Reactive Data Access with RxJava... and N1QL!

This talk, given at JDK.io (recording may be available through them at some point) is a whirlwind introduction to RxJava, N1QL and reactive programming in general. It also talks about resiliency and various patterns you are likely to need.

Transcript

1. Reactive Data Access with RxJava
   …and N1QL!
   Michael Nitschinger, Couchbase Inc.
   @daschl
   

   View Slide

2. ©2016 Couchbase Inc. 2
   § New challenges
   §  React to user load
   §  React to failure
   §  Be responsive under load and failure
   § Need new solutions
   §  Decoupled, event-driven architectures
   §  Optimal resource utilization
   Reactive?
   2
   

   View Slide

3. ©2016 Couchbase Inc. 3
   3
   

   View Slide

4. ©2016 Couchbase Inc. 4
   §  RxJava 1
   §  Reactive Streams (http://www.reactive-streams.org/)
   §  RxJava 2
   §  Reactor 3
   §  Akka Streams
   The Reactive Landscape
   4
   

   View Slide

5. ©2016 Couchbase Inc. 5
   Detour: N1QL
   5
   §  Declarative Query Language
   §  Extends SQL for JSON
   §  Query UI Built-In
   §  Full SDK Support
   §  Based on performant Global Secondary Indexes
   

   View Slide

6. ©2016 Couchbase Inc. 6
   Detour: N1QL
   6
   

   View Slide

7. ©2016 Couchbase Inc. 7
   Detour: N1QL
   7
   

   View Slide

8. ©2016 Couchbase Inc. 8
   Detour: Async Java Client Stack
   8
   

   View Slide

9. ©2016 Couchbase Inc. 9
   Detour: Smart Batching
   Source: http://mechanical-sympathy.blogspot.co.at/2011/10/smart-batching.html
   9
   

   View Slide

10. ©2016 Couchbase Inc. 10
    Detour: Smart Batching
    Source: http://mechanical-sympathy.blogspot.co.at/2011/10/smart-batching.html
    10
    

    View Slide

11. RxJava 101
    A Gentle Introduction
    

    View Slide

12. ©2016 Couchbase Inc. 12
    §  Java implementation for Reactive Extensions
    https://github.com/ReactiveX
    §  A library to compose asynchronous and event-driven
    programs through observable sequences.
    RxJava: Introduction
    single multiple
    sync T Iterable
    async Future Observable
    12
    

    View Slide

13. ©2016 Couchbase Inc. 13
    §  Observables are the duals of Iterables
    §  They describe both Latency and Error side effects.
    RxJava: Introduction
    event Iterable (pull) Observable (push)
    data retrieval T next() onNext(T)
    error discovery throws Exception onError(Exception)
    completion returns onCompleted()
    13
    

    View Slide

14. ©2016 Couchbase Inc. 14
    Consuming Observables
    14
    §  The Subscriber subscribes
    and receives events.
    §  A cold Observable
    starts when subscribed.
    §  onNext can be called
    0..N times
    

    View Slide

15. ©2016 Couchbase Inc. 15
    Single
    15
    §  Behaves like an Observable
    §  Emits only a single successful value or an error
    

    View Slide

16. ©2016 Couchbase Inc. 16
    Completable
    16
    §  Signals Completion or Error for a deferred computation.
    

