Simon Heath offers four strategies for actualising the promised cost savings of cloud computing.
In traditional environments, IT budgets were predominately built from past procurement activities. This meant if you spent $100,000 on a renewal for product X, you would have a similar line item added to next year’s budget. Whilst some adjustment for CPI might be allowed, the line manager would then be expected to manage to these fixed cost structures. Under such a budget construct the degree of micromanagement from the financial team was light touch, commonly based on a monthly review meeting to make sure all the numbers were aligning. Someone simplified I know but in general quite straight forward.
However, for organisations transitioning to cloud services, this traditional approach to IT budgets is being challenged – primarily because the cloud is effectively a variable cost model more closely linked to business activity and how IT operations satisfy this demand technically, needs to change.
Any organisation embarking on a move to cloud must ensure the business processes supporting financial budgeting are similarly transformed across the key stakeholders. Failure to do this typically results in unforeseen budget creep, followed by a rapid loss in business confidence with how IT operations are managing their budgets, something you just don’t want to happen.
To mitigate against this outcome, it is critical for IT to become more sophisticated and transparent around its financial management practices across the organisation. So, to achieve this here are my three top recommendations for starting such a journey whilst also reaping the benefits of migrating your IT infrastructure to the cloud.
1.Keep it transparent for everyone
In the old days, while Finance had rigid control over the fixed IT budget, there wasn’t the need to much transparency. The organisation may have invested capital in equipment to place in its data centre and that was the end of it unless IT negotiated for more capacity. If the equipment was not being used to capacity, it had been paid for so there was nothing to recover, after all that spare capacity was a sunk cost to support for future organic growth.
Cloud, however, is OpEx and variable. Monthly fees are unpredictable and might suddenly balloon – which is an unfamiliar concept for most finance teams. This means that it is important for IT to give Finance transparency over-usage – because transparency creates trust. They must also demonstrate they are effectively managing cloud expenditure, to ensure what growth is occurring is justified and implemented in an optimised configuration to just support the business demand.
This calls for astute technical and financial planning. Unfortunately, a lot of IT departments don’t actually monitor and calibrate their cloud usage carefully in order to reap the cost benefits over fixed, owned infrastructure. Here’s one example:
- In the old days, you’d spin up a machine for development and it wouldn’t cost any extra to be available 24/7 because it was deployed on existing hardware that happens to have spare capacity.
- When you spin up a development server in the cloud, you could elect to pay, run the resources, for only 12 hours a day, five days a week when developers are actually working, saving the business money.
Just by asking the question: “When do we actually need to use this dev server?” you can save 60% or more of its cost!
Similarly, you can save on storage by keeping careful track of what you’ve got. Instead of renting more and more cloud capacity just because you can, why not continue to archive and delete as you did when you had to buy storage outright?
2. Set lifecycle gates to check savings
Design Governance around traditional applications was based on their hardware lifecycle. You would set up the architecture according to your security, performance, capacity and other requirements. Even refresh cycles would often be set in concrete based on the cost of maintaining or upgrading the infrastructure.
With the cloud, the architecture process is continuous and typically part of a methodology, such as the AWS Well-Architected Framework. With the rapid rate of new or updated services released daily by the public cloud providers, in excess of 1 ever 4 hours for AWS in 2018 for example, the value of regular revalidation of the design to further optimised costs should not be ignored. To do this effectively your staff must be technically trained to understand new capabilities available on cloud – because you are no longer restricted only to the capability of the technology you’ve deployed. More importantly, processes need to be established to ensure these reviews have the appropriate triggers and resource capacity to efficiently execute.
As already mentioned, in 2018 alone, Amazon Web Services (AWS) issued over 1,600 changes in the form of new or updates to their service catalogue. These changes might be as simple as a generation change in an EC2 processor family which, if adopted, can deliver an increase in CPU performance and lower instance cost by as much as 10%.
Essentially, your infrastructure can constantly evolve with new cloud capabilities, without changing the end-user experience. But to benefit from this evolution, you must stay abreast of innovation and constantly review potential savings and economies as part of your normal operating processes.
3. Trace cloud resources to manage them
Cloud resources may be virtual, but you must still track them as though they were physical resources. A good practice to achieve this is by adopting a good tagging methodology, to allow resource expenditure to be broken down and filtered into expenditure buckets that make sense to the business. This might cover categorises such as Application Name, Project Number, Owner, Environment, and Service Level, it’s not a prescriptive formula, but when set up correctly becomes a powerful tool to manage your cloud resources and provide more granular information to finance.
It also helps you manage different categories of resources and take advantage of different cloud license constructs. Is your payroll system heavily used doing no critical processing for three days twice a month? Then consider using a ‘spot price’ model for those high capacity days. By having an effective tagging model such actions can be easily implemented.
Ultimately a good tagging taxonomy should provide traceability – enabling the grouping of costs in business language. In the above example, by tagging cloud resources against your payroll system you can report on usage in order to do what is typically referred to as “show back” of current consumption and even forecast future consumption. Doing this in traditional models has been challenging and somewhat irrelevant when operating on sunk cost equipment. However, in a Cloud environment, it becomes an effective tool to provide visibility back to the ultimate owner of the service to open further dialogue with them on cost optimisation strategies.
4. Reserved Instance (RI) Model Strategy
Adoption of a RI strategy is a sound way to achieve savings on constantly used resources for customers and each public cloud provider have this capability. The new monthly pricing structure update from Microsoft allows customers the ability achieve the RI savings, which can be as high as 80% but still more accurately align billing to monthly consumption to smooth out your cash flow.
When looking at adopting a RI strategy each customer needs to have a sound understand of their consumption patterns and also the intricacies of the Reserved Instance Model being evaluated. In the case of Azure RI some key points to consider;
> RI discounts typically apply to the compute component of a resource and not other areas such as storage, network and most importantly Windows Server Licensing., so these need to be excluded as part of any TCO
> Additional just keep in mind a pay by the month RI strategy is still based on a one or three year commit, so complete flexibility you get with Pay-As-You-Go is removed, and changes to your commit might result in early termination fees.
Microsoft has a useful webpage to provide more detailed information about RI and the new pay by the month strategy here https://azure.microsoft.com/en-au/pricing/reserved-vm-instances/#faq
. Alternative our Cloud Consulting Cost Management team can provide the services to help a customer define, build or review a RI strategy to maximum savings.
Need expert help?
As mentioned above, it takes considerable skill and knowledge to keep on top of cloud offerings, because your options are continually expanding. You are not confined to one cloud supplier, either.
I’ve covered off a few key considerations to managing your organisations cloud expenditure, but the critical message is if you don’t uplift you people and processes when adopting cloud as your technology platform achieving an optimised cost structure will be challenging if not impossible.
A partner like SXiQ has significant experience with helping organisations transform their technology, people, and processes to ensure cloud adoption is sustainable and those “Sticker Shock” moments are avoided. We see too many organisations contemplating a retreat back to on-premise models, but which on closer analysis have not completed this end to end transformation within the IT function.
To learn more about our managed and cloud operations services, contact us.
About the Author
Simon Heath is CTO at SXiQ. With decades in the information technology and services industry, he has strong skills in IT strategy, professional services, cloud transformation and data centre management.