response to event ) • Serverless compute service (Automatically manages the compute resources ) • Stateless (each request is independent and not related to earlier request)
a fleet of proxy machines to capture data uploads. 2. For each upload, enqueue a job to process it. 3. Provision a second fleet of machines to read and process jobs. 4. Pick a deployment solution. 5. Plan capacity 6. Accounting for fault tolerance, good long term utilization and burst capability. 7. Monitor 24x7x365 for capacity, health and security etc. 8. Keep OS and Language runtimes patched and up to date. 9. Migrate to new instance type over time
Cost effective and efficient 3 No infrastructure management Capacity, Scaling, DR, Monitoring, Logging, Security Focus on business logic Pay only for what you use ( subsecond metering, 100ms) Run code in standard language
use • Simple resource allocation ( from 128 MB up to 1.5gb) • Flexible Authorization • Built In Editor and Third party plugin (eclipse , Visual Studio) • Standard Programming Model ( process, thread, socket, AWS SDK) • Built In Monitoring and Logging • Stateless Nature
(micro-service) • Limit your function size • Limit function execution duration ( no more than 300 seconds) • Dont assume function will reuse underlying compute infrastructure • Use built-in CloudWatch monitoring of your Lambda functions to view and optimize function. • 500 mb temp storage capacity limit • Delete old Lambda functions that you are not longer using.
- stateless, trigger-based code execution Triggered by events • Direct sync and async API calls • AWS service integration • Third party triggers • Many more …. Make it easy to • Perform data-driven auditing, analysis and notification • Build back-end services that perform at scale
upload packaged zip ◦ Third party plugin (Eclipse, Visual Studio) • Standard Programing model ◦ Use process, threads and sockets normally ◦ AWS SDK built in ( NodeJs & Python) • Built-In Monitoring and logging ◦ Metrix for requests, errors and throttles ◦ Built-in logs to AWS CloudWatch Logs • Stateless ◦ Persist data using external storage ◦ No dependency or access to underlying infrastructure