Comparison: AWS Lambda vs Google Cloud Functions vs Azure Functions

Technology

February 27, 2023

Overview

AWS Lambda, Google Cloud Functions and Azure Functions are all event-driven computing services offered by major cloud providers that allow you to run code without having to provision and manage servers. Here's a comparison of the three services on different aspects.

‍

Pricing

The cost of using AWS Lambda, Google Cloud Functions, and Azure Functions can vary depending on several factors, such as the number of invocations, the duration of each execution, the memory allocation, and the number of executions. Here's a high-level comparison of the pricing models for each service (Assuming Region Asia Pacific)

Pricing Model	AWS Lambda	Google Cloud Functions	Azure Functions
Price per Request	Free for first 1M requests, then $0.20 per 1M requests	First 2M requests per month are free, then $0.40 per 1M requests	First 1M requests per month are free, then $0.20 per 1M requests
Price per GB- Second	$0.00001667 per GB-second	$0.0000025 per GB-second	$0.000016 per GB-second
Free Tier	1M free requests per month and 400,000 GB-seconds of compute time per month	2M free requests per month and 400,000 GB-seconds of compute time per month	1M free requests per month and 400,000 GB-seconds of compute time per month

‍

We can estimate the expenses by using our actual request patterns for the three service providers in the following situation:

Scenario 1: Request pattern for well-established companies such as Clappia.

Average Lambda Requests: 49.8 M per month

Average Allocated memory: 128MB per Invocation

Average Execution Time: 500ms per Invocation

Cost (in USD)	AWS Lambda	Google Cloud Functions	Azure Functions
Invocations	$9.76	$19.12	$9.80
CPU Utilization	$45.22	$66.92	$43.40
Others	nil	RAM (Gib-second) : $9.49 Minimum number of instance(in this case 1): $1.92	nil
Total (in USD)	$54.98 / month	$97.45 / month	$53.20 / month

‍

Scenario 2: Some early age Startups ( Hooli, Aviato etc 🙂)

Average Lambda Requests: 2M per month

Average Allocated memory: 128MB per Invocation

Average Execution Time: 500ms / per Invocation

Cost (in USD)	AWS Lambda	Google Cloud Functions	Azure Functions
Invocations	$0.20	$0.0	$0.20
CPU Utilization	$0.0	$0.0	$0.0
Others	nil	RAM (Gib-second) : $0.0 Minimum number of instance(in this case 1): $1.92	nil
Total (in USD)	$0.20 / month	$1.92 / month	$0.20 / month

It is crucial to keep in mind that the costs outlined above are general approximations, and the actual expenses incurred will vary depending on the specific usage patterns of each service.

‍

Language Support‍

Language	AWS Lambda	Google Cloud Functions	Azure Functions
C#	Yes	Yes	Yes
Java	Yes	Yes	Yes
Python	Yes	Yes	Yes
Go	Yes	Yes	Yes
Node.js	Yes	Yes	Yes
Ruby	Yes	No	No
PHP	Yes	No	Yes
PowerShell	Yes	No	Yes
Custom Runtime	Yes	Yes	Yes

‍

Scale

AWS Lambda, Google Cloud Functions, and Azure Functions can all automatically scale to meet the demands of incoming requests. However, the exact scale capabilities may vary between services.

Scaling Capability	AWS Lambda	Google Cloud Functions	Azure Functions
Scaling Type	Automatic	Automatic	Automatic
Scaling Limit	No limit, can request limit increase	2nd gen Cloud Functions have a default limit of 100 for both HTTP and event-driven functions. 1st gen Cloud Functions have a default limit of 3000 for event-driven functions, but there is no default limit for 1st gen HTTP Cloud Functions.	Up to 200 instances per function per region
Scaling Speed	MilliSeconds	Seconds	Seconds
Scaling Granularity	Per function or per account	Per function	Per function
Scaling Policy	Concurrent executions	Request per second	Request per instance
Warm Start	Supported	Supported	Supported

‍

Integrations

Capability	AWS Lambda	Google Cloud Functions	Azure Functions
API Gateway Integration	Yes	Yes	Yes
Compute Platform	AWS Elastic Container Service, AWS Elastic Kubernetes Service, Amazon EC2,	Google Kubernetes Engine, Compute Engine	Azure Kubernetes Service, Azure Container Instances
Messaging Service	Amazon Simple Notification Service, Amazon Simple Queue Service	Cloud Pub/Sub, Firebase Cloud Messaging	Azure Service Bus, Azure Event Hubs
Database Integration	Amazon RDS, Amazon DynamoDB,	Cloud SQL, Cloud Spanner	Azure Cosmos DB, Azure SQL Database
DevOps Integration	AWS CodePipeline, AWS CodeCommit	Cloud Build, Cloud Source Repositories	Azure DevOps, Azure Repos
Event Trigger Sources	Lambda can be integrated with over 140 AWS services directly or through the Amazon EventBridge event bus. The commonly utilized event sources for Lambda are: Amazon API Gateway Amazon CloudWatch Events Amazon CloudWatch Logs Amazon DynamoDB Streams Amazon EventBridge Amazon Kinesis Data Streams Amazon S3 Amazon SNS Amazon SQS AWS Step Functions	Cloud Functions (2nd gen) offers support for various triggers, which includes: HTTP triggers Event triggers, such as: Pub/Sub triggers Cloud Storage triggers Generalized Eventarc triggers Any event type supported by Eventarc, including 90+ event sources via Cloud Audit Logs.	Over 20 sources including Event Grid, Blob Storage, and more

‍

Monitoring and Logging

‍

AWS Lambda: AWS Lambda provides CloudWatch for monitoring and logging. CloudWatch allows you to monitor the performance of your functions and view logs for troubleshooting and debugging. You can also set up alarms to notify you of any issues with your functions.

Google Cloud Functions: Google Cloud Functions integrates with StackDriver for monitoring and logging. StackDriver provides a unified view of logs and metrics from your cloud resources, and allows you to set up alerts and dashboards for quick and easy access to key metrics.

Azure Functions: Azure Functions provides Application Insights for monitoring and logging. Application Insights provides real-time monitoring of your functions and allows you to view logs and diagnose issues. You can also set up custom alerts and dashboards to keep track of the health and performance of your functions.

‍

Execution Duration and other Constraints‍

Service	AWS Lambda	Google Cloud Functions	Azure Functions
Max Execution Duration	900 sec or 15 mins	540 sec or 9 mins	600 sec or 10mins
Max Memory	10GB	8GB	3.5GB
Max Concurrent Executions	1000	1000	Varies by plan

Note: the maximum concurrent executions on AWS Lambda and Google Cloud Functions are the default limits, which can be increased by contacting the provider's support team.

‍

Conclusion

After comparing the three services, AWS Lambda is the preferred choice for us. The cost of using AWS Lambda is significantly lower than the other two providers, especially for established companies that have high request patterns. Additionally, AWS Lambda supports a wide range of programming languages and has practically no limits on scaling. AWS Lambda also offers extensive integrations with other AWS services, making it a highly versatile service for developers. While Google Cloud Functions and Azure Functions also have their unique strengths, AWS Lambda is the most cost-effective and versatile option for event-driven computing.

‍