Deep Dive: Amazon DynamoDB

Transcript

• 1. ©2015, Amazon Web Services, Inc. or its affiliates. All rights reserved Deep Dive: Amazon DynamoDB Siva Raghupathy Principal Solutions Architect Amazon Web Services
• 2. Agenda • Tables, API, data types, indexes • Scaling • Data modeling • Scenarios and best practices • DynamoDB Streams • Reference architecture
• 3. Amazon DynamoDB • Managed NoSQL database service • Supports both document and key-value data models • Highly scalable • Consistent, single-digit millisecond latency at any scale • Highly available—3x replication • Simple and powerful API
• 4. Tables, API, Data Types
• 5. Table Table Items Attributes Hash Key Range Key Mandatory Key-value access pattern Determines data distribution Optional Model 1:N relationships Enables rich query capabilities All items for a hash key ==, <, >, >=, <= “begins with” “between” sorted results counts top/bottom N values paged responses
• 6. • CreateTable • UpdateTable • DeleteTable • DescribeTable • ListTables • GetItem • Query • Scan • BatchGetItem • PutItem • UpdateItem • DeleteItem • BatchWriteItem • ListStreams • DescribeStream • GetShardIterator • GetRecords Table and item API Stream API DynamoDB In preview
• 7. Data types • String (S) • Number (N) • Binary (B) • String Set (SS) • Number Set (NS) • Binary Set (BS) • Boolean (BOOL) • Null (NULL) • List (L) • Map (M) Used for storing nested JSON documents
• 8. 00 55 A954 AA FF Hash table • Hash key uniquely identifies an item • Hash key is used for building an unordered hash index • Table can be partitioned for scale 00 FF Id = 1 Name = Jim Hash (1) = 7B Id = 2 Name = Andy Dept = Engg Hash (2) = 48 Id = 3 Name = Kim Dept = Ops Hash (3) = CD Key Space
• 9. Partitions are three-way replicated Id = 2 Name = Andy Dept = Engg Id = 3 Name = Kim Dept = Ops Id = 1 Name = Jim Id = 2 Name = Andy Dept = Engg Id = 3 Name = Kim Dept = Ops Id = 1 Name = Jim Id = 2 Name = Andy Dept = Engg Id = 3 Name = Kim Dept = Ops Id = 1 Name = Jim Replica 1 Replica 2 Replica 3 Partition 1 Partition 2 Partition N
• 10. Hash-range table • Hash key and range key together uniquely identify an Item • Within unordered hash index, data is sorted by the range key • No limit on the number of items (∞) per hash key – Except if you have local secondary indexes 00:0 FF:∞ Hash (2) = 48 Customer# = 2 Order# = 10 Item = Pen Customer# = 2 Order# = 11 Item = Shoes Customer# = 1 Order# = 10 Item = Toy Customer# = 1 Order# = 11 Item = Boots Hash (1) = 7B Customer# = 3 Order# = 10 Item = Book Customer# = 3 Order# = 11 Item = Paper Hash (3) = CD 55 A9:∞54:∞ AA Partition 1 Partition 2 Partition 3
• 11. Table examples case class CameraRecord( cameraId: Int, // hash key ownerId: Int, subscribers: Set[Int], hoursOfRecording: Int, ... ) case class Cuepoint( cameraId: Int, // hash key timestamp: Long, // range key type: String, ... )HashKey RangeKey Value Key Segment 1234554343254 Key Segment1 1231231433235
• 12. Indexes
• 13. Local secondary index (LSI) • Alternate range key attribute • Index is local to a hash key (or partition) A1 (hash) A3 (range) A2 (table key) A1 (hash) A2 (range) A3 A4 A5 LSIs A1 (hash) A4 (range) A2 (table key) A3 (projected) Table KEYS_ONLY INCLUDE A3 A1 (hash) A5 (range) A2 (table key) A3 (projected) A4 (projected) ALL 10 GB max per hash key, i.e. LSIs limit the # of range keys!
