For the last few years, you could have been forgiven for thinking that eventually every single company in the world would be using public cloud for all of their IT needs and the on-premises hardware, private cloud and colocation industries would eventually either transition to provide services on AWS, Azure or Google Cloud or disappear.
“We have a cloud first policy” became a mantra for everyone from SMEs to large enterprises and from charities to local government. If you were involved in selling hardware, hosting, or data centre services it started to feel like a change of direction was necessary before your market left you behind.
Public cloud does offer some compelling benefits, there’s no doubt about that. If you need to be able to scale up and down to meet seasonal changes in your business, then auto-scaling is for you. If you are a start-up growing extremely rapidly and every penny needs to go into maintaining that growth, the capex cycle of buying your own hardware makes no sense at all. If you can redesign your workloads to take full advantage of the native databases, services and applications, then you can make amazing things happen very quickly and cost effectively when certain trigger events occur. Public cloud is amazing, and it is here for the long term.
Which raises the question, why are many companies, especially large enterprises, repatriating workloads to their own, or third-party, data centres? Why would 72% of respondents to the 2021 State of Hybrid Cloud Survey say that they had moved applications back on-premises?
There are a number of reasons why cloud repatriation happens, including security, management and service quality issues, but for many enterprise customers the main driver is financial. While the hyperscale cloud providers were growing there was a ‘land grab’ for customers, which created huge downward price pressure on the services, especially those that were seen as being commodities, such as virtual servers and storage, rather than unique value add services like Lambda.
However, in the last couple of years the cloud providers have had to switch to being independently profitable rather than having their growth funded by other parts of their parent organisations, so the trend now is for prices to increase steadily.
The obvious prime candidate workloads for repatriation are those that require low-latency, have particularly high compute requirements, are graphic intensive or have large, fast storage needs. If they are also relatively stable and grow at a predicable rate, then it can save a lot of money bringing those workloads back into your own or a hosting partner’s data centre.
In 2018 IDC published a report that found that companies were increasing both public cloud and on-premises private cloud spend. The next year they reported that 80% of companies were planning to repatriate 50% of workloads. By 2021 the Virtana survey linked above showed that companies had very much carried out those plans. There have been reports of companies saving between 30% and 50% on repatriated workloads, which is why Dropbox decided to go on-premises in 2017, saving an estimated $75M.
Navigating through the complexities of what should be in the public cloud, what should be on-premises and which workloads could be better hosted by a third party is time consuming and requires an in depth understanding of all of the possible end states. Advanced have the expertise, knowledge and certifications in public cloud, private cloud, SaaS and on-premises solutions along with real world experience gained through years of guiding customers to help them plan for future, with every workload located in the best possible place.
If you’d like to speak to an expert to discuss your cloud environment and workloads, simply get in touch today.