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

Infinispan, 10 ans et toutes ses dents

Katia Aresti
November 13, 2018

Infinispan, 10 ans et toutes ses dents

Infinispan est un cache distribué/datagrid open-source qui est né en Octobre 2008… Où sommes-nous 10 ans plus tard ? Mon constat après presque 2 ans comme développeuse dans l’équipe est le suivant: beaucoup des devs méconnaissent les capacités et usages des datagrids/caches distribués, et d’Infinispan en particulier.

En quoi Infinispan peut enrichir vos applications? Quels sont les usages, bonnes pratiques, et patterns à mettre en place pour répondre aux besoins des architectures modernes ? Est-ce Infinispan Reactive? Avec quels frameworks l’utiliser, et pourquoi ? Comment déployer tout cela sur des conteneurs et les orchestrer avec kubernetes/openshift ? Que pouvez vous attendre de la version 10 qui sortira en 2019 ?

Katia Aresti

November 13, 2018
Tweet

More Decks by Katia Aresti

Other Decks in Technology

Transcript

  1. Infinispan embarqué APP-1 Users APP-2 Users APP-3 Users 1, Julien

    2, Oihana 3, Elaia 1, Julien 2, Oihana 3, Elaia 1, Julien 2, Oihana 3, Elaia Replicated
  2. Infinispan embarqué APP-1 Users APP-2 Users APP-3 Users 1, Julien

    2, Oihana 3, Elaia 1, Julien 2, Oihana 3, Elaia 1, Julien 2, Oihana 3, Elaia Replicated
  3. W ta Pr u r Reb o s Mon in

    Eve B s 2. Con f o d s ID 4. Con f o d s RE 3. Sen d es RE 1. Sen d es ID
  4. Microservices need to scale out (up and down): very elastic

    • Scaling state in the app? • State scaling or compute scaling? Deploying new version (A/B or canary) with no disruption • State? Which state • Basket, last articles seen, HTTP session etc Stateless app Stateless app Stateless app Stateless app Magic Data Stateless app The challenge
  5. Application v1 Shopping Cart S-Z / A-F Application v1Application v1

    Application v2 Shopping Cart G-L / M-R Shopping Cart M-R / S-Z Shopping Cart A-F / G-L Infinispan 1 2 3 4 https://12factor.net/fr/
  6. Application v1 Shared state Application v1 Application Shared state DC1

    Load balancer Application v1 Shared state Application v1 Application Shared state DC2 Serves customers A-M Serves customers N-Z Customers A-Z Customers A-Z Writes to A-M Writes to N-Z Benefits Low latency Move complexity of “in-memory” state clustering outside of the application complexity
  7. Cluster of microservices with dependencies • Latency accumulation What if

    one goes down µ-service 1 µ-service 2 µ-service 3 µ-service 4 One service to fail them all
  8. There is a service for that Cache A-F Application 1

    v2 Application 2 Cache G-L Cache M-R Cache S-Z Application 1 Approach
  9. 3 4 THE DATA transport.opendata.ch + sbb.ch { "x":"8290840" ,"y":"47483629"

    ,"name":"IR 1978" ,"poly":[ {"x":"8290840","y":"47483629",...} , {"x":"8290193","y":"47483647"...,"msec":"2000" , ...] }