Posted on October 10, 2017 at 12:00 PM

---

AWS Cloud pricing pitfalls

by INVOKE Team

Most of us know that 3 key pieces of modern computing are:

1. CPU
2. Memory (RAM & Hard Disk)
3. I/O

All cloud hosting providers price the usage of these key components (either hourly (or) minute (or) seconds based on provider).

Let us explore how & what

AWS price

the usage of these components, for their cloud consumers:

In my opinion, AWS consumers can categorize these charges into 2 types.

1. Easy to understand
2. Complicated details

Easy to understand:
I call these fixed charges, because these are the charges we can calculate straight forward, without too many technical details

For example, if you want 4 core, 16gb RAM windows machine, pay X$ per hour (assuming on-demand). If you want 4 core, 16gb RAM linux machine, pay Y$ per hour (assuming on-demand).

Most of the times, CPU part is not that complicated in cloud world.

Complicated details:
Understanding billing for these complicated components (storage & IO) require little bit technical and hands-on

AWS expertise

.s

For example, storage costs of AWS snapshots. Though you take multiple snapshots of a volume, you would be charged same, IF there is no change in volume data.

Based on an answer from AWS support team :

For example, let's say you have a 3GiB volume that you put 1GiB on and take a snapshot, S1. Then let's say you put an additional 1GiB on it (so that you're using 2GiB) and take another snapshot, S2. You would be charged for the 2GiB (less after compression) of total unique data. If you deleted S1, your charges wouldn't go down, because all that data was still needed by S2. If you deleted S2, your charges would go down to 1GiB, the amount that was needed for S1.

For example, let's say you have a 3GiB volume that you put 1GiB on and take a snapshot, S1. Then let's say you put an additional 1GiB on it (so that you're using 2GiB) and take another snapshot, S2. You would be charged for the 2GiB (less after compression) of total unique data. If you deleted S1, your charges wouldn't go down, because all that data was still needed by S2. If you deleted S2, your charges would go down to 1GiB, the amount that was needed for S1.

Charges aren't tracked per snapshot. We look at the total amount of unique data for all the snapshots you have. So if you have a volume (V1), take a snapshot of it (S1), create a new volume (V2), from the snapshot and then create a snapshot (S2) of volume (V2), you will not be charged twice for the data shared between snapshots S1 and S2.

Pricing for AWS volumes would be (somewhat) opposite to snapshots, you would be charged for the storage you have provisioned (though you use small part of it).

If you are using DynamoDB resource, storage rules & pricing would be completely different. So, in summary, cloud consumers can’t simply assume every storage component will be billed SAME, pick appropriate options to optimize your cloud costs (or) talk to US to get expert cloud consulting services.

The last complicated billing component is I/O, most of the times cloud consumers overlook this cost. Based on our analysis, we found that I/O would be 10-15% for most of the consumers and with careful analysis, at least > 5% charges can be reduced.

Just because we talked about Elastic IP, I would like to bring that, EIP is free ONLY if the associated EC2 is in running state, otherwise you would be charged for EIP.

Some of the pricing details could be found here

Cloud consumers need to think about lot of these intricacies to optimize their cloud hosting costs. Looking for help to optimize your cloud costs, leave your details here, we will touch base with you.

Would like to optimize your cloud hosting costs? contact us.