Price cuts and posturing in the public cloud war
The price battle among commodity cloud players is set to rage on, but how much impact is it really having for enterprise customers?
The price war among commodity cloud players is set to rage on, but how much impact is it really having for enterprise customers?
Tit-for-tat price cutting has continued among commodity public cloud infrastructure-as-a-service (IaaS) players throughout 2014.
The big three – Amazon Web Services (AWS), Microsoft and Google – have the infrastructure, range of offerings and financial clout to support highly-scalable, global public cloud services, and appear to be locked in a battle to the bottom on price, which is unlikely to let up in the coming years.
"The price war is now perpetual and will carry on until the cost of disk capacity, memory capacity and processor capacity stop falling – in other words, until Moore’s law breaks down," said analyst David Bradshaw, IDC’s research manager for software and services in Europe.
Google kicked off the war in earnest in March 2014, with announcements of a 32% cut across the board for its IaaS platform Google Compute Engine, as well as slashing the price of storage by an average of 68%. It also offered discounts of up to 53% (with no upfront commitment) when a virtual machine is used for a sustained period.
More on cloud computing
- Choosing between AWS versus Google versus Azure cloud storage
- Why AWS is being held back in the public sector
- Companies driving growth of cloud
- How resilient is the public cloud?
AWS responded days later with reductions of between 7% and 40% on its own offerings. Microsoft followed suit in April 2014, trumpeting cuts of up to 35% on compute and 65% on storage.
More recently, we’ve seen further salvos fired by all three players. Google cut across-the-board prices for Google Compute Engine again in October 2014 – this time by 10%. It followed this up in November 2014 with further reductions of up to 79% on specific services such as persistent disks.
Microsoft also cut prices on a range of Azure services in October 2014. And at the beginning of December 2014 AWS simplified its reserved instance payment structure into three tiers – pay all upfront, partial payment upfront or no upfront payments – offering discounts of between 30% and 75% to customers prepared to cough up for one to three years of use in advance.
AWS model more appealing to CFOs
While AWS’s reserved instance pricing structure is still not as straightforward as Google’s no-commitment sustained use discount policy, 451 Resarch senior analyst for digital economics Owen Rogers said it is now a lot simpler and will appeal to CFOs more than the old model.
If enterprises are confident they will use a reserved instance over a term, they can make substantial savings and get guaranteed availability
Owen Rogers, 451 Research
"If enterprises are confident they will use a reserved instance over a term, they can make substantial savings and get guaranteed availability, so it can be a worthwhile investment," he said.
As by far-and-away the IaaS market leader – with almost double the share of Microsoft and Google combined, according to Synergy Research Group’s most recent figures – AWS doesn’t need to use as much cost-cutting firepower as its ankle-biting rivals to keep them at bay, at least for the time being.
The company was the pioneer of commodity cloud services, has had a huge headstart and continues to develop and augment a compelling set of features and services.
Second-placed Microsoft, meanwhile, seems to be benefitting from CEO Satya Nadella’s cloud-first strategy, growing its IaaS business by 164% in the second quarter of 2014 compared with AWS’s 49% and Google’s 47%.
Google was late to the party, but has certainly shaken the market up in 2014 with its pricing innovations, and few would write off the chances of the IT industry’s most powerful player growing to a far more dominant position in commodity cloud provision, particularly since the market is growing ever bigger, ever faster.
Cloud savings depend on application
"I’d never bet against any of these players. All three have access to enough resources to do what it takes," said IDC’s Bradshaw. "Let’s not forget data volumes are growing continuously. Technology developments such as the internet of things will lead to even larger volumes of data being collected and that data has to be stored and processed somewhere."
The exact savings achieved depend on the application and its use case
Owen Rogers, 451 Research
Yet the effect of the price war has so far not been as dramatic for customers as headlines might suggest. 451 Research produces the Cloud Pricing Index (CPI), which compares changes to the cost of a typical application over time, averaged from a range of providers’ quotations for virtual machines, storage, database, bandwidth, support, and so on.
451’s Rogers said: "Although there has been a lot of excitement around pricing changes, we found that since the last CPI measure in October 2014, the CPI has only declined by about 1%. The average monthly price for our typical app was $2.56 per hour in October 2014 and in December 2014 it was $2.53.
"So although any price cut is good news for consumers, enterprises need to remember their costs will amount to more than just virtual machines and storage, and they need to ensure they are aware of all their financial liabilities. The exact savings achieved depend on the application and its use case."
Sizing up AWS vs. Azure vs. Google Compute Engine