• 14. Global secondary index (GSI) • Alternate hash (+range) key • Index is across all table hash keys (partitions) A1 (hash) A2 A3 A4 A5 GSIs A5 (hash) A4 (range) A1 (table key) A3 (projected) Table INCLUDE A3 A4 (hash) A5 (range) A1 (table key) A2 (projected) A3 (projected) ALL A2 (hash) A1 (table key) KEYS_ONLY RCUs/WCUs provisioned separately for GSIs Online indexing
• 15. How do GSI updates work? Table Primary table Primary table Primary table Primary table Global Secondary Index Client 2. Asynchronous update (in progress) If GSIs don’t have enough write capacity, table writes will be throttled!
• 16. LSI or GSI? • LSI can be modeled as a GSI • If data size in an item collection > 10 GB, use GSI • If eventual consistency is okay for your scenario, use GSI!
• 17. Scaling
• 18. Scaling • Throughput – Provision any amount of throughput to a table • Size – Add any number of items to a table • Max item size is 400 KB • LSIs limit the number of range keys due to 10 GB limit • Scaling is achieved through partitioning
• 19. Throughput • Provisioned at the table level – Write capacity units (WCUs) are measured in 1 KB per second – Read capacity units (RCUs) are measured in 4 KB per second • RCUs measure strictly consistent reads • Eventually consistent reads cost 1/2 of consistent reads • Read and write throughput limits are independent WCURCU
• 20. Partitioning math # 𝑜𝑓 𝑃𝑎𝑟𝑡𝑖𝑡𝑖𝑜𝑛𝑠 = 𝑇𝑎𝑏𝑙𝑒 𝑆𝑖𝑧𝑒 𝑖𝑛 𝐺𝐵 10 𝐺𝐵(𝑓𝑜𝑟 𝑠𝑖𝑧𝑒) # 𝑜𝑓 𝑃𝑎𝑟𝑡𝑖𝑡𝑖𝑜𝑛𝑠 (𝑓𝑜𝑟 𝑡ℎ𝑟𝑜𝑢𝑔ℎ𝑝𝑢𝑡) = 𝑅𝐶𝑈𝑓𝑜𝑟 𝑟𝑒𝑎𝑑𝑠 3000 𝑅𝐶𝑈 + 𝑊𝐶𝑈𝑓𝑜𝑟 𝑤𝑟𝑖𝑡𝑒𝑠 1000 𝑊𝐶𝑈 (𝑓𝑜𝑟 𝑡ℎ𝑟𝑜𝑢𝑔ℎ𝑝𝑢𝑡)(𝑓𝑜𝑟 𝑠𝑖𝑧𝑒)(𝑡𝑜𝑡𝑎𝑙) In the future, these details might change…
• 21. Partitioning example # 𝑜𝑓 𝑃𝑎𝑟𝑡𝑖𝑡𝑖𝑜𝑛𝑠 = 8 𝐺𝐵 10 𝐺𝐵 = 0.8 = 1 (𝑓𝑜𝑟 𝑠𝑖𝑧𝑒) # 𝑜𝑓 𝑃𝑎𝑟𝑡𝑖𝑡𝑖𝑜𝑛𝑠 (𝑓𝑜𝑟 𝑡ℎ𝑟𝑜𝑢𝑔ℎ𝑝𝑢𝑡) = 5000 𝑅𝐶𝑈 3000 𝑅𝐶𝑈 + 500 𝑊𝐶𝑈 1000 𝑊𝐶𝑈 = 2.17 = 3 Table size = 8 GB, RCUs = 5000, WCUs = 500 (𝑡𝑜𝑡𝑎𝑙) RCUs per partition = 5000/3 = 1666.67 WCUs per partition = 500/3 = 166.67 Data/partition = 10/3 = 3.33 GB RCUs and WCUs are uniformly spread across partitions
• 22. Getting the most out of DynamoDB throughput “To get the most out of DynamoDB throughput, create tables where the hash key element has a large number of distinct values, and values are requested fairly uniformly, as randomly as possible.” —DynamoDB Developer Guide • Space: access is evenly spread over the key-space • Time: requests arrive evenly spaced in time
• 23. Example: hot keys Partition Time Heat
• 24. Example: periodic spike
• 25. How does DynamoDB handle bursts? • DynamoDB saves 300 seconds of unused capacity per partition • This is used when a partition runs out of provisioned throughput due to bursts – Provided excess capacity is available at the node
• 26. Burst capacity is built-in 0 400 800 1200 1600 CapacityUnits Time Provisioned Consumed “Save up” unused capacity Consume saved up capacity Burst capacity: 300 seconds (1200 × 300 = 3600 CU)
• 27. Burst capacity may not be sufficient 0 400 800 1200 1600 CapacityUnits Time Provisioned Consumed Attempted Burst capacity: 300 seconds (1200 × 300 = 3600 CU) Throttled requests Don’t completely depend on burst capacity… provision sufficient throughput
