Understand how various AWS cost optimization tools differ

There are currently four distinct approaches to cost optimization, along with hybrid variations of these in the marketplace today. While each may deliver on the promise of ‘cost savings’, understanding the way the savings are generated may help you make better choices for your business.

OpEx to CapEx for savings

  • Savings come with lock-in contracts for use of types of cloud resources in a region for a 1-3 year time frame.
  • Most tools in this category including AWS Trusted Advisor can advise ‘reservation’ at utilizations over 10% for 1-3 years to deliver cost savings. Would a resource contributing at a 10% level be given 100% pay? We don’t think so!
  • Up-front cash payments on underutilized resources is a way to generate discounts on future OpEx spend.
  • The downside of reservation is missing out on newer technologies that are usually cheaper with better performance than older reserved ones. (Moore’s law at rising exponential levels with cloud innovation)
  • Similar to past CapEx type spend for traditional on-premise hosting, except the cloud customer gets nothing at the end of the 3-year term.

Cloud Orchestrators

  • Switch on/off various resources based on time of day.
  • Orchestrate bringing up of storage, compute and other cloud resources based on fixed schedule – example 9-5 for developers.
  • Requires rigid planning and sticking to a planned work time.
  • Savings come from switching off cloud resources during times of no use.
  • The downside here is flexibility to alter schedules when orchestrated events are at play.

Spot Instance Price Attrition

  • Spot instance pricing at 70%-80% discounts are available at times based on supply and demand in the cloud.
  • Expectation setting is to terminate workloads on a moments notice when demand increases (example 9-5 when development work occurs).
  • Custom hypervisor addons or agents allow the workload to switch from one spot instance to another with a reduced total outage of computing.
  • Savings come from the use of discounted price of spot instances with the possibility of an abrupt end to a workload.
  • The downside is abrupt shutdowns that are inevitable not suited for production workloads.

Workload Optimization

  • Right-sizing of EC2, RDS, Block storage and other storage like S3.
  • Reviewing the cost of alternative system types for cost savings
  • Savings come from running the workload on just the right type of workload and switching up when the workload is able to consume that capacity.
  • Reservations and lock in contracts are used sparingly to arrive at cost savings discounts.
  • The downside is overfitting or underfitting of resource types.

CloudSqueeze uses Workload Optimization approach along with AI based differentiators

We differ from others in the workload optimization category by the following unique features:

1. Three minute computation time with AI:

If you run less than $1K/month or use $1M/month, our AI process completes all the necessary computation across all your resources in about three minutes time. We scale up our ability to parallel task with serverless technologies so you don’t have to wait for stale data from overnight processing like a vast majority of competing options. Nothing to install.

2. Secure cloud generated data only with no agents or access to your OS:

We run all the computational jobs with just cloud generated data. No access to your private key secured systems or access to your secure data is ever sought with our approach. With just read-only access to the cloud generated data (CloudWatch), we perform all the necessary computation in three minutes.

3. AI based prediction for scale up or down

We leverage deep learning of cloud generated data to predict when you need to scale up or scale down. Our systems can correlate configuration changes and its impact on cloud resources and costs.

4. Spending changes and velocity measurements:

We compute your cloud savings opportunities daily at a set time of your choosing and provide any cost surprises as a change list, that you can assess if with your operating spend or not. Examples - changes from high network consumption, new EC2, RDS deployed accidentally, etc.

5. No cash upfront type reservations:

When the utilization of your workload matches a cloud resource we do recommend reservations for computation of savings. We price our reservations based on no up-front-cash basis. You can increase the savings by choosing to do a cash down for part or all of the year or more and further maximize savings.

6. No service based business conflicts:

Early on, as we studied MSP and cloud vendors service models we made a decision to stay focused on tools without deriving revenue from services which can tilt the product towards services. We believe our customers can make these changes easily, with our easy to follow instructions and cloud formation typescripts that can be executed by DevOps skilled resources internally. We have competent certified staff internally and we intend to train DevOps rather than develop a service business model.

7. A community and customer referral driven marketing:

Our customers are our largest promoters, so much so that we incentivize and reward our customers for word of mouth referrals. We are a community of cloud users who love the cloud and want to see reduced wastage in the cloud.

Create an account - It's free!
No credit card needed

Find us also at