OpenStack private cloud deployments may work out worse value for money than commercial alternatives, because of the scarcity of engineers with experience of using the open-source platform.
That’s according to 451 Research’s latest Cloud Price Index, which set out to shine a light on the complex pricing models of private cloud for the first time.
This information is relatively hard for users to come by, Owen Rogers, senior analyst at 451 Research, told Computer Weekly. He said the procurement of private cloud services tended to follow the same path as on-premise deployments.
“It is far more challenging to collect private cloud pricing information, as public cloud details are all online and easy to get,” said Rogers.
“But we still anecdotally hear that private cloud is purchased in the same way that IT has always been, so there are requests for proposal, negotiations and custom pricing involved.”
Read more about OpenStack
Researchers approach private cloud suppliers
Analysts at 451 Research sent out a hypothetical requests for proposal (RFP) to 20 service providers and software orchestration suppliers for a small-scale private cloud, featuring 500 virtual machines (VM) running across 25 nodes.
By publishing the results of its research, the organisation hopes to bring a greater degree of transparency to the sector, and help users make better informed decisions when weighing up prospective private cloud providers.
The research from 451 Research revealed the cost of running a private cloud powered by Microsoft, Red Hat or VMware varied by around half a cent, working out at about $0.10 per VM hour.
Running an OpenStack distribution would work out around 20% cheaper than these commercial competitors (at $0.08 per VM hour) 451’s findings showed, but – despite the lower cost – the total cost of ownership (TCO) could work out more expensive for users in the long run.
This is because there are fewer skilled OpenStack engineers available, compared to those with experience in supporting commercial private cloud offerings. This means users face having to spend a lot for engineers with the skills to manage the open-source cloud offering.
“451 Research believes that, for a typical deployment, buyers could hire 3% more engineers to support a commercial cloud environment, and still have a lower TCO compared to an OpenStack distribution,” said the report.
There are a lot of good reasons for choosing OpenStack, particularly when it comes to the risk of lock-in
OpenStack adoption drives down costs
However, as adoption of OpenStack increases, and the number of skilled support engineers grows, the situation is likely to improve, said Rogers.
“Even though it might be more expensive in TCO terms to use an OpenStack distribution now, as time goes on, the expense associated with moving will become a lot less, and the value of using an unlocked-in platform will really come to the fore,” said Rogers.
“As more OpenStack engineers enter the market and the technology becomes more mature too, then it’s perfectly likely the tipping point will go the other way and it will become better value than some of the more commercial technologies.”
Meanwhile, Jonathan Bryce, executive director of the OpenStack Foundation, told Computer Weekly the relative shortage of Openstack-literate engineers is being keenly addressed.
"We’re already seeing a lot of programmes to make the talent pool bigger around OpenStack," he said, and as adoption of the technology increases, the number of people with experience using it will naturally grow.
"Whenever there is a big technology shift, there is always a lag in engineering talent, and it's a temporary thing."
While TCO is an important part of weighing up which provider to go with, it should never be the sole deciding factor, Rogers added.
“TCO calculations should just be one part of a broader picture, and – of course – you should investigate the specific value you’re getting from each supplier,” he said.
To emphasise this point, he said users stand to gain many benefits by going down the OpenStack route.
“There are a lot of good reasons for choosing OpenStack, particularly when it comes to the risk of lock-in, which can be a problem if users opt for one of the commercial, more proprietary suppliers,” he said.
“An important point to make is that, when investing in any IT technology, you should look at what will happen in five years’ time and the financial cost or security penalty of moving data to a new provider.”
This is a viewed echoed by Bryce, who said: "When you make a call on what cloud platform to use, that is a 5-10-year decision. It’s a foundational piece of technology for your business, so you want something that will give you a lot of choice and a continuous pipeline of innovation. And I think people definitely see that coming out of the OpenStack community."
Managed private cloud vs. in-house support
The research also compared the TCO of running a managed private cloud against one supported in-house by an organisation’s own IT team, and established a so-called “golden ratio” about how many VMs an engineering team can comfortably support to achieve an advantageous TCO.
“If a private cloud is being successfully operated with a ratio of at least one engineer to about 100 virtual machines, it may be deriving better TCO than the average-priced, managed private cloud,” the 451 Research report said.
“If each engineer is not able to fully support this minimum amount, an average managed cloud is likely to be cheaper.”