• 28. What causes throttling? • If sustained throughput goes beyond provisioned throughput per partition • From the example before: – Table created with 5000 RCUs, 500 WCUs – RCUs per partition = 1666.67 – WCUs per partition = 166.67 – If sustained throughput > (1666 RCUs or 166 WCUs) per key or partition, DynamoDB may throttle requests • Solution: Increase provisioned throughput
• 29. What causes throttling? • Non-uniform workloads – Hot keys/hot partitions – Very large bursts • Dilution of throughout across partitions caused by mixing hot data with cold data – Use a table per time period for storing time series data so WCUs and RCUs are applied to the hot data set
• 30. Data Modeling Store data based on how you will access it!
• 31. 1:1 relationships or key-values • Use a table or GSI with a hash key • Use GetItem or BatchGetItem API Example: Given a user or email, get attributes Users Table Hash key Attributes UserId = bob Email = bob@gmail.com, JoinDate = 2011-11-15 UserId = fred Email = fred@yahoo.com, JoinDate = 2011-12-01 Users-Email-GSI Hash key Attributes Email = bob@gmail.com UserId = bob, JoinDate = 2011-11-15 Email = fred@yahoo.com UserId = fred, JoinDate = 2011-12-01
• 32. 1:N relationships or parent-children • Use a table or GSI with hash and range key • Use Query API Example: – Given a device, find all readings between epoch X, Y Device-measurements Hash Key Range key Attributes DeviceId = 1 epoch = 5513A97C Temperature = 30, pressure = 90 DeviceId = 1 epoch = 5513A9DB Temperature = 30, pressure = 90
• 33. N:M relationships • Use a table and GSI with hash and range key elements switched • Use Query API Example: Given a user, find all games. Or given a game, find all users. User-Games-Table Hash Key Range key UserId = bob GameId = Game1 UserId = fred GameId = Game2 UserId = bob GameId = Game3 Game-Users-GSI Hash Key Range key GameId = Game1 UserId = bob GameId = Game2 UserId = fred GameId = Game3 UserId = bob
• 34. Documents (JSON) • New data types (M, L, BOOL, NULL) introduced to support JSON • Document SDKs – Simple programming model – Conversion to/from JSON – Java, JavaScript, Ruby, .NET • Cannot index (S,N) elements of a JSON object stored in M – They need to be modeled as top-level table attributes to be used in LSIs and GSIs Javascript DynamoDB string S number N boolean BOOL null NULL array L object M
• 35. Rich expressions • Projection expression – Query/Get/Scan: ProductReviews.FiveStar[0] • Filter expression – Query/Scan: #V > :num (#V is a place holder for keyword VIEWS) • Conditional expression – Put/Update/DeleteItem: attribute_not_exists (#pr.FiveStar) • Update expression – UpdateItem: set Replies = Replies + :num
• 36. Scenarios and Best Practices
• 37. Event Logging Storing time series data
• 38. Time series tables Events_table_2015_April Event_id (Hash key) Timestamp (range key) Attribute1 …. Attribute N Events_table_2015_March Event_id (Hash key) Timestamp (range key) Attribute1 …. Attribute N Events_table_2015_Feburary Event_id (Hash key) Timestamp (range key) Attribute1 …. Attribute N Events_table_2015_January Event_id (Hash key) Timestamp (range key) Attribute1 …. Attribute N RCUs = 1000 WCUs = 100 RCUs = 10000 WCUs = 10000 RCUs = 100 WCUs = 1 RCUs = 10 WCUs = 1 Current table Older tables HotdataColddata Don’t mix hot and cold data; archive cold data to Amazon S3
