Skip to content

Latest commit

 

History

History
93 lines (80 loc) · 4.79 KB

File metadata and controls

93 lines (80 loc) · 4.79 KB

Cost estimation

You are responsible for the cost of the AWS services used while running this solution.

Assumptions

General
Avg clicks/edits per model creation 40
DynamoDB entry size (KB) / model training 40
Processing Step Docker image size (GB) 0.8
Avg Lambda Duration (ms)/call 350
Lambda Avg Memory (MB) 256
Machine Learning
Processing (Feature Engineering Step)
Instance Type ml.m5.xlarge
Number instances / Job 1
Avg Mins / Job 10
Storage (MB) / Job 200
 
Model Training Step
Instance Type ml.m5.xlarge
Number of instances 1
Avg Mins / Job 12
Storage (MB) / Job 100
 
Inference (Endpoints)
Instance Type ml.m5.large
Number of instances 1
Avg Mins Active / Endpoint 60
Avg Model Size (MB) 5
 
Pipeline Events
Avg Number of Events / Job 21
 
Model Training
Total Number of Model Trainings/mo 2000

Cost

Metadata and Model training artifacts Cost
S3 $14.65 586GB
DynamoDB $0.285 0.08GB
Total $14.94
Per Model $0.0074675
Feature Engineering Step Cost
Amazon ECR (Processing Docker Image) $0.08 1 GB/MO
ML Storage $46.92 391GB
Processing Job $96.00 333 Hours
 
Model Training Step
SageMaker Training Jobs $115.20 400 Hours
ML Storage $32.93 195GB
Event Triggers $0.04 42,000 Events
Lambda Triggers From Events $0.07 42,000 Executions
 
Inference
Model Endpoints Active $288.00 2,000 Hours
ML Storage $1.68 10 GB
 
 
Total $ $289.79
Per Model $ $0.14489583
Web UI Cost
API Gateway $0.34 80,000 Requests
Lambda (handle requests) $0.13 80,000 Executions
S3 Hosting $0.43
Total $0.56 $ 0.56

Total

Monthly cost Fixed Variable Subtotal
Monthly Grand Total $15 $291 $306
Per Model $ 0.0075000 $0.1455000 $0.1530000
Annual cost Fixed Variable Subtotal
Annual Grand Total $80 $3,492 $3,672
Per Model $ 0.0900000 $1.7460000 $1.8360000



IMPORTANT: Prices are subject to change. For full details, refer to the pricing webpage for each AWS service you will be using in this solution.