Public cloud is gaining traction with more housing-sector deployments each year, but potential cloud customers have the difficulty of justifying the costs and risks of moving from their existing on-premise infrastructures. Private-cloud IT infrastructure takes significant time, risk and cost to deliver, and over time will have received many upgrades through software and hardware improvements so service owners are loathe to discard an entire IT asset that still provides value and would require significant migration effort to replicate in the cloud.
A ‘big bang’ replacement approach will inevitably result in less than optimal outcomes; replacing equipment with remaining serviceable life is not just financially wasteful but may also cause more disruption than necessary, particularly where any change also equals risk to a critical IT service.
If a carefully-planned approach to deciding which solutions require replacement is carried out and other factors, such as warranties, supportability, performance capabilities and capacity, downstream impacts of failures and the ability to recover from them, are all considered then a more selective approach to maintaining and augmenting on-premise solutions can be applied. The careful integration into a public-cloud solution such as Microsoft Azure can be used to create a hybrid cloud to mitigate constraints while avoiding the need for broader and more indiscriminate transformation activities.
The term ‘hybrid cloud’ often includes public-cloud infrastructure, on-premise or hosted private cloud and some form of WAN, all of which can be established with minimal changes to the existing on-premise environment. While WAN connectivity can be expensive, in reality the cost is greatly impacted by performance and reliability, so with careful planning, the first services to leverage public cloud can be tolerant of the occasional WAN performance issue. The public-cloud setup is also relatively inexpensive and fast if your supplier is using good practice such as ‘infrastructure as code’ to ensure its deliveries are repeatable, reliable, secure and fast.
A low upfront investment means the levels of obligation are low (no rushed migrations) and there are fewer reasons not to commit. By creating a hybrid-cloud solution, you have suddenly created access to a huge number of services waiting to be used with no cost until leveraged – they are simply in the public cloud, waiting until they’re needed.
Now the question is how and what to begin consuming from the public side of your hybrid cloud, and this is where many people stumble. The answer lies in the challenges faced by your on-premise environments and the new capabilities best aligned to the business demands on your IT services.
Before we go any further, let’s point out that cloud unleashes a vast amount of capability that’s on tap and ready for consumption at any time. However, it’s very important to still apply your IT service principals of strategy, service design, service transition and operations because if left unchecked, IT administrators can unwittingly open up huge security holes through the wrong placement of a unprotected public IP address, the incorrect granting access of rights through, say, Azure Active Directory, or poor planning and implementation of network security.
Our recommended starting point is to build on the three components of a hybrid cloud (public cloud, on-premise hosted private cloud and WAN) by starting with a thorough design for your hybrid cloud, with a strong governance and security theme.
We can briefly consider some typical demands that could be easy wins and useful ways of dipping your corporate toe into an initially tepid-seeming pool of public cloud.
Test and development environments are a great example of ‘low risk with high benefit’. They might be in your on-premise Hypervisor and storage environments now; instead consider using public-cloud development and testing environments as and when needed, with lower SLAs and typically 50 per cent cheaper.
DR is usually a delicate topic in IT departments – there are burdensome processes such as routine rehearsals, scripting of failovers and impacts on operational systems to consider, yet all of these can be addressed using public cloud-based DR and replication.
Operational costs can be forecasted with incredible accuracy and in great detail, allowing you to set multi-year budgets and charge services back to departmental budgets. Plan carefully the use of pay as you go and ‘reserved instance’ pricing – some workloads are particularly suited to pay as you go if you use automation to de-allocate and shrink according to varying demand. However public-cloud providers like commitment and reward customers with prices that can be around 70 per cent cheaper, so think about workloads that are always running and carefully plan what size virtual machine is needed to meet the necessary demand then reserve the instance. That said, virtual machines are often the most expensive option and there are usually non-server-based alternatives such as web applications, load balancers and Office 365-based solutions.
The setup complexity and cost for your hybrid cloud can be commensurate with your early levels of demand and commitment. As your business and IT departments become more familiar, assured and organised to consume more public-cloud services then they can be configured as and when needed to facilitate this greater level of adoption. Setup and operational costs can be accurately forecasted and controlled, and security and resilience can also be granularly applied on a system-by-system basis.
Hybrid cloud is the best of both worlds, combining the powers of both on-premise/private cloud and public-cloud services, with the ability to move services between environments depending on where the need is best served. In short, hybrid cloud enables businesses to deliver incredible services to their customers while greatly improving flexibility and reducing risk.
Jason Birchall is the founder and managing director of A4S Cloud Solutions.