For many tech services, the pricing models are fairly standard. Both IT and Finance departments can discuss the variables using a common language.
Take for example bandwidth for Internet services. There are factors such as the required bandwidth, quality of service, agile local supplier versus established national vendor, etc. All this is translated into a possible one-off charge followed by a predictable monthly fee.
If only cloud pricing had developed on a similar trajectory.
On-Demand Cloud Pricing Versus Cloud Reservations
Let’s examine the leading public cloud provider – Amazon Web Services (AWS). For their EC2 and RDS services Amazon offers on-demand and reserved pricing:
On-Demand: Zero upfront commitment but a relatively high rate for hourly usage.
Reservations: Upfront fee and commitment in exchange for a lower hourly rate. (Note: For heavy reservations, the hourly rate is charged even if the instance is never used.)
However, determining which pricing option will deliver the best fit for expected capacity is no walk in the park. There is a wide variety of virtual machine types (instances) , reservation levels, and other options to choose from. In order to benefit from the lower hourly fees via reservations, enterprises, and especially SMBs, have to estimate their cloud requirements ahead of time and map that to all the pricing models available.
But how can IT predict their company’s evolving cloud requirements up to a year in advance? After all, weather forecasters are never expected to predict next winter’s weather while this year’s snow is still fresh on the ground.
This is where MSPs (Managed Service Providers) and VARs (Valued Added Resellers) come in. As middlemen they provide cloud consumers with two critical functions:
- Consulting services to help the customer set up and manage their cloud deployment.
- Lower or more predictable pricing than the customer can obtain from AWS directly as explained below.
Three Pricing Models
Cloudyn works with MSPs and VARs by providing cloud optimization and cloud cost management services. While working closely with MSPs to achieve savings, we identified three main pricing models that they offer to their customers.
- Wholesale Mark-up
- Leveraged Instances
- Fixed Rate Commitment
The most traditional model is where a wholesaler buys in bulk from AWS and then adds on a mark-up to cover the cost of doing business.
Some MSPs have a very large client base and hence negotiate cheaper pricing from AWS. On top of this, they can buy reservations when warranted to lower their costs even more.
They then can turn around and sell AWS services at the same price as AWS does (plus EC2 at on-demand pricing), giving them a large profit margin. They can even offer even more competitive prices to their clients than AWS while still maintaining a healthy margin.
In addition to making a direct profit on AWS services, this model can also include additional fees for consulting, deployment, set-up and management, etc.
MSPs that do not have access to wholesale pricing are still able to provide their customers with an attractive proposition. They do this by analyzing the projected aggregate cloud usage of their clients and they then bulk purchase reserved instances. These instances are “shared” by their clientele who benefit from prices that while not as low as actual reserved instance rates, are still lower than on-demand, without the upfront payment required for reservations.
It is important to note that this type of vendor takes on risk by purchasing reservations in bulk even though it cannot guarantee that its clients’ usage will financially justify that purchase.
To protect themselves, the MSPs use a combination of mark-up pricing and fees for different services. However, it is not a simple percentage mark-up, as the overall costs are spread over the entire client base.
From the point of view of AWS, the MSP’s clients are a single client, as the MSP uses consolidated billing to allocate the costs based on each client’s actual usage.
In short, Leveraged Instances is a pricing model that enables smaller MSP players to get into the game without an official reseller relationship with AWS.
Fixed Rate Commitment
In this model, the MSP buys reservations and then charges the customer a fixed hourly rate that is lower than on-demand prices. The customer benefits from the lower rates without having to make a large, upfront “capital investment” that reservations typically require. The MSP mitigates their risk by requiring a minimum amount of usage from the customer, which covers the MSP’s upfront payment to AWS. With numerous customers using many of the same instance types, the MSP can reach break-even quite quickly and profit from the spread between the actual hourly rate paid to AWS and what they charge their customers.
Discounted Upfront Fee is a version of Leveraged Instances, as these MSPs do not purchase at wholesale prices, and they use consolidated billing to distribute costs to their customers.
The pricing of cloud services and reserved instances has evolved into a complex proposition resulting in many IT departments struggling to determine the best way forward. Therefore the consultancy services provided by MSPs and VARs is becoming a growing necessity for enterprises moving to the cloud.
Cloudyn currently works with MSPs and enterprises to manage and optimize cloud costs for AWS, Google Cloud and OpenStack.