insights

5 steps to gaining control of your cloud infrastructure costs

March 19, 2020

Most organisations have migrated part or in some cases all of their IT out of the data centre and into the cloud. 

Most likely your organisation is one of them. 

In the current climate, cost optimisation is even more vital than normal. Reducing running costs and operating a lean organisation will be key to staying in business.

Do you know how much we are spending on public cloud infrastructure? 

Do you know where cost savings can be made during these uncertain times?

It’s likely that your answers to the questions above are No, No and No. But you are not alone.

According to this year’s State of the Cloud Survey optimising cloud use for cost-saving continues to be the top initiative for enterprises [1]. However, in the latest Cloud Management Report, only 30% of organisations stated they were able to manage and optimise their cloud costs [2].

If it’s such a high priority how is it that less than a third of organisations can do it effectively?

With companies such as Netflix spending a reported £18.5 million per month on Amazon AWS services alone [3] it’s easy to see how important optimising cloud spend can be.

So how does an organisation gain control over their cloud spend while still maintaining the numerous benefits that migration to public cloud infrastructure brings?

How to get control of your cloud infrastructure costs

There are broadly three approaches to reduce cloud infrastructure costs:

  1.  Use less resources: Cloud vendors generally charge using a pay as you go model. i.e. you pay for what you use. Due to this it’s obvious that the less you use the cheaper it gets. Compute and storage are the main culprits but this principle applies to all cloud services that your organisation uses.
  2.  Pay less for your existing resources: Cloud vendors provide options that organisations can use to run the same estate for less. This approach includes discounted pricing models and using alternative products or services to reduce costs..
  3. Do more with less: optimise your architecture to get more out of your existing resources

To deliver cost savings, organisations can apply one or more of these seemingly simple approaches  To achieve this there are five key steps that IT departments must take.

1. Know what you use: Get visibility of your cloud usage

It sounds obvious but unless you know what resources your organisation’s  cloud infrastructure is using it’s going to be impossible to optimise usage and cut costs.

All the major cloud providers have tools of differing capability for producing billing reports and performing cost analysis. Make sure you use them. These automated analysis tools also make recommendations that can be easily applied. Though they are quite sophisticated, as always, with automated analysis you should fact check all recommendations before putting them into operation.

2. Know where you spend: Tag resources

All major cloud vendors provide ways of labelling or tagging resources. This will help you to attribute the resources to the applications, services and departments that use them.

Make sure all resources are tagged with the department or cost centre using that resource and an environment name, such as development, test and live.

Tagging enables greater insight from usage data. Empowered with better understanding of the usage it is possible to drive cost savings through automation. Organisations can automate shutdown and startup of services and environments when they are not needed. For example, organisations can achieve significant savings by automating the shutdown of all development resources out of working hours.

Tagging will also make it easier to identify unauthorised use of cloud within your organisation. So called “Shadow IT” can occur in organisations where the IT department is slow to respond to user demand. This happens for a variety of reasons including cumbersome governance processes, slow moving commercial teams etc. Identifying the use of Shadow IT will help improve data governance and compliance while reducing cyber security risk. Mistakes in these areas can cause significant reputational damage and come with a high financial cost.

3. Understand your utilisation: Right size resources

Right-sizing resources with auto scaling enables cost savings by removing the need for deployments scaled for maximum load, with excess resources allocated but unused most of the time.

The benefit of immediate scalability in the cloud comes with a pitfall – the potential to incur significant unexpected costs when a system is manually scaled up during a period of high system load and not scaled back. It can be too easy to address poor system performance with more resources rather than by application and / or system optimisation. Scaling up compute resources is not the long term answer to fixing a poorly architected application.

Use monitoring and auto-scaling functions to scale up or scale out compute resources when systems are under high load. Auto-scaling will scale down when demand drops.

Visibility, analysis and understanding of your resource usage and system performance is key to achieving value and return on investment in cloud infrastructure.

4. Pay less for resources

Once you are armed with a better understanding of your cloud usage, you can further optimise costs by paying less for what you need.

By taking advantage of cheaper resources immediate cost savings can be made.

The top three cloud vendors all have a system to sell unused compute capacity at a discount. You set the maximum price you are willing to pay and when the price drops below this threshold your workload runs. 

AWS and Azure call these Spot instances or Spot VMs. Google uses the term preemptible VM instance. Compute resources of this type do not support all workloads. They are suitable where the time of task completion is not critical and where the workload can stop and start without issue. Good candidates are data transformation and data analysis.  Significant cost savings, up to 90%, can be achieved if your workload can be supported.

Further discounts can be achieved by committing to a level of compute usage up front. This feature should be used when load is sustained and predictable. Committed use or reserved instance pricing can achieve discounts of up to 75% compared to pay-as-you-go pricing.

5. Optimise usage and share knowledge across the organisation

It’s becoming increasingly common to co-ordinate cloud usage between departments [2]. Moving cloud management into a centralised function enables economies of scale across the enterprise and avoids siloed IT that can occur in some organisations.

The central team, sometimes called a Cloud Centre of Excellence, can enable cost savings by ensuring that optimisations are applied across the organisation and ensuring common pitfalls of cloud migration are avoided. The team will also provide architecture patterns to enable efficient resource utilisation through better infrastructure and application design. By recommending approaches using Platform as a Service (PaaS) options, instead of Infrastructure as a Service (IaaS) significant savings can be made in management, support and licensing costs.

Cloud Optimisation from 6point6

6point6 have a proven track record in providing cost-effective digital transformation. Our in-house team of consultants with expertise in all three major cloud platforms can help your organisation optimise its cloud infrastructure and save your organisation money.

From an initial assessment of your cloud usage through to full DevOps automation of your cloud infrastructure and everything in between, 6point6 can help your organisation do more for less.

References

[1] RightScale 2019 State Of The Cloud Report

[2] Cloud Management Report (https://click.cloudcheckr.com/rs/222-ENM-584/images/CloudCheckr-White-Paper-The-Cloud-Infrastructure-Report-2020.pdf)

[3] Cloud Costs – Who Spends What on AWS?

Email Phil Fordham for more information or ask questions.

PHIL FORDHAM 

Solution Architect

Phil Fordham
Solution Architect