The dominant three public cloud players, AWS, Google and Microsoft Azure have spent 2015 positioning their products as enterprise ready for mission critical systems.
Outages have hit all three companies, with some high-profile cases. September saw AWS suffer from significant error rates which impacted Netflix, Tinder and Wink, December has proved to be a bad month for Azure with a major outage across Europe for Office 365 and Google fared no better with an outage that lasted almost a day.
As users took to Twitter to air their frustrations during these outages, the company’s took to their health dashboards and cloud status pages to keep customers informed.
The question is, how much have these outages impacted these companies and their goal of running mission critical systems in the cloud. Have these outages slowed momentum?
If outages are not slowing momentum, then what they might be doing is offering the opportunity to other vendors to offer better SLAs or higher availability.
While these outages may not hurt the market adoption as a whole, they are offering other cloud service providers a way to compete, although the lessons being learnt are that cloud providers must offer greater transparency, no matter how small.
The larger cloud providers are providing greater transparency into their error reporting. Google for example reveals details of precisely what went wrong, along with a time frame of what actions they took to fix an error.
Cloud growth of the top three public cloud providers appears to not be hurt; AWS in Q3 2015, brought in $2.08bn in quarterly revenue, a 78% year over year rise. Google’s cloud revenue is grouped up "other revenues" with that division bringing in $1.89bn in the last quarter an 11% growth.
Microsoft is on target for cloud revenue to surpass revenue of $6.3bn for the year according to its last quarter, an 8% growth.
The numbers are clear in showing growth in the cloud units, but problems remain with transparency, outages and perhaps with the perception of cloud and on-premise.
Speaking to CBR, Owen Rogers, research director, 451 Research, said: "There seems to be a rosy retrospective going on, people looking back at on premise and seem to remember it being highly available and highly secure."
The problems of downtime and security existed with on-premise and they still exist with cloud, although cloud appears to be receiving a lot more criticism for it or certainly a lot more attention; the vast majority of cloud outages are widely reported
Rogers says: "A lot of it is that because you are handing over responsibility to someone else, you expect more."
With handing over control to someone else, there comes a lot of pressure in having selected the right provider, as Rogers said, "If it goes wrong they can lose their job."
There is a strong likelihood that cloud outages will persist in 2016 despite all cloud providers striving for 100% availability.
Cloud providers are aware of this and according to Rogers, have been very upfront about the need to architect applications across multiple availability regions.
While an outage in one region may impact you, if you have architected across multiple availability zones then you are unlikely to be severely impacted.
Rogers said: "People aren’t architecting across multi availability zones and that tends to be the problem."
Problems lie on both sides of the customer and cloud service provider relationship. Cloud has perhaps been oversold as something that will offer 100% availability, which when it isn’t met causes anger.
Customers for their part need to adapt to the changing world of cloud and follow the advice from cloud providers to architect correctly.
Despite the problems the cloud market appears unaffected.
