Businesses investigating how to reduce EC2 costs on Amazon´s cloud compute platform (Amazon Web Services or AWS) have several options available to them. First, they can review the purchasing options for their existing resources. For example, the reassignment of instances to either Amazon´s Spot Instances or Reserved Instances pricing plan rather than On Demand can reduce EC2 costs if businesses have development projects that are not time-sensitive, or have if businesses have the organizational foresight to plan up to three years ahead. More on this below.
AWS customers should look for wasted spend on other resources that are not being used, such as orphaned volumes, snapshots, load balancers, and object storage; overprovisioned and oversized resources such as volumes, database warehouses, relational databases; idle resources whether used for non-production or production purposes; legacy resource types; and more.
Alternate ideas for how to reduce AWS costs include turning On Demand instances off when they’re not needed or not being used. This can be accomplished in several ways, such as reassigning resources to develop scheduling scripts – or asking developers to log into the AWS console each day – so that non-production AWS instances can be switched off when not required. These are not ideal solutions, as the reassignment of development resources can be counter-productive, and manual on/off processes are subject to human error. An automated on/off solution is necessary in order to make this strategy work (more on this below).
The final alternative is the implementation of software to reduce EC2 costs automatically. This too can have its disadvantages if the software does not allow you to manage existing instances without redeploying them as a new role, if it fails to give you a full tool set for account governance, or if it only tells you where the money has been spent – rather than helping your business to reduce EC2 costs.
AWS Pricing and EC2 Instance Pricing
First, let’s look at the variety of options to illustrate the complexity of Amazon EC2 pricing. EC2 is, of course, AWS’s core computing product. The “default option” for purchasing, so to speak, is On Demand. When you choose to buy AWS instances On Demand, you pay by the hour or second, with no long-term commitments. There are myriad options for customization of instance types, allowing users to optimize based on compute, memory, accelerated computing, storage, or a balance between all these options.
Here’s a brief overview of the AWS EC2 instance types:
- t2 instance type – general purpose, burstable, good for changing workloads.
- m5 instance type – general purpose, balanced, good for consistent workloads
- c5 instance type – compute optimized, high ratio of compute to memory.
- r4 instance type – memory optimized, good for in-memory database
- x1e instance type – memory optimized, good for full in-memory applications
- p3 instance type – accelerated computing, good for graphics processing and other GPU uses
- h1 instance type – storage optimized, HDD backed, balance of compute and memory
- i3 instance type – storage optimized, SDD backed, balance of compute and memory
- d2 instance type – storage optimized, highest disk ratio
Additionally, On Demand instances give you choices for operating system and region, for even more variety in prices. With this many choices, it can be difficult to predict any given EC2 instance cost in advance, and an AWS calculator may be useful. For example, a t3.xlarge on Linux in the US East (Northern Virginia) region costs $0.0823 per hour as of this writing. The same t3.xlarge, but on Windows with SQL Enterprise in Asia Pacific (Tokyo) costs $1.7912 per hour – more than 21 times the cost for the same core instance type.
These all assume you are choosing to pay by the second/hour. However, in addition to the On Demand purchasing options, there are a few other options that can be employed by savvy users to reduce costs.
Spot and Reserved Instances
AWS offers several other purchasing options, in particular, Spot and Reserved Instances.
AWS Spot Pricing Option
The AWS Spot Instance pricing option allow users to purchase spare capacity at a discount – up to 90%. Prices are set by AWS and adjust based on long-term trends in supply and demand for the capacity. Spot instances are best for fault-tolerant applications such as big data, containerized workloads, CI/CD, web services, and high-performance computing.
AWS Reserved Instance Pricing Option
AWS Reserved Instances are not actually a separate kind of instance. Rather, they’re a way to buy capacity in advance at a discount. Think of them as a credit against your account. If you purchase a reservation for a t3.large, any t3.large you provision in the applicable account will be counted against your reservation.
The AWS Reserved Instance Pricing option offers up to 75% discount compared to On Demand. However, to get the maximum discount, you must commit to 3 years of usage and pay all upfront. A lesser discount can be obtained by purchasing for a shorter duration and/or by paying for less upfront.
This can be a useful option for predictable workloads, such as production sites, if you are confident that your infrastructure needs are not going to change during the contract period. However, workloads with the ability to be turned off, such as those used for non-production or pre-production, are better off being purchased On Demand and turned off when not needed.
