Learn You Some Riak

An intro to Riak talk given at the Atlanta Ruby Users Group on April 11, 2012.

The focus of the talk was on some simple Riak usage with the Ruby driver, some high-level overview of how Riak works, reasons you might want to use Riak in your application, and demo off a sample toy app for pulling tweets of the Twitter Firehose for analysis using Riak MapReduce queries.

Transcript

1. Learn
   You
   Some
   Riak
   

   View Slide

2. @wfarr
   github.com/wfarr
   

   View Slide

3. Learn
   You
   Some
   Riak
   

   View Slide

4. Story
   Time
   

   View Slide

5. View Slide

6. TL; DR
   

   View Slide

7. TL; DR
   Disclaimer: You should totally read this paper as
   soon as you get home and bask in its glory.
   

   View Slide

8. Dynamo
   

   View Slide

9. Buckets,
   Keys,
   Values
   

   View Slide

10. Fault-
    Tolerant
    

    View Slide

11. Masterless
    

    View Slide

12. What is this
    magic?
    

    View Slide

13. CAP
    Theorem
    

    View Slide

14. Consistency
    

    View Slide

15. All Nodes
    See Data at
    the Same
    Time
    

    View Slide

16. Availability
    

    View Slide

17. Every DB Request
    Gets a Response
    for Success or
    Failure,
    Guaranteed
    

    View Slide

18. Partition
    Tolerance
    

    View Slide

19. Your DB keeps
    working despite
    arbitrary message
    loss or failure of a
    part of the system
    

    View Slide

20. All Good for
    Different Things
    

    View Slide

21. You Only Get To
    Have 2 of the 3
    

    View Slide

22. Dynamo Chooses
    Availability and
    Partition Tolerance
    

    View Slide

23. What
    is
    Riak?
    

    View Slide

24. Riak is a
    Dynamo
    

    View Slide

25. Bucket-
    Key: Value
    

    View Slide

26. Entries
    abc object
    bcd object
    cde object
    def object
    Logs
    abc object
    bcd object
    cde object
    def object
    

    View Slide

27. The Ring
    

    View Slide

28. View Slide

29. VNodes
    

    View Slide

30. View Slide

31. Querying
    

    View Slide

32. Map
    Reduce
    

    View Slide

33. Riak::MapReduce.new(client)
    

    View Slide

34. mr.filter("tweets") do
    matches "^testeroftests-"
    end
    

    View Slide

35. fn = "function (v) {
    return [
    JSON.parse(v.values[0].data).text
    ];
    }"
    

    View Slide

36. mr.map(fn, :keep => true)
    

    View Slide

37. mr.run
    

    View Slide

38. Search
    

    View Slide

39. Full-text
    Search
    

    View Slide

40. Lucene
    Syntax
    

    View Slide

41. client.search "tweets",
    "retweeted:true"
    

    View Slide

42. Supports
    Manual
    Indexing
    

    View Slide

43. client.index "tweets",
    {
    id: "abcde",
    text: "#webscale"
    }
    

    View Slide

44. Supports
    Auto
    Indexing
    

    View Slide

45. t = client[‘tweets’]
    t.is_indexed?
    t.enable_index!
    t.disable_index!
    

    View Slide

46. Queries
    Nodes
    Intelligently
    

    View Slide

47. Tradeoffs
    

    View Slide

48. Consistency
    Availability
    Partition Tolerance
    

    View Slide

49. Consistency
    Availability
    Partition Tolerance
    

    View Slide

50. The Good
    

    View Slide

51. Riak gets to be
    masterless
    

    View Slide

52. Riak gets to be
    fault-tolerant
    

    View Slide

53. Riak gets to be
    easy to scale
    

    View Slide

54. Riak gets to be
    easy to manage
    

    View Slide

55. The Bad
    

    View Slide

56. Riak is “only”
    eventually
    consistent
    

    View Slide

57. Understand Your
    Tradeoffs
    

    View Slide

58. Your Tradeoffs
    Might Not Be
    Someone Else’s
    

    View Slide

59. There is no
    silver bullet
    

    View Slide

60. View Slide

61. “The most boring
    database you’ll
    ever run in
    production.”
    @pharkmillups
    

    View Slide

62. Boring
    Makes
    Devs Happy
    

    View Slide

63. Boring
    Makes
    Ops Happy
    

    View Slide

64. Boring
    is
    Awesome
    

    View Slide

65. Devs
    Ops
    

    View Slide

66. Ops
    Devs
    

    View Slide

67. Example
    Use
    Cases
    

    View Slide

68. Session
    Storage
    

    View Slide

69. Private
    S3-like
    Storage
    

    View Slide

70. Huge
    Amounts
    of Rich
    Media
    

    View Slide

71. Caching
    Layer
    

    View Slide

72. Simple
    Horizontal
    Scaling
    

    View Slide

73. Logging
    Systems
    

    View Slide

74. Maybe Not
    Use
    Cases
    

    View Slide

75. Realtime
    

    View Slide

76. Replacing
    Stuff
    That Isn’t
    Broken
    

    View Slide

77. You Can
    Use
    Multiple
    Databases!
    

    View Slide

78. Who
    Already
    Uses
    Riak?
    

    View Slide

79. View Slide

80. View Slide

81. Demo
    

    View Slide

82. Let’s
    Pretend...
    

    View Slide

83. View Slide

84. Single
    Server
    

    View Slide

85. Nope
    

    View Slide

86. Sharding
    

    View Slide

87. Nope
    

    View Slide

88. “building a
    distributed
    system ass first”
    @jnewland
    

    View Slide

89. Go
    Horizontal
    

    View Slide

90. View Slide

91. Madness?
    

    View Slide

92. Nope
    just big data
    

    View Slide

93. Scenario
    

    View Slide

94. Questions?
    

    View Slide

95. Thanks!
    Will Farrington
    speakerdeck.com/u/wfarr
    github.com/wfarr/tweetscale
    

    View Slide