Cloud economics: what are the real costs of cloud?
Looking to place your company’s workloads into the cloud? Just how much do you expect to pay for that? What if you want to ramp up capacity, or reduce it? Do you know what this will cost in advance?
While cloud providers often loudly proclaim the low cost of entry to their offerings and the convenience of the pay-as-you-go model of computing, is that really the full picture when it comes to costs? Not so, according to John Abel, senior business director for Oracle.
Abel said that while the pay-as-you-go model has many advantages, it does not communicate the full story.
“It’s a little like how mobile phone usage is priced, you have pay-as-you-go versus contracts. It’s well known that if you have small consumable usage then pay-as-you-go is completely adequate. You can control the spend and renew it when you want,” he said.
“But if you’re going to equip hundreds of employees with mobile phones, you’d want to move to phone contracts pretty quickly, and essentially the same is true of how cloud services are priced.”
Abel offers the model of ‘dev ops, low ops and no ops’ to describe how to best work out what cloud model suits an individual company. Hidden costs can escalate as companies climb the stack.
“This loosely maps onto Infrastructure-as-a-service (IaaS), platform-as-a-service (PaaS) and software-as-a-service (SaaS). As you move up the stack, a lot of the costs you would be incurring are being incurred by the cloud vendor,” he said.
“When cloud started out, companies realised that when it came to IaaS, it was so cheap they didn’t have to have a data centre. But most cloud vendors have a pay-as-you-go model which means that as you scale, network costs become quite difficult to manage if you’re not careful. Essentially there can be hidden costs to this model.”
“For example, if you deploy services in one data centre and then you use them in a second data centre as well, then you will pay for network costs across the two data centres. That’s not always transparent,” said Abel.
If a company moves to PaaS for its operations, a lot of the patching of the software and the support structures such as security is done for it.
“Most customers that start with an IaaS landscape still have a heavy amount of cost around operations, patching and application support which they didn’t think they were going to need. As they go up the stack they get less and less IT capability and move from operations to service management,” he said.
“When you get to the very top of the stack, to SaaS, which is where you get a capability through a log-in screen, then that’s a ‘no ops’ situation.”
According to a report published by 451 Research, 41% of all enterprise workloads are currently running in some type of public or private cloud. By mid-2018, it is expected that this will rise to 60%, and consequently a majority of enterprise workloads will run in the cloud in the near term. These are significant figures, with significant implications.