When the first era of the digital transformation occurred, a wave of businesses moved their data to the public cloud. Cloud management reduced their need for in-house data centres, provided better service, and reduced the costs associated with investing in up-to-date hardware. Public clouds were built around the concept that you only pay for the resources you use and therefore as business needs changed, costs could be scaled accordingly – a huge draw for developing businesses.
However, there is much more to running a service on the public cloud than server capacity alone and, unfortunately, if organisations are not careful their cloud costs can quickly spiral out of control. Indeed, a 2018 report by Gartner found that ‘By 2020, organisations that lack cost optimisation processes will average 40% overspend in public cloud’. This could clearly amount to a significant (and avoidable) cost to businesses.
The problem
From Reserved Instances to unblended rates (i.e. normal rates) and on-demand costs, attempting to make sense of a cloud bill can be confusing, even for CIOs. Working out what line of the business might be losing money and how spend can be optimised is not an easy process. Unfortunately, it is very easy to waste money if businesses don’t balance a structure we’ve come to call A, B, C – Architecture, Business Requirements, and Costs.
Balancing the Architecture is centred around how often the servers need to be running. If the majority of the work in the cloud takes place for example between 9:00-18:00 their capacity can be scaled back overnight, or on weekends. This ensures the server doesn’t become redundant when it’s not being used and avoids businesses racking up big bills during idle time. In the same way as in a house, turning off the lights when you’re not in the room saves energy and money and applies to cloud resources too.
Reviewing the Business requirements is also essential to determine the optimal cloud function. This is because rightsizing is the main cost waste in the public cloud management today. In general, businesses have a tendency to oversize their network interface to make sure they are prepared to accommodate peak traffic.
However, if we assume for a moment that the cloud is sized for a big event like Black Friday, it is essential that after the event is over the business scales back their use of the server to save costs. While business may be hesitant to reduce their cloud size in case of performance issues, without scaling the performance required and tailoring this around business requirements, power is wasted.
The last part of the structure looks at how Costs can be controlled from the outset, simply by leveraging usage discounts and choosing the best geographical region for cost for the business. Through analysing what hourly commitment is appropriate for the business and what percent coverage is required, it becomes clearer whether an upfront, partial upfront or no upfront payment is best suited to save costs. Another consideration is if a one year or three year commitment will save the most money.
Finally, cloud providers like AWS and Google Cloud have different resource cost regions due to demand. For example, services in the Northern Virginia region are cheaper than in Ireland. Given this, there is a need to find zones that are priced-appropriately for the customer, in order to optimise the cost from where their cloud services operates.
By controlling these three aspects, businesses be confident they are controlling costs at a full organisation level.
The solution
Addressing this complex, evolving problem also requires an accurate reporting tool to monitor, optimise, and forecast costs.
The best tools work by ingesting data from whatever cloud platforms you use today, third party tools you may use for performance management, as well as your data centre environments. This way you have a complete view of your entire cloud environment in one place.
Once all this data is collected, it is much easier to understand which business lines are overusing the cloud and it is possible to set up alerts to notify you if a daily cost is above average, or your monthly cloud budget is reaching its threshold. This way, you can optimize cost and prevent unusual costs before they become unmanageable.
If you are interested in cutting your cloud spending, investing in a third-party cost monitoring platform is a good idea. This is because while many cloud vendors provide their own monitoring tools to help customers keep track of usage, helping you spend less isn’t always in their best interests.
