451 Research has released some very interesting research this week, comparing the cost of ownership between commercial private cloud offerings and OpenStack distributions – where it has taken into account the price of skills and compared this to how many virtual machines an engineer would have to manage to make the investment 'valuable'.
The most surprising thing to come out of the results is that 451's use case of a small-scale enterprise revealed that commercial private cloud builds from the likes of Vmware, Red Hat and Microsoft are only marginally more expensive than OpenStack distributions, and actually work out cheaper when taking into account the price of OpenStack engineers in the Bay area.
And although commercial cloud offerings are inherently more proprietary in nature, versus the freedom and flexibility that OpenStack theoretically offers, the research may give buyers pause for thought.
Now, before we go on, it's worth pointing out that there are some limitations to the results. 451 Research obviously can't produce comparisons for every enterprise size and need, and so the numbers reflect a buyer that would be looking to implement a private cloud of 25 nodes and 500 virtual machines. It's just worth bearing this in mind, as 451 isn't suggesting that the results reflect every enterprise buyer's situation.
However, the level of detail in the research is noteworthy – when talking about cloud price and value we often deal with anecdotal evidence. It's great to have some hard numbers to support this.
But let's dive into the numbers. 451 Research found that the typical cost of their small-scale enterprise private cloud scenario powered by Vmware, Red Hat and Microsoft are all within half a cent of each other at about $0.10 per virtual-machine hour.
Whereas OpenStack distributions cost, on average, $0.08 per VM hour – equating to a 20 per cent saving. Not to be sniffed at.
But when you take into account other factors – such as the price of OpenStack engineers (priced at an average of $126,000 per engineer, versus $91,000 for a Vmware, Microsoft or Red Hat engineer) – 451 estimates that for a typical deployment, buyers could hire 3 per cent more engineers to support a commercial cloud environment and still have a lower cost of ownership compared to an OpenStack distribution.Dr Owen Rogers, senior analyst at 451 Research and author of the report, said:
Finding an OpenStack engineer is a tough and expensive task that is impacting today’s cloud-buying decisions. Commercial offerings, OpenStack distributions and managed services all have their strengths and weaknesses, but the important factors are features, enterprise readiness and the availability of specialists who understand how to keep a deployment operational. Buyers need to balance all of these aspects with a long-term strategic view – as well as TCO – to determine the best course of action for their needs.
Decisions also need to include the risks associated with lock-in should prices rise, support dwindle or features be decommissioned,” Rogers added. “As OpenStack matures and the pool of available engineering staff increases, buyers can expect the TCO of deploying OpenStack to improve.
Rogers also looked at what he describes as the 'golden ratio', which details the value of a private cloud versus a managed cloud, based on how many virtual machines an engineer can successfully manage.
For example, the research found that for commercial providers, the point at which it is more valuable to have it managed by somebody else is when an engineer cannot manage more than 91 virtual machines. If your engineer needs to be managing 150 virtual machines, but can't, it is better value in terms of total cost of ownership to have it managed by somebody else.
This golden ratio for OpenStack distribution is 118 virtual machines per engineer. For a DIY OpenStack cloud, the number of virtual machines an engineer would have to manage in house to make it valuable is 106.
If an enterprise was considering the price of public cloud, the golden ratio grows significantly. 451 Research found that for a self-managed private cloud to be cheaper than public cloud, an average golden ratio of 250 virtual machines per engineer is required. Any less and public cloud is probably better value.
However, despite the cost of ownership savings that commercial providers seem to be able to deliver, Rogers doesn't write off OpenStack for buyers. There are a couple of paragraphs in the report that I think are worth highlighting for those considering their cloud options – particularly in relation to the future of OpenStack and its lack of reliance on a particular commercial provider. The report states:
Legacy offerings may be the better all-rounders at the moment, but if support dwindles, pricingincreases or reliability fades, consumers are trapped (at least to some degree) into using their existing provider and technology – this is a business risk, with a financial implication. OpenStack, in theory, should allow consumers to move as they please without being stuck to a particular vendor, although the number of reference architectures available suggests this might not be as easy in practice – the price of this freedom is greater TCO and a less mature capability. Does OpenStack's removal of lock-in reflect its higher TCO? Enterprises need to ask themselves this question when looking at their overall business strategy.
It goes on to add:
In terms of TCO, proprietary offerings – at the moment – seem to provide the best value for our requirements. However, since they are commercial technologies, this may not always be the case. Price increases, product decommissioning and other factors may increase TCO, with no simple, inexpensive way out for consumers. This is less of a problem with OpenStack, where the promise of interoperability and portability that the industry touts should allow consumers to move between providers and offerings, which will lead to lower market costs.
Commercial offerings are largely more mature, have more features, and have been widely tested in the market. OpenStack is a disruptive force supported by a wide range of blue-chip vendors, withpurportedly less lock-in, but still not without its risks (e.g. fragmentation). As more skilled OpenStack engineers enter the labor market, the TCO benefits of the proprietary software packages will be reduced as salaries are reduced as well. Enterprise buyers evaluating and negotiating with vendors will want to bear this in mind.
The report is extensive and detailed – to get a copy, click here. Very useful information for buyers that want the hard facts relating to the cost of ownership. However, it's worth mentioning that in this day and age it isn't just about numbers anymore. It's also about the flexibility, scale, freedom and agility that a deployment can give you. Those are things that the numbers won't ever be able to tell you, but are equally important.