Understanding how AWS pricing works can be difficult. In recent years, the advent of the Public Cloud has profoundly transformed IT. The development of such technology is still changing the face of business and the way companies process and store data.
The purpose of this article is to help you better understand the basics of AWS pricing to help you better understand your AWS bill. We will talk about the three main types of billing used, then we will dive into the main principles of Amazon Web Services pricing (data-transfer, on-demand, reserved instances...). Finally, you will be introduced to the advantages of the free offer offered by AWS.
Public Cloud pricing: a real advantage?
You may know this, but if in doubt, a reminder will do no harm: the public cloud brings several key advantages over a traditional on-premise architecture. To sum up quickly, migration to the cloud promises three main benefits to businesses: more flexibility, reduced time-to-market and better control of their expenses.
In reality, the last statement sounds more like a marketing promise than a tangible benefit. Yet the Cloud is really suited to optimising IT spending through a flexible pricing model that can be tailored to business needs. But for many public cloud users, understanding pricing policies and monthly bills is a real pain.
As a result, expenses are often out of control, invoices explode and the migration strategy can be questioned.
The case of AWS pricing
As the world market leader, AWS does not fall through the cracks. The great diversity of the offer, the multiplicity of services and the wide range of instances, favours the complexity of the AWS offer.
On its pricing page, Amazon Web Services describes its offering as similar to "services like running water and electricity ": you only pay for what you use. To achieve this, AWS offers three main types of billing:
- request payment
- payment based on reservation of services
At first glance, nothing too complicated. But this is without counting on the many specific features of each product category offered by AWS. With the widest range on the market, AWS inevitably offers pricing that is fuzzy and difficult to decipher.
The basic principles of AWS pricing
Concretely, as an AWS user, you will encounter three sources of billing: billing for computing power (compute), billing for data storage (storage) and billing for data transfer (data-transfer). On your invoice, you will therefore generally find three types of data relating to these three main items.
The vagueness associated with data transfers
The billing for computing power and storage is usually fixed and clearly stated when subscribing to the AWS product. You pay per hour for the computing power, depending on your usage time. You pay per GB for storage. Simple enough to understand.
On the contrary, data transfer is sometimes very difficult to decipher and can sometimes create unpleasant surprises, or even frustration. AWS is not always transparent in this respect.
To sum up briefly: you pay no surcharge for incoming data transfers or for data transfers between two AWS services in the same region. Outbound data transfers are aggregated across services and billed at a rate defined by AWS, based on your location and product region.
On your invoice, the costs related to data transfers are indicated in the "AWS Data Transfer Out" section. It is also interesting to note that the larger your transfers, the lower your cost per GB will be.
AWS's different billing models
Each AWS product has its own specificities and, therefore, its own pricing. What looks like a gas plant can be a formidable source of optimization when you master its workings.
To clarify the situation a little, here are the 5 billing models used by AWS. These 4 models are references in the AWS universe and, although each product may have different prices, they allow you to better decipher your monthly bill and better prepare your optimization strategy.
Payment on demand
On-demand billing for AWS products is synonymous with "no obligation". Simply put: you have no commitment to the product you consume. You simply pay for the computing power you use, or the amount of storage needed to store your data sets. No long-term contracts, no advance payments: you live your relationship with AWS on a day-to-day basis, according to your daily needs.
Dedicated instances (Amazon EC2 only)
When you consume computing power at AWS (or any other cloud provider), a multitude of other AWS users may also depend on this computing power without your knowledge. It is a principle of mutualisation specific to the Cloud.
Dedicated instances allow you to take advantage of your own hardware and only dedicated to your activity. The pricing of these instances depends on two things: hourly usage fees based on the instance you choose, and dedicated fees based on the region of your instance (charged by the hour, so you pay a fixed price no matter how many instances you run).
Spot Instances (Amazon EC2 only)
Amazon EC2 Spot Instances are a kind of recycling of unused instances in the AWS Cloud. They are very suitable for static and fault-tolerant applications. These Spot instances are generally very advantageous in terms of pricing, as they are marketed with discounts of up to 90% of the initial price. A good opportunity to make significant savings.
The Reserved Bodies
Well-known to AWS users, Reserved Instances are a great source of savings for large Cloud users. Concretely, you have the opportunity to book a stock of instances in advance, over a period of 1 or 3 years, with a discount of up to 75%.
This model goes against the principle of payment on demand, but allows those who are able to anticipate their needs to better manage their expenses and significantly reduce their bill.
Saving Plans are a kind of recent evolution of the Reserved Bodies. While instances only allow the reservation of very specific instances and the establishment of very precise plans, Saving Plans allow you to commit to a family of instances. Much more flexible, therefore, for companies that need more room for manoeuvre.
If you want to know more about integrating Saving Plans into your Cloud strategy, we have written an article on the subject on our blog.
The free offer: a real opportunity to get off to a good start with AWS
Going into the Cloud can sometimes generate fears and that's normal. Cloud providers are not always very transparent and the diversity of the offer often leads to confusion. To get off to a good start with AWS, the free offer is a real opportunity and a real guarantee of peace of mind.
This is not a very limited freemium version of a much more powerful tool. AWS' free offer allows you to really enter the world of the Cloud and start developing a coherent strategy over time.
To sum up, the offer is divided into two parts: on the one hand, free products reserved for new users and valid for 12 months from the date of registration; on the other hand, products open to all users, with no time limit but rather a limit on use (e.g. 200 million queries per month on AWS DynamoDB).
Pricing by AWS product family
Mastering the basics of AWS invoicing already allows you to better control your Cloud expenses and be more serene when it comes to receiving your monthly invoice. However, this does not guarantee that development processes for AWS users will be fully optimized.
In order to better understand the pricing of each AWS product, we are preparing a series of articles dedicated to each of the 4 main families of Cloud products (the most used): compute, storage, database and network.
Find progressively below the links to these various articles that will be published in the coming weeks. To stay informed of their release, we advise you to follow us on social networks!
Let's start with a big piece: 1 - Pricing of the Compute AWS range