. This gap presents a unique opportunity for managed service providers (MSPs) to address the growing demand for expertise in this area.
Insights into cloud resource management
A survey of 700 engineering leaders (350) and developers (350) from the U.S. and the United Kingdom (UK) conducted by Harness, a provider of a platform for building and deploying software, found that less than half have access to real-time insights into idle cloud resources (43 percent), unused or orphaned resources (39 percent), and over- or under-provisioned workloads (33 percent).
Additionally, more than half of the application developers (55 percent) said purchasing decisions are ultimately based on guesswork, and only 32 percent have fully automated practices to enforce cost savings. In fact, half of all respondents (52 percent) admitted there is a disconnect between their application development teams and FinOps functions within their organization.
The report estimates that approximately 21 percent of cloud infrastructure spending—about $44.5 billion—is being wasted. Much of this waste results from over-provisioning cloud resources and a general lack of proactive management. For instance, 71 percent of developers do not utilize spot orchestration, 61 percent fail to right size instances, 58 percent do not take advantage of reserved instances or savings plans, and 48 percent neglect to track and shut down idle resources.
The push for efficiency in the AI era
A separate analysis of more than 2,100 cloud-native application environments conducted by CAST artificial intelligence (AI), a platform provider for managing Kubernetes clusters, finds similar issues, with average CPU utilization rates of just 10 percent and memory utilization rates of only 23 percent. Overall, the report finds the over-provisioning gap between requested resources and the infrastructure used is 40 percent for CPUs and 57 percent for memory.
One of IT’s dirty secrets is that there has always been a high waste of IT infrastructure resources, even in on-premises IT environments. Software developers routinely overprovision infrastructure to ensure applications are available during peak load time. The issue is that peak load occurrences are often rare, causing much of the provisioned IT infrastructure to go unused.
Many MSPs have recognized this issue for some time. However, given the current economic uncertainty, the demand for organizations to reduce cloud costs by becoming more efficient has reached an all-time high. At the same time, cloud service providers in the age of AI are focusing on optimizing resource consumption. The growing demand for AI workloads on graphics processing units (GPUs) has also created pressure to free up space in data centers.
MSPs are uniquely equipped with the skills and expertise needed to address these dual challenges. The key opportunity now is ensuring that organizations seeking to optimize their cloud environments are aware of the value MSPs bring.
Photo: lzf / Shutterstock