• 39. Use a table per time period • Pre-create daily, weekly, monthly tables • Provision required throughput for current table • Writes go to the current table • Turn off (or reduce) throughput for older tables Dealing with time series data
• 40. Product Catalog Popular items (read)
• 41. Partition 1 2000 RCUs Partition K 2000 RCUs Partition M 2000 RCUs Partition 50 2000 RCU Scaling bottlenecks Product A Product B Shoppers ProductCatalog Table 100,000 𝑅𝐶𝑈 50 𝑃𝑎𝑟𝑡𝑖𝑡𝑖𝑜𝑛𝑠 ≈ 𝟐𝟎𝟎𝟎 𝑅𝐶𝑈 𝑝𝑒𝑟 𝑝𝑎𝑟𝑡𝑖𝑡𝑖𝑜𝑛 SELECT Id, Description, ... FROM ProductCatalog WHERE Id="POPULAR_PRODUCT"
• 42. RequestsPerSecond Item Primary Key Request Distribution Per Hash Key DynamoDB Requests
• 43. Partition 1 Partition 2 ProductCatalog Table User DynamoDB User SELECT Id, Description, ... FROM ProductCatalog WHERE Id="POPULAR_PRODUCT"
• 44. RequestsPerSecond Item Primary Key Request Distribution Per Hash Key DynamoDB Requests Cache Hits
• 45. Messaging App Large items Filters vs. indexes M:N Modeling—inbox and outbox
• 46. Messages Table Messages App David SELECT * FROM Messages WHERE Recipient='David' LIMIT 50 ORDER BY Date DESC Inbox SELECT * FROM Messages WHERE Sender ='David' LIMIT 50 ORDER BY Date DESC Outbox
• 47. Recipient Date Sender Message David 2014-10-02 Bob … … 48 more messages for David … David 2014-10-03 Alice … Alice 2014-09-28 Bob … Alice 2014-10-01 Carol … Large and small attributes mixed (Many more messages) David Messages Table 50 items × 256 KB each Large message bodies Attachments SELECT * FROM Messages WHERE Recipient='David' LIMIT 50 ORDER BY Date DESC Inbox
• 48. Computing inbox query cost Items evaluated by query Average item size Conversion ratio Eventually consistent reads
• 49. Recipient Date Sender Subject MsgId David 2014-10-02 Bob Hi!… afed David 2014-10-03 Alice RE: The… 3kf8 Alice 2014-09-28 Bob FW: Ok… 9d2b Alice 2014-10-01 Carol Hi!... ct7r Separate the bulk data Inbox-GSI Messages Table MsgId Body 9d2b … 3kf8 … ct7r … afed … David 1. Query Inbox-GSI: 1 RCU 2. BatchGetItem Messages: 1600 RCU (50 separate items at 256 KB) (50 sequential items at 128 bytes) Uniformly distributes large item reads
• 50. Inbox GSI
• 51. Simplified writes David PutItem { MsgId: 123, Body: ..., Recipient: Steve, Sender: David, Date: 2014-10-23, ... } Inbox Global secondary index Messages Table
• 52. Outbox Sender Outbox GSI SELECT * FROM Messages WHERE Sender ='David' LIMIT 50 ORDER BY Date DESC
• 53. Messaging app Messages Table David Inbox Global secondary index Inbox Outbox Global secondary index Outbox
• 54. • Reduce one-to-many item sizes • Configure secondary index projections • Use GSIs to model M:N relationship between sender and recipient Distribute large items Querying many large items at once InboxMessagesOutbox
• 55. Multiplayer Online Gaming Query filters vs. composite key indexes
• 56. GameId Date Host Opponent Status d9bl3 2014-10-02 David Alice DONE 72f49 2014-09-30 Alice Bob PENDING o2pnb 2014-10-08 Bob Carol IN_PROGRESS b932s 2014-10-03 Carol Bob PENDING ef9ca 2014-10-03 David Bob IN_PROGRESS Games Table Multiplayer online game data
• 57. Query for incoming game requests • DynamoDB indexes provide hash and range • What about queries for two equalities and a range? SELECT * FROM Game WHERE Opponent='Bob‘ AND Status=‘PENDING' ORDER BY Date DESC (hash) (range) (?)
