In order to get the most efficiency out of cloud, it is important that companies are building the right foundation and organisational governance approaches for cloud management.
This will include areas such as security & access management, governance & policy, service blueprint & orchestration, observability & self healing, capacity & resource planning and cost optimisation. The hot item we would like to focus on as part of this blog is cost optimisation. Managing the financial operations of your cloud platform is usually key to the business case.
Many organisations we talk with have now made cost optimisation of cloud resources a priority. As there has been massive growth in public cloud services, this is not surprising as it is unlikely to slow down anytime soon. Gartner is predicting public cloud service revenue to grow to over $330B USD by 2022. This represents a growth rate of over 50% from 2019 figures and further highlights the need to ensure you are managing cloud resources and associated costs sensibly.
Of financial and IT leaders surveyed by 451 research, over 80% of Respondents acknowledge poor cloud financial management could impact their business poorly. If you have ever looked at the base billing and usage monthly reports from the cloud providers, they can be very long and very difficult to interpret and use as a management tool.
If though, you employ the right management practices, there are ways to optimise your resource usage and also the company spend. The following are a set of approaches to help.
Focus Areas for Better Resource and Cost Optimisation
Cloud management will require proactive use of tools or reporting for regular, ideally daily review of the resource usage across your organisation. This may not only be IT but may also be the business direct spend. You need to ensure the net covers all these resources being used and regulate and automate where possible the monitoring of resource usage and hence cost.
The following areas are the main parts where we see most value as you put together this monitoring.
1. Identify Zero usage or Orphaned Resources
This is the low hanging fruit that will enable immediate savings. There may need to be a minimum period of inactivity that you instigate as standard for defining when to terminate that resource. Likewise, you may need to put a date on the orphaning to determine eligibility for termination.
2. Right Sizing Underutilised Resources
On average only 50% of VMs are correctly sized. Taking an instance that is too large from say 10% average CPU utilisation to a smaller instance will save on spend.
3. Plot Activity Over Time
Using a visual tool to plot computing demand over time can help to identify the peaks and troughs. This information should highlight where the resources could be augmented over a period of time or conversely scaled back through a regular period.
4. Purchase Reserve or Spot Instances
It may help to group your instances into those with a heuristic of long constant use. These groups may be good candidates that could be changed to a reserved instance type. They are discounted based on a specific long time commitment. Alternatively for spikes of non time critical processing, consider buying spot instances. These also come at a significant discount. Both AWS and Azure offer these instance categories.
5. Consider Long Term Invest or On-prem return
In what appears to be a shift back in model (to a more traditional data centre offering) some cloud providers are offering a savings plan based on a commitment to a consistent usage amount (measured in $/hour) for a one or three year term. AWS has started this offering just recently. Alternatively if your company is making the 3 year commitment, maybe you also consider a shift of this compute back to on-prem.
6. Multi-Cloud Leverage
In an effort to broker your payload services to the most efficient harbour or cloud provider, multi-cloud may be able to help. Note though that it may also lower your ability to negotiate a volume discount. Additionally having skills in the organisation to be able to work with multiple providers and also in the orchestration tools may outweigh the benefits of being able to shift (particularly if your organisation is small).
The most important of these would have to be items 1 and 2. Building this into your weekly or even daily reporting will help to be in control of the organisation cloud spend before the shock of a monthly billing report. The good news is that the data to support these decisions is available from the cloud providers on a real time basis and can be ingested into your own reporting data warehouses or tools for cloud management.
Two of the leading Cloud Management Platform tools that can help particularly in these areas of cost optimisation are Cloudability and CloudCheckr. New Relic One comes at cloud from an observability perspective for management but can also help with some of the cost optimisation management.
Companies need to spend some time setting themselves up for effective organisational governance for cloud management. From this position, cost can be managed. Clint Boulton from CIO.com says “.... it all comes back to cost. In an era of mass digital transformation, CIOs are being asked to reduce run-the-business costs while fostering more innovation. That’s pushed many head-long into cloud systems without forethought as to whether the applications will actually cost less to run in the cloud than on-premises.”
Summary
In order to make effective use of cloud, companies need to treat the financial operations governance approach differently to the way they have previously when they had their own data centre, or an IT data centre provider. The beauty of cloud is that if you have the right data and can make timely decisions, you can make real time optimisation of resource and associated cost (unlike previously with on-prem). You just need to do the homework and have this reporting management process embedded into the cycle of operations.
We have talked briefly about 6 areas that we think are important considerations for better cost management. These though go hand in hand with a broader set of governance areas for cloud that should form the backbone of your cloud operations.
Your AWS, Azure and GCP monthly news roundup
Join our list to receive The Radar delivered directly to your inbox on a monthly basis.