How to manage a 24×7 private cloud with one engineer

Recent reports show that over 80% of organisations are either transitioning to a private cloud or considering building their own, making cloud software solutions such as Openstack the focal point for the majority of organisations. This has also led to increasing interest in cloud-native applications and Kubernetes adoption. How is it possible to achieve an efficient deployment of your cloud and operate it at the cost of only one full time engineer?

Why do we need digital transformation? Really! 

In the last several years, we have witnessed the creation of many technologies, starting with the cloud and going further to machine learning, artificial intelligence, IoT, big data, robotics, automation and much more. The more the tech evolves, the more organizations thrive to adopt these technologies seeking digital transformation and disrupting industries along their journey, all for the benefit of better serving their consumers.

With every technology having its own requirements, costs and benefits, the only common aspect between any technology you decide to invest in is one thing: it is all based on achieving a business goal that will help your organization better position itself in the market. You might be luckily taking advantage of leading your field, or looking to better serve your customers, or even keeping up with tough competition. Whatever your motive is, the aim will always be to realise a business goal out of your investment.

What are the overlooked costs?

As much as organisations benefit from digital transformation, everything comes at a cost. With cost here, I mean to refer to the friction that opposes the successful implementation and smooth operation of the technology rather than the financial cost of acquiring the technology. This leads to the most important overlooked question when considering a digital transformation: Who cleans the windows after raising a shiny artifact? The massive acceleration in evolving technology has made it even more challenging to achieve sustainable operations and has added a technical debt in post-deployment that organisations should not be immersed in rather than keeping their focus on activities that drive impact. 

“When we pick up one end of a stick, we pick up the other end”, Stephen Covey said in his famous bestseller The 7 Habits of Highly Effective People. The truth is that the rule applies to every aspect of life, but is it debatable in technology adoption? Is it possible to achieve a successful digital transformation journey without having to pay its costs? Is it possible to remove the friction opposing your journey, and avoid increasing the complexity of your operations, growing your team exponentially to keep the lights on while spending months on finding the right talent? 

How to pick up only one end of the stick, your goal?

We all want to pick up the gain and avoid the cost whenever possible, it’s human nature! Think of your car insurance for a moment, what makes you want to pay the insurance company for their service (apart from being mandatory)? They simply pay your cheque, and may even provide you with a temporary ride until your car is ready to hit the streets again. They might even come to pick it up and hand it over at your doorstep after it’s fixed. The important thing is that you don’t worry much about what’s happening behind the scenes, as you have more important things in life to take care of rather than spending time, effort and money getting your car fixed. You bought the car to get you around conveniently and fixing it was not part of your goals, but a consequence of buying the car instead. You simply let the insurance company take care of this, and hold the other end of the stick for you.

Building a private cloud can be complicated, but operating a private cloud efficiently is definitely challenging. Taking Openstack as an example, being the preferred choice for organisations building private and hybrid/multi-clouds and being the most widely used open source cloud software. The post-deployment operations are said to be complex due to requirements of expertise in different layers of the stack, in addition to regular firmware upgrades that require proper planning, backup and fallback strategies to ensure safe upgrades. What if you further require to deploy Kubernetes on top of Openstack to further enable focus on developing your cloud-native applications? Luckily, there’s a similar “car insurance” story when building your private cloud using Openstack or deploying Kubernetes, but let’s first wrap up what you’re looking for at that stage:

• Access to a pool of specialised experts, at a reasonable cost
• 24×7 operations and support
• Full visibility and control over the costs
• Secure your data, infrastructure and applications
• Efficient deployment and even more efficient operations
• Shared risks and minimum failures and downtimes
• An up-to-date cloud infrastructure
• And most importantly, you need to maintain focus on the business goal rather than commodity activities that do not drive impact. 

Managed IT services can provide you with the opportunity to offload many commodity activities and benefit from the knowledge and experience brought by the managed service provider’s team of cloud experts, and help you maintain your focus on driving your business. 

Let’s Talk Numbers! 

It’s time to think about efficiency, and the financial cost of outsourcing to a managed service provider compared to hiring a dedicated team. So first, let us breakdown the operational costs of a self-managed cloud:

• For 24x 7 operations throughout the year, you require coverage of 8,760 hours
• A full-time equivalent (FTE), after removing weekends and PTO, has an annual availability of 1,776 hours (222 days x 8 hours/day)
• Accordingly, 5 FTE are required to operate the cloud with no redundancy
• For production-level SLAs, 2x FTEs are required at the same time for redundancy, and accordingly, 10x FTEs is the minimum number of engineers required to operate 24x 7 
• The average annual income of a Cloud Operations Engineer is ~ $ 98.8K, and the minimum is   ~$ 72K
• Considering the minimum income, operating a private cloud will require an annual human resources budget of at least $ 720K. This does account for any turnover, time and costs for training and development or any other work-related activities.

What if you want to start your cloud small, and scale as the business grows? Would you want to add this massive overhead to your expansion plans rather than allocate it to core activities that will support your successful journey? 

Now Really, How can I Manage a 24×7 Private Cloud with one Engineer?

Let’s get this straight, it is not possible to operate a private cloud with only one engineer. This is due to the architecture of private cloud solutions having many integrated components, each requiring a unique technical skillset and expertise. Although this is true, you can still operate your private cloud at the cost of one engineer by partnering with the right managed service provider (MSP). 

An MSP can help you significantly minimise costs and accelerate the adoption of the cloud while maintaining your focus on what you do best, driving your business. This is possible because an MSP simply has the ability to leverage the same pool of specialists to serve different customers. This way, multiple organizations benefit from running their 24×7 operations by experienced professionals at a much lower cost compared to hiring the same engineers exclusively.

If we consider Canonical’s offering for cost comparison, a managed Openstack service is provided at the cost of $ 5,475 per host annually. The minimum number of nodes to build an Openstack cloud is 12, making the total annual cost for operating your cloud less than $ 66,000, which is $ 6,000 less than the minimum annual income of one full-time cloud operations engineer. This allows your team of IT specialists to focus more on innovation and strategy rather than keeping the lights on. 

Want a fully managed private cloud?

Kubernetes or OpenStack? On public clouds, on your premises or hosted? Tell us what your preference is!

If you want to learn more, watch the “Hosted private cloud infrastructure: A Cost Analysis” webinar.