Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Serverless Computing @ Jax 2017

Serverless Computing @ Jax 2017

Niko Köbler

May 10, 2017
Tweet

More Decks by Niko Köbler

Other Decks in Programming

Transcript

  1. Serverless is about the of servers. Build and run applications

    servers. management without thinking about
  2. But: Serverless is NoOps! You still have to deal with

    things like: Monitoring, Metrics, State, Dependencies, Deployment, Debugging, Network, ... not
  3. You are still for the (end) results! Keep in mind!

    No ops cost != no cost of ownership! responsible DevOps
  4. OF COMPUTING Physical Servers in datacenters Virtual Machines in datacenters

    Virtual Machines in Cloud Containers on Virtual Machines Serverless as the unit of compute EVOLUTION
  5. COMPUTE Functions are the unit of deployment and scaling. No

    machines, VMs, or containers visible in the programming model. Permanent storage lives elsewhere. Scales per request; Users cannot over- or under-provision capacity. Never pay for idle (no cold servers/containers or their costs). Implicitly fault-tolerant because functions can run anywhere. BYOC - Bring Your Own Code. Metrics and logging are a universal right. SERVERLESS MANIFESTO
  6. AWS Event Driven JVM-based languages (Java 8), JavaScript (Node.js), Python,

    C#, (more to come) Versions, Step Functions, Encrypted Environment Variables Pay as you use - in 100ms blocks (e.g. 100ms w/ 128MB memory = USD 0.000000208) LAMBDA
  7. AWS EVENTS S3 Cloud Watch Config AWS IoT Cognito SES

    API Gateway (no complete list) LAMBDA DynamoDB Cloud Formation Kinesis Code Commit SNS Echo / Alexa Cron
  8. GATEWAY "Reverse Proxy" HTTPS interface for Lambda functions and other

    services Authorization Model validation Caching, Throttling, DDoS attack protection, etc. API
  9. DYNAMO Fully Managed NoSQL Data Store Schemaless Data Model DynamoDB

    Streams Record TTL Seamless Scalabiltiy, "no limits" Pay by Read-/Write-Capacity-Units and used storage, not by time DB
  10. Provider since Environment AWS Lambda 2014 JavaScript, Python, Java, C#

    Microso AzureFunctions 2016 JavaScript, Python, Java, C#, PHP Google CloudFunctions 2016 JavaScript IBM Bluemix w/ OpenWhisk* 2014 JavaScript, Python, Swi * Apache OpenWhisk is also available for on-premise usage as OSS under ASL2 COMPETITORS
  11. BUT POWERFUL No need to run/manage Application Servers Fast results,

    minimal effort, minimal TTM(!) Low costs, minimal financial risk Automatic Scalabiltiy and Availability "out-of-the-box" SERVERLESS
  12. THANK ! ANY ? Slides: Niko Köbler So ware-Architect, Developer

    & Trainer | | YOU QUESTIONS http://bit.ly/sls-jax-2017 [email protected] www.n-k.de @dasniko http://serverlessbuch.de