Cost Optimisation with Cloud

Many companies tend to waste some of their IT budgets without having the recurring cost optimisation measures being followed. The benefits of cloud infrastructure have made our lives easier, but it comes with a challenge, too. The ability to scale up and down on demand has allowed resource procurement to transition from sole ownership of the finance or procurement team to stakeholders across IT, DevOps and others. This makes it difficult to track the capacity utilisation in our Infrastructure. A cost optimised infra gives us the control over cost and it continuously optimise your spend, while building scalable applications with promised SLA and low latencies to meet customer's needs.

"By following a right approach, you can save huge cost with promised performance and SLA"

The guiding principles to get started with cost optimisation in your infrastructure are as follows:

1) Identify waste in existing infrastructure 

As you begin with the practice of cost optimisation from scratch, audit the existing infrastructure first for applications/resources/instances/storage which are under-utilized. All Cloud providers offer different types of compute and storage resources with different payment options and tiers. Listing few common practices as initial steps into cost optimisation

• Understand the the top areas/components where your application is consuming the most cost
• Virtual machines which are not being used and needs to be decommissioned
• Resources which were initially spinned up for training or practice are unused now to be removed
• Unused load balancers to be removed
• Unused & unattached storages to be deleted
• Under-utilized resources where the actual capacity need is lower than what is being spinned up should be downsized.
• Removal of obsolete snapshots
• Migrating the cold data to lower cost storage.
• Exploring if hosting in a different region or availability zone could reduce costs and give high performance 
• Check the possibility of shutting down the non-prod virtual machines on weekends or other non-working days.

 2) Use Reserved Instances

By using Reserved Instances feature in cloud, we can significantly reduce the cost of VMs. Understand your infrastructure and decide if you can use reserved instances. This will be a big step towards reducing the cost of instances.

3) Autoscaling

Autoscaling feature in cloud computing lets us scale the resources, virtual machines up & down on the basis of defined parameters like load, utilization etc. Hence, better cost management. Understand the nature of the traffic that your application has and come up with the practice to scale your application resources up and down.

4) Set up alerts

By setting up the cost threshold and triggering the alarm whenever the cost exceeds we can track the cost going beyond the expected amount. Also, using the respective cost analyser tool for your underlying cloud (Azure Advisor, AWS cost explorer etc) gives us a better picture and recommendations to reduce the cost and optimise the performance.

5) Tagging the resources

Tags are just the labels which help us organize our resources. This is always a good practice to give a tag to all the resources in order to keep a track especially while auditing. Tagging can be done on the basis of owner, environment, purpose (Ex: what is the cost spent on a particular application).

6) Containerise your infrastructure

Companies are migrating towards Microservices architecture using Kubernetes, docker from monolithic architecture. The benefits of containers & orchestration are extremely wide in terms of Cost savings, Scalability, Portability, faster deployments, better performance, easier management.

7) Monitor & predict the cost every month

Cloud cost management is not a one-and-done process. As there are some low hanging fruits to harvest that can get cloud users started on the journey to intelligent cloud optimization but it is very important to keep a close watch on the cost being spent regularly and notice the deviations between the predicted amount and actual amount.

 8) Approval based access

It is always a good practice to have a gatekeeper or approver to approve the spinning up of resources that incurs cost and prevent from provisioning unapproved resources from the marketplace.

Conclusion

The proper cost optimisation is all about high performance with minimised risk and costs while giving utmost concern to the benefit of the business. Having said that industries are moving to cloud computing at a very fast pace, it is very crucial to standardize, control and optimise the cloud practices and have optimum use of services of your respective cloud providers.