Analytics company Pepperdata released a new study highlighting the unexpected costs that crop up when enterprises look to adopt cloud platforms.
The “Big Data Cloud Technology Report” online study was done in November and polled 750 senior enterprise IT professionals from a range of automotive, advertising, finance, and healthcare companies about the ways enterprises use the cloud to run big data applications and workloads.
For 2020, one in every three respondents said they had spent somewhere between 20% and 40% more than they budgeted, and 64% of respondents cited “cost management and containment” as their main worry when it comes to running cloud big data technologies and applications. One in 12 respondents said they ended up spending 40% more than they budgeted for cloud services.
A majority of the survey’s respondents told Pepperdata researchers that to “better optimize current cloud resources” was their main priority for big data cloud initiatives.
“This research shows us the importance of visibility into big data workloads. It also highlights the need for automated optimization as a means to control runaway costs,” said Ash Munshi, CEO of Pepperdata.
“Significantly more than half the companies surveyed indicate lack of control on cloud spend, making optimization and visibility the keys to getting costs under control.”
More than 34% of respondents said they planned to spend between $500,000 and $1 million in 2020 on big data analytics in the cloud, with 26.4% spending somewhere between $1 million and $2 million.
Almost 17% of respondents were spending less than $500,000 and 15.4% spent between $2 million and $10 million. Just 7.2% said they spent more than $10 million.
Less than 45% of respondents expected to stay on budget and 34.4% of respondents said they were planning to go over budget by somewhere between 20% and 40%. More than 8% said they expected to exceed their cloud budget by 40% or more.
The survey found that private clouds are still very popular, with nearly 47% of respondents opting for it over hybrid or public clouds, which both had less than 30%.
Outside of cost management and containment, respondents cited a number of other concerns they have about running big data applications in the cloud, like increased complexity (33%) and lack of control (13.4%).
In order to test workload and application performance, 29.8% of respondents said they use an application performance monitoring solution while 28.1% use cloud provider tools. More than 19% use a manual monitoring tool or a homegrown solution.
A majority of the survey respondents use ITOps and business units as well as line-of-business developers to manage the support and troubleshooting of cloud platforms, with 43.5% of IT leaders saying they used the shared support model. For 35% of respondents, support stayed solely with ITOps and 21.6% said support was managed by the development organization within business units.
When asked which types of applications or workloads consumed the most resources, 28.7% mentioned Hive while 26.9% cited Spark. More than 16% use MapReduce and 10.6% use Tez.
2020 saw an explosion in interest from enterprises in cloud platforms as organizations moved a significant amount of their systems online due to the coronavirus pandemic.
“Even with the best cloud strategy, the flexibility of the cloud can make managing resources more difficult,” Munshi added. “What’s needed is a system that will alert you when spend is excessive, while providing you with sufficient details so you can quickly understand the reasons for the cost overrun. Ideally, the system should also provide granular recommendations on fixing these issues.”