Architecture with NoSQL: Experiences of a SaaS startup

Presentation at WICSA 2012 Industry day.

Transcript

1. Architecture with
   NoSQL
   Experiences of a SaaS startup
   Ville Lautanala
   

   View Slide

2. View Slide

3. •Single page JavaScript front-end
   •Thousands of concurrent users
   •Three replicated databases
   •X servers on two continents
   •99.98% availability
   Facts
   

   View Slide

4. Streaming
   API
   MongoDB
   Redis
   + IRC, Twitter, RSS, and Email
   integrations
   WebSocket
   API
   Message Backend
   PostgreSQL
   Rails App
   

   View Slide

5. MongoDB
   • Document-oriented
   • JSON based queries
   • Scalable reads and
   writes
   • Replication and
   failover built-in
   • No transactions, no
   JOINs
   

   View Slide

6. SSD == Awesome
   

   View Slide

7. View Slide

8. Chef
   

   View Slide

9. Manage servers
   with code
   1. Edit recipes
   2. knife upload
   3. ???
   4. Profit!!!
   

   View Slide

10. Redis: in-memory DB
    •Key-value store
    •Data-structure server
    •Network accessible shared memory
    •Non-blocking event-bus (pub/sub)
    

    View Slide

11. Flowdock message
    processing
    Redis Queue
    Redis
    Pubsub
    MongoDB
    Flowdock Backend
    

    View Slide

12. Redis is SPoF
    

    View Slide

13. ZooKeeper
    

    View Slide

14. ZooKeeper?
    •Highly-available key-value store
    •Enables distributed coordination
    •Listen for value changes
    •Eventual consistency with quorum
    semantics
    

    View Slide

15. Node Manager
    Node Manager
    Redis Node
    Redis Node
    ZooKeeper
    Monitor
    Update
    Redis Failover
    

    View Slide

16. Zoos are kept.
    

    View Slide

17. Chef ZooKeeper
    Configuration files
    Dynamic configuration
    variables
    Server installation Failover handling
    Chef vs ZooKeeper
    

    View Slide

18. Network fail
    Flowdock
    Comcast
    Problems
    

    View Slide

19. Taking control of
    the routes
    1. Submarine communications cable
    2. 3rd party CDN/Load Balancers
    3. AWS as SSL endpoint
    

    View Slide

20. This Is How We
    Roll
    •GeoDNS chooses endpoint
    •Encrypted VPN mesh between
    servers
    •Servers near client act as SSL
    endpoint
    •Request processed near data
    

    View Slide

21. 150ms faster page
    requests from U.S.
    

    View Slide

22. Conclusions
    •MongoDB a bit like MySQL 15 years
    ago
    •Some NoSQL databases are
    suitable for distributed setups
    •Hardware and network issues
    needs to be addressed
    

    View Slide

23. Thanks!
    

    View Slide