ParkMyCloud´s Software to Reduce EC2 Costs
ParkMyCloud´s software to reduce EC2 costs is a SaaS application that reduces costs through machine learning and automation. It is focused around taking action to reduce wasted spend, particularly that on non-production or pre-production resources that are running when they are not being used – up to 65% of the week or more – and resources that are oversized.
The platform provides a single dashboard view of all a business´s AWS instances to allow for the informed optimization of resources. The software automates the process of manually switching off existing and new non-production AWS instances when not required – eliminating the need for businesses to develop their own scheduling scripts, and the risk that non-production AWS instances will be left running when not required due to human error.
The software works by temporarily stopping – or “parking” – non-production AWS instances in accordance with a schedule. The schedule can be automatically generated based on utilization history, or manually set by the user or administrators.
To set the schedule, users connect to their AWS service through the ParkMyCloud dashboard with a dedicated IAM role that requires minimal permissions. The software locates the business´s AWS instances and makes recommendations for parking based on utilization history, names, and tags. A policy engine can automatically apply rules for schedules and team access.
Users can accept the recommendations or adjust them as necessary, after which the UI reveals a grid displaying days of the week and times of the day. The user highlights the days and hours on which the selected non-production AWS instances will not be required and accepts the schedule. The UI then displays the projected savings the business will achieve by parking the non-production AWS instances and, once the schedule takes effect, the projected savings become actual savings.
ParkMyCloud also provides “rightsizing” recommendations. Many EC2 instances are provisioned at a larger capacity than actually needed. Larger instance sizes cost more, so this is a frequent – and frequently overlooked – source of waste. All recommendations are based on actual utilization history.
How Much Can be Saved with EC2 Cost Reduction Software?
How much your business can save with EC2 cost reduction software depends on the type of instances deployed on Amazon´s cloud compute platform, their pricing plans, and how many AWS instances you are able to park and rightsizing. Typically businesses can use EC2 cost reduction software to park all development, testing, and staging instances during weekends and for at least 12 hours on weekdays – reducing EC2 costs on non-production AWS instances by 65%. The number of oversized resources varies, but just by reducing an instance by one size, you can reduce the cost by 50%. Downsizing two sizes saves 75%.
Depending on the specific circumstances of your business´s cloud computing operations, the savings can be even greater. International software development company Foster Moore uses ParkMyCloud to park non-production AWS instances by default. The company has set up a governance structure that gives team members access to their particular instances whenever they want to use them by snoozing the parking schedule. As a result, Foster Moore has reduced its total EC2 costs by more than $3.5 million.
As well as providing a means for governance and reducing costs, the implementation of EC2 cost reduction software can deliver security benefits to businesses. While non-production AWS instances are parked, they are inaccessible. Consequently there is no risk of data being hacked or compromised, and no need for network security teams to monitor them or run unnecessary virus scans – saving further resources.
Free Fourteen-Day Trial of ParkMyCloud
If your business is investigating how to reduce EC2 costs on Amazon´s cloud compute platform, we invite you to try a free thirty-day trial of ParkMyCloud. Our offer consists of fully-enabled EC2 cost reduction software, with full customer and technical support, and no obligation to continue using ParkMyCloud to reduce EC2 costs once the free trial has finished.
Automate Cloud Cost Control
in 15 Minutes
Free Trial + Free Forever Tier
- Immediate Access
- No Agents
- No Downloads
- Fully Functional
- No Obligation
- Confidentiality Guaranteed
- No Credit Card Required
There are no set up fees, no credit cards required and no contracts to sign. You simply register for an account using your email address and can start evaluating ParkMyCloud in your own environment and reducing your business´s EC2 costs within fifteen minutes. The best part of our free trial offer is all the money your business saves by using ParkMyCloud is yours to keep.
If you would like to find out more about our free offer before registering for your trial, or if you have any questions about how to reduce EC2 costs, do not hesitate to contact us. Our team will be happy to answer your questions and guide you through the features of our EC2 cost reduction software before you start using ParkMyCloud and reducing your business´s EC2 costs by parking non-production AWS instances when not required.
- Many of the ways how to reduce EC2 costs have their disadvantages.
- Parking EC2 instances with ParkMyCloud is both reliable and cost-effective.
- Our scheduling software provides total visibility of all a business´s cloud accounts.
- ParkMyCloud offers secondary benefits such as enhanced budget and capacity planning.
- Speak with us today about a free trial of ParkMyCloud and find out how much you could save.