No doubt IT cloud budgets play an important role in our cloud utilization. Although AWS cloud payment conditions are flexible enough to let us pay for what we use, matching our spending to our workload cannot be taken for granted!
So far, we have seen that managing cloud spend forms an issue of concern for 79% of all cloud-adopting companies interviewed by the Flexera 2021 State of the Cloud Report. In fact, it appears to be a top business concern that comes second only to security.
In today’s article, we will discuss the main AWS reserved instance types which can be used to match our workloads. We will see what AWS reserved instances are and how they can help us optimize our cloud budget. We will review key AWS recommendations, as we will also touch upon what other AWS cloud practitioners and analysts say.
If you would first like to learn more about AWS cost-optimization strategies in general, please take a look at our dedicated article AWS Basics: The Top 10 AWS Cost Optimization Strategies and Tools. Now, let’s get started!
What are Amazon’s Reserved Instances?
The option for purchasing cloud capacity in advance (i.e. reserving capacity) is one of the most popular ways for saving money that AWS provides to its customers. With reserving capacity, the idea of saving money is simple – the more you pay upfront, the larger the discount.
You can purchase AWS reserved capacity in three main ways – with all, partial or no up-front payment. In all three ways, you will be able to free some funds for other needs. At the same time, if you dispose of capacity that you no longer need you are enabled to sell it in the Reserved Instance Marketplace. Hence, AWS reserved instances are a great concept to consider in your overall cloud budget optimization journey.
Although investing in reserved capacity is available for a range of services in AWS, its most popular applications include:
- Amazon EC2 Reserved Instances – you are entitled to a significant discount (up to 75%) for a specific availability zone, as you may choose between standard, convertible and scheduled EC2 reserved instances depending on the regularity of your usage and needs;
- Amazon RDS Reserved Instances – you can save up to 69% of your regular on-demand costs by preliminary reserving a database instance for a term of one to three years.
Would you like to get a feel for how EC2 reserved instances translate to figures and what their related challenges involve? Then, don’t hesitate to drop us a line!
Are there any Differences between AWS Reserved Instances and AWS Saving Plans?
Both AWS reserved instances and AWS savings plans are popular ways for optimizing costs in AWS. Yet, the way they work is a little different. With AWS savings plans the way you save is by taking advantage of a discounted hourly rate in return for a commitment to regular cloud usage over a specific period.
Here is what a basic comparison between the two models further reveals:
- reserved instances can be bought and sold in the AWS Marketplace, whereas savings plans cannot;
- convertible reserved instances can let you increase your commitment without changing the term of your contract, whereas savings plans allow additions only against a new contract;
- reserved instances generally cover more services but they have narrower technical and geographic applications.
You can explore these comparative aspects in further detail here.
How Does Right-sizing Matter for Cloud Budgets?
AWS can offer plenty of cost-saving options but what matters here is the way we utilize these against our needs. And this does not happen by itself. One of the proven ways of matching our workload performance and capacity requirements to our costs is by right-sizing.
The earlier we start right-sizing, the better for our cloud budget optimization. As our cloud needs may be continuously changing, we better integrate the process into our regular activities. The AWS White Book on Right-Sizing advises that we right-size at least once monthly. Also, we can try and launch the process by following this basic three-step approach:
- Analyze your current use of cloud services by gathering sufficient performance data to identify your workload and its peaks;
- Pick up metrics that you would like to observe (the most common metrics here appear vCPU utilization, memory utilization, network utilization, and ephemeral disk use);
- Specify the strategies and tools for evaluating costs against actual usage. You can take advantage of both native AWS tools (such as CloudWatch, Cost Explorer, Trusted Advisor) and similar third-party tools.
Now, let’s see what other key tips the AWS White Books puts forward to further support us in this process.
5 Key AWS Tips for Right-sizing
Here are the 5 key AWS strategies that can help us identify right-sizing options for our EC2 and RDS reserved instances in particular:
- Use performance data – you can start your data analysis by focusing on recent data which covers at least half of the period you are looking at; then, choose the metrics you are going to monitor depending on the type of instance you are considering;
- Analyze usage needs – you could try to identify your usage pattern by distinguishing among different states of load such as a steady load, variable (but predictable) load, production and development load (which may be automatically turned off outside business hours), and temporary load;
- Turn off idle instances – you should stop paying for instances that you no longer use. If your analysis shows that you need to stop them, also check for any associated volumes, as you would continue to be charged for them if you are dealing with a database instance;
- Select the right instance family – this strategy includes the option for migrating to a different instance model within or outside the same instance family;
- Right-size database instances – you can do so by scaling your database instance up and down but you should also consider factors such as decoupling your storage and instance type, timing, etc.
You can explore these in further detail here.
Reserved instance types are one of the most popular AWS cost-saving models as they can effectively support our cloud budget optimization. One of the proven approaches to pursue the latter is right-sizing, which is best to be introduced as a regular activity in our IT operation. If you still feel new in how this works, feel free to contact us! We will be happy to be your partner in your cloud budget optimization journey.