Redis: Data Cheeseburgers

A short introduction to Redis.

Definitely check out v2 of this talk! http://speakerdeck.com/u/qrush/p/redis-data-cheeseburgers-2

Given at BarCampRochester3 in 2010. http://barcamproc.org/

Transcript

1. Redis
   Data Cheeseburgers
   

   View Slide

2. Redis
   Data Cheeseburgers
   1/43 - 3%
   

   View Slide

3. Nick Quaranto
   2/43 - 5%
   

   View Slide

4. 3/43 - 7%
   

   View Slide

5. rubygems.org
   4/43 - 10%
   

   View Slide

6. @qrush
   nick@quaran.to
   litanyagainstfear.com
   5/43 - 12%
   

   View Slide

7. Redis
   6/43 - 14%
   

   View Slide

8. no
   7/43 - 17%
   

   View Slide

9. yes
   8/43 - 19%
   

   View Slide

10. persist data as you think of it
    9/43 - 21%
    

    View Slide

11. LRANGE 0 -1
    basics
    data types
    use cases
    10/43 - 24%
    

    View Slide

12. basics
    11/43 - 26%
    

    View Slide

13. history
    12/43 - 28%
    

    View Slide

14. installing
    13/43 - 31%
    

    View Slide

15. drivers
    ruby
    python
    java
    php
    node.js
    ...
    14/43 - 33%
    

    View Slide

16. memory
    keys, values
    keys must stay in memory
    values can be swapped
    really stable
    15/43 - 35%
    

    View Slide

17. speed
    ANSI C
    110,000 SETs/second
    81,000 GETs/second
    pipelining
    16/43 - 38%
    

    View Slide

18. misc
    17/43 - 40%
    

    View Slide

19. data types
    18/43 - 42%
    

    View Slide

20. strings
    like memcached
    simple data
    expiration
    renaming
    multiple get/set
    19/43 - 45%
    

    View Slide

21. > set fries done
    OK
    > get fries
    done
    > del fries
    1
    > set fries done
    OK
    > expire 60
    1
    > get fries
    nil
    20/43 - 47%
    

    View Slide

22. counters
    incr
    incrby
    decr
    decrby
    completely atomic
    21/43 - 49%
    

    View Slide

23. > incr burgers
    1
    > incrby burgers 42
    43
    > decr burgers
    42
    22/43 - 52%
    

    View Slide

24. lists
    push/pop
    length
    random access
    ranges
    23/43 - 54%
    

    View Slide

25. > rpush order burger
    1
    > rpush order hotdog
    2
    > rpush order fries
    3
    > lrange order 0 -1
    1. burger
    2. hotdog
    3. fries
    > lpop order
    burger
    > rpop order
    fries
    > lindex order 0
    hotdog
    24/43 - 56%
    

    View Slide

26. sets
    unique elements
    basic functionality from lists
    intersect
    union
    difference
    25/43 - 59%
    

    View Slide

27. > sadd meat bacon
    1
    > sadd meat turkey
    1
    > sadd toppings bacon
    1
    > sadd toppings bacon
    0
    > sinter meat toppings
    1. bacon
    > sdiff meat toppings
    1. turkey
    26/43 - 61%
    

    View Slide

28. sorted sets (zset)
    {key => float}
    atomic increments
    ranges by score or rank
    union/intersect
    27/43 - 63%
    

    View Slide

29. > zadd menu 4.99 burger
    1
    > zadd menu 2.99 shake
    1
    > zadd menu 1.99 fries
    1
    > zrank menu fries
    0
    > zscore menu fries
    1.99
    > zrange menu 0 -1
    1. fries
    2. shake
    3. burger
    > zrangebyscore menu 2 5
    1. shake
    2. burger
    > zremrangebyscore menu 2 5
    28/43 - 66%
    

    View Slide

30. hashes
    {key => value}
    atomic increments like zset
    set, get, del, exists, len
    get all keys, values, or both
    29/43 - 68%
    

    View Slide

31. > hset orders nick burger
    1
    > hset orders john fries
    1
    > hget orders nick
    burger
    > hvals orders
    1. burger
    2. fries
    > hgetall orders
    1. nick
    2. burger
    3. john
    4. fries
    30/43 - 70%
    

    View Slide

32. use cases
    31/43 - 73%
    

    View Slide

33. boggle
    32/43 - 75%
    

    View Slide

34. boggle
    game id, score = counter
    guesses, dictionary = set
    words for a game = list
    33/43 - 77%
    

    View Slide

35. def load
    $redis.del "dict"
    dict = File.read("/usr/share/dict/words"
    dict.split.each do |word|
    if word =~ /^[a-z]{3,8}$/
    $redis.sadd "dict", word.upcase
    end
    end
    end
    def word?(word)
    $redis.sismember("dict", word)
    end
    34/43 - 80%
    

    View Slide

36. queues
    35/43 - 82%
    

    View Slide

37. queues
    36/43 - 84%
    

    View Slide

38. analytics
    37/43 - 87%
    

    View Slide

39. analytics: gem downloads
    38/43 - 89%
    

    View Slide

40. on a gem download
    $redis.incr "downloads"
    $redis.incr "downloads:rubygem:rails"
    key = "downloads:version:rails-2.3.5"
    $redis.incr key
    $redis.zincrby "downloads:today", 1, key
    39/43 - 91%
    

    View Slide

41. daily rollover
    date = "2010-04-02"
    $redis.rename "dls:today", "dls:yesterday"
    dls = Hash[*$redis.zrange("dls:yesterday",
    0, -1,
    "withscores")]
    dls.each do |key, score|
    ver = versions[key]
    $redis.hincrby ver, date, score
    $redis.hincrby ver.rubygem, date, score
    end
    40/43 - 94%
    

    View Slide

42. wrapup
    41/43 - 96%
    

    View Slide

43. redis is...
    really damn fast
    extremely useful
    waiting for you!
    42/43 - 98%
    

    View Slide