Are you an enterprise IT professional? Do you agree? Join the discussion today!
A report published Wednesday that includes survey responses from over 740 senior IT professionals globally suggests that the majority of IT professionals believe typical service level agreements (SLAs) built around availability and performance fail to address the risks of moving and managing applications in the cloud.
The survey, commissioned by Compuware and carried out by Research In Action, found 79 per cent of survey respondents believed their SLAs were “too simplistic” and failed to address key risks associated with moving and managing cloud apps.
“Entrusting mission critical business applications that drive revenue and critical business processes require ultimate trust and accountability in a cloud provider,” said Michael Masterson, director of cloud solutions for Compuware APM’s business unit. “
Vanity metrics like simple uptime do not capture well-known issues such as ‘noisy neighbours,’ which can be detrimental to traditional enterprise apps that were not designed to scale and fail horizontally.”
The report confirms that there is still a great deal of fear, uncertainty and doubt among enterprise IT professionals when it comes to cloud.
Nearly three quarters (73 per cent) said they believe their cloud providers could be hiding problems at an infrastructure or platform level that impact on the performance of applications. 62 per cent said limited visibility into the infrastructure made it difficult to both monitor their application performance and troubleshoot problems in the cloud.
“The truth behind performance is what the end customer or user feels – that is all that really matters,” Masterson told Business Cloud News. “The vendor should be responsive, but the general rule seems to be innocent until proven guilty, with fingers usually pointing back to the application or database vs the infrastructure.”
“Customers that encounter this often discover it on their own, then blame the provider for poor performance, when the reality is that it’s another tenant over consuming shared resources.”
Masterson explained that the results suggest an underlying problem with the way cloud service providers often make guarantees on their services.
“There are two primary issues with “technical guarantees”. First, if everyone tries to grab maximum rates at once. It’s like a run on the bank and very quickly you’ll discover that the provider (the bank) doesn’t actually have all that reserve bandwidth (the cash) on hand. Instead, they’ve made assumptions and provisioned based on estimated usage and normalised workloads – thereby offering you capacity at a discount compared to dedicated hardware,” he said.
But “users are not typically normally distributed – they have spikes and troughs in demand. Software is inherently virtual and should be able to match the demand with appropriate supply – much more so than tradition fixed goods and complex logistics and fulfilment systems. Yet they run into the same bottlenecks, often with catastrophic failure,” he added.