• 58. Secondary Index Opponent Date GameId Status Host Alice 2014-10-02 d9bl3 DONE David Carol 2014-10-08 o2pnb IN_PROGRESS Bob Bob 2014-09-30 72f49 PENDING Alice Bob 2014-10-03 b932s PENDING Carol Bob 2014-10-03 ef9ca IN_PROGRESS David Approach 1: Query filter Bob
• 59. Secondary Index Approach 1: Query filter Bob Opponent Date GameId Status Host Alice 2014-10-02 d9bl3 DONE David Carol 2014-10-08 o2pnb IN_PROGRESS Bob Bob 2014-09-30 72f49 PENDING Alice Bob 2014-10-03 b932s PENDING Carol Bob 2014-10-03 ef9ca IN_PROGRESS David SELECT * FROM Game WHERE Opponent='Bob' ORDER BY Date DESC FILTER ON Status='PENDING' (filtered out)
• 60. Needle in a haystack Bob
• 61. • Send back less data “on the wire” • Simplify application code • Simple SQL-like expressions – AND, OR, NOT, () Use query filter Your index isn’t entirely selective
• 62. Approach 2: Composite key StatusDate DONE_2014-10-02 IN_PROGRESS_2014-10-08 IN_PROGRESS_2014-10-03 PENDING_2014-09-30 PENDING_2014-10-03 Status DONE IN_PROGRESS IN_PROGRESS PENDING PENDING Date 2014-10-02 2014-10-08 2014-10-03 2014-10-03 2014-09-30
• 63. Secondary Index Approach 2: Composite key Opponent StatusDate GameId Host Alice DONE_2014-10-02 d9bl3 David Carol IN_PROGRESS_2014-10-08 o2pnb Bob Bob IN_PROGRESS_2014-10-03 ef9ca David Bob PENDING_2014-09-30 72f49 Alice Bob PENDING_2014-10-03 b932s Carol
• 64. Opponent StatusDate GameId Host Alice DONE_2014-10-02 d9bl3 David Carol IN_PROGRESS_2014-10-08 o2pnb Bob Bob IN_PROGRESS_2014-10-03 ef9ca David Bob PENDING_2014-09-30 72f49 Alice Bob PENDING_2014-10-03 b932s Carol Secondary Index Approach 2: Composite key Bob SELECT * FROM Game WHERE Opponent='Bob' AND StatusDate BEGINS_WITH 'PENDING'
• 65. Needle in a sorted haystack Bob
• 66. Sparse indexes Id (Hash) User Game Score Date Award 1 Bob G1 1300 2012-12-23 2 Bob G1 1450 2012-12-23 3 Jay G1 1600 2012-12-24 4 Mary G1 2000 2012-10-24 Champ 5 Ryan G2 123 2012-03-10 6 Jones G2 345 2012-03-20 Game-scores-table Award (Hash) Id User Score Champ 4 Mary 2000 Award-GSI Scan sparse hash GSIs
• 67. • Concatenate attributes to form useful secondary index keys • Take advantage of sparse indexes Replace filter with indexes You want to optimize a query as much as possible Status + Date
• 68. Real-Time Voting Write-heavy items
• 69. Requirements for voting • Allow each person to vote only once • No changing votes • Real-time aggregation • Voter analytics, demographics
• 70. Real-time voting architecture AggregateVotes Table Voters RawVotes Table Voting App
• 71. Partition 1 1000 WCUs Partition K 1000 WCUs Partition M 1000 WCUs Partition N 1000 WCUs Votes Table Candidate A Candidate B Scaling bottlenecks Voters Provision 200,000 WCUs
• 72. Write sharding Candidate A_2 Candidate B_1 Candidate B_2 Candidate B_3 Candidate B_5 Candidate B_4 Candidate B_7 Candidate B_6 Candidate A_1 Candidate A_3 Candidate A_4 Candidate A_7 Candidate B_8 Candidate A_6 Candidate A_8 Candidate A_5 Voter Votes Table
• 73. Write sharding Candidate A_2 Candidate B_1 Candidate B_2 Candidate B_3 Candidate B_5 Candidate B_4 Candidate B_7 Candidate B_6 Candidate A_1 Candidate A_3 Candidate A_4 Candidate A_7 Candidate B_8 UpdateItem: “CandidateA_” + rand(0, 10) ADD 1 to Votes Candidate A_6 Candidate A_8 Candidate A_5 Voter Votes Table