    View Slide

17. ©2016 Couchbase Inc. 17
    RxJava: Creating Observables
    just
    17
    

    View Slide

18. ©2016 Couchbase Inc. 18
    RxJava: Creating Observables
    18
    

    View Slide

19. ©2016 Couchbase Inc. 19
    RxJava: Creating Observables
    19
    

    View Slide

20. ©2016 Couchbase Inc. 20
    RxJava: Creating Observables
    20
    

    View Slide

21. ©2016 Couchbase Inc. 21
    RxJava: Filtering Observables
    21
    

    View Slide

22. ©2016 Couchbase Inc. 22
    RxJava: Filtering Observables
    22
    

    View Slide

23. ©2016 Couchbase Inc. 23
    RxJava: Filtering Observables
    23
    

    View Slide

24. ©2016 Couchbase Inc. 24
    RxJava: Filtering Observables
    24
    

    View Slide

25. ©2016 Couchbase Inc. 25
    RxJava: Transforming Observables
    25
    

    View Slide

26. ©2016 Couchbase Inc. 26
    RxJava: Transforming Observables
    26
    

    View Slide

27. ©2016 Couchbase Inc. 27
    RxJava: Transforming Observables
    27
    

    View Slide

28. ©2016 Couchbase Inc. 28
    RxJava: Transforming Observables
    28
    

    View Slide

29. ©2016 Couchbase Inc. 29
    RxJava: Transforming Observables
    29
    

    View Slide

30. ©2016 Couchbase Inc. 30
    RxJava: Transforming Observables
    30
    

    View Slide

31. ©2016 Couchbase Inc. 31
    RxJava: Transforming Observables
    31
    

    View Slide

32. Reactive N1QL
    Cranking Queries up to Eleven!
    

    View Slide

33. ©2016 Couchbase Inc. 33
    § Simple
    §  Executes a single N1QL statement, no options available.
    § Parameterized
    §  Executes a parameterized query with positional or named params
    §  Prepared Query supported through adhoc(false)
    Query Types
    33
    

    View Slide

34. ©2016 Couchbase Inc. 34
    § (Async) N1qlQueryResult
    §  Observable rows()
    §  Observable errors()
    §  Observable info()
    §  Observable signature()
    §  Observable finalSuccess()
    §  Observable status()
    §  boolean parseSuccess()
    §  String requestId()
    §  String clientContextId()
    Query Response
    34
    

    View Slide

35. ©2016 Couchbase Inc. 35
    Simple Query
    35
    

    View Slide

36. ©2016 Couchbase Inc. 36
    Parameterized Query
    36
    § Named Params
    

    View Slide

37. ©2016 Couchbase Inc. 37
    Parametrized Query
    37
    § Positional Params
    

    View Slide

38. ©2016 Couchbase Inc. 38
    Index Handling
    38
    

    View Slide

39. ©2016 Couchbase Inc. 39
    Index Handling
    39
    

    View Slide

40. ©2016 Couchbase Inc. 40
    Conditional Index Creation
    40
    

    View Slide

41. ©2016 Couchbase Inc. 41
    Transparent Query Caching
    41
    

    View Slide

42. Error Handling
    with RxJava
    

    View Slide

43. ©2016 Couchbase Inc. 43
    § Things will go wrong, so better plan for it
    § Do not trust integration points (including the SDK)
    § Synchronous retry & fallback is too damn hard
    § RxJava provides (almost) everything you need to
    §  fallback
    §  retry
    §  fail fast
    Preparing to Fail
    43
    

    View Slide

44. ©2016 Couchbase Inc. 44
    Timeouts
    44
    § The network is unreliable
    § Servers fail
    § The SDK contains bugs
    § Always specify timeouts and deal with them!
    § The synchronous wrapper defines them for you all the time.
    

    View Slide

45. ©2016 Couchbase Inc. 45
    Timeouts: Simple
    45
    

    View Slide

46. ©2016 Couchbase Inc. 46
    Timeouts: Synchronous API
    46
    

    View Slide

47. ©2016 Couchbase Inc. 47
    Timeouts: Complex Example
    47
    

    View Slide

48. ©2016 Couchbase Inc. 48
    Timeouts: Complex Example
    48
    

    View Slide

49. ©2016 Couchbase Inc. 49
    Coordinated Retry
    49
    §  Fail fast
    §  Don’t let your system get stuck
    §  Shed load with backpressure
    § Retry
    §  immediately won’t help
    §  either linear or exponential back-off necessary
    §  have a strategy if retry also doesn’t work (Fallbacks!)
    

    View Slide

50. ©2016 Couchbase Inc. 50
    Coordinated Fallback
    50
    

    View Slide

51. ©2016 Couchbase Inc. 51
    Coordinated Retry: Builder
    51
    §  Declarative API instead of complicated retryWhen
    

    View Slide

52. Thank you.
    

    View Slide