• 74. Votes Table Shard aggregation Candidate A_2 Candidate B_1 Candidate B_2 Candidate B_3 Candidate B_5 Candidate B_4 Candidate B_7 Candidate B_6 Candidate A_1 Candidate A_3 Candidate A_4 Candidate A_5 Candidate A_6 Candidate A_8 Candidate A_7 Candidate B_8 Periodic Process Candidate A Total: 2.5M 1. Sum 2. Store Voter
• 75. • Trade off read cost for write scalability • Consider throughput per hash key and per partition Shard write-heavy hash keys Your write workload is not horizontally scalable
• 76. Correctness in voting UserId Candidate Date Alice A 2013-10-02 Bob B 2013-10-02 Eve B 2013-10-02 Chuck A 2013-10-02 RawVotes Table Segment Votes A_1 23 B_2 12 B_1 14 A_2 25 AggregateVotes Table Voter 1. Record vote and de-dupe; retry 2. Increment candidate counter
• 77. Correctness in aggregation? UserId Candidate Date Alice A 2013-10-02 Bob B 2013-10-02 Eve B 2013-10-02 Chuck A 2013-10-02 RawVotes Table Segment Votes A_1 23 B_2 12 B_1 14 A_2 25 AggregateVotes Table Voter
• 78. DynamoDB Streams
• 79. • Stream of updates to a table • Asynchronous • Exactly once • Strictly ordered – Per item • Highly durable • Scale with table • 24-hour lifetime • Sub-second latency DynamoDB Streams
• 80. View Type Destination Old image—before update Name = John, Destination = Mars New image—after update Name = John, Destination = Pluto Old and new images Name = John, Destination = Mars Name = John, Destination = Pluto Keys only Name = John View types UpdateItem (Name = John, Destination = Pluto)
• 81. Stream Table Partition 1 Partition 2 Partition 3 Partition 4 Partition 5 Table Shard 1 Shard 2 Shard 3 Shard 4 KCL Worker KCL Worker KCL Worker KCL Worker Amazon Kinesis Client Library Application DynamoDB Client Application Updates DynamoDB Streams and Amazon Kinesis Client Library
• 82. DynamoDB Streams Open Source Cross- Region Replication Library Asia Pacific (Sydney) EU (Ireland) Replica US East (N. Virginia) Cross-region replication
• 83. DynamoDB Streams and AWS Lambda
• 84. Real-time voting architecture (improved) AggregateVotes Table Amazon Redshift Amazon EMR Your Amazon Kinesis– Enabled App Voters RawVotes TableVoting App RawVotes DynamoDB Stream
• 85. Real-time voting architecture AggregateVotes Table Amazon Redshift Amazon EMR Your Amazon Kinesis- Enabled App Voters RawVotes TableVoting App RawVotes DynamoDB Stream
• 86. Real-time voting architecture AggregateVotes Table Amazon Redshift Amazon EMR Your Amazon Kinesis- Enabled app Voters RawVotes TableVoting App RawVotes DynamoDB Stream
• 87. Real-time voting architecture AggregateVotes Table Amazon Redshift Amazon EMR Your Amazon Kinesis– Enabled App Voters RawVotes TableVoting app RawVotes DynamoDB Stream
• 88. Real-time voting architecture AggregateVotes Table Amazon Redshift Amazon EMR Your Amazon Kinesis– Enabled App Voters RawVotes TableVoting app RawVotes DynamoDB Stream
• 89. Analytics with DynamoDB Streams • Collect and de-dupe data in DynamoDB • Aggregate data in-memory and flush periodically Performing real-time aggregation and analytics
• 90. Architecture
• 91. Reference Architecture
• 92. SAN FRANCISCO