Apache Jackrabbit Oak is a new JCR implementation with a completely new architecture. Based on concepts like eventual consistency and multi-version concurrency control, and borrowing ideas from distributed version control systems and cloud-scale databases, the Oak architecture is a major leap ahead for Jackrabbit. This presentation describes the Oak architecture and shows what it means for the scalability and performance of modern content applications. Changes to existing Jackrabbit functionality are described and the migration process is explained.

1. APACHE SLING & FRIENDS TECH MEETUP
BERLIN, 22-24 SEPTEMBER 2014
Oak, the Architecture of the new Repository
Michael Dürig, Adobe Research

2. Design goals
 Scalable
 Big repositories
 Clustering
 Customisable, flexible
 OSGi friendly
adaptTo() 2014 2

3. Outline
 CRUD
 Changes
 Search
adaptTo() 2014 3

4. Tree model
a d
b c
adaptTo() 2014 4

5. Updating
?
x
a d
b c
adaptTo() 2014 5

6. MVCC
HEAD
r1: / r2: /
a d
b c
r1: /d r2: /d
r1: /a/b
r2: /a/b
r2: /d/x
adaptTo() 2014 6

7. Refresh and Garbage Collection
adaptTo() 2014 7

8. Refresh
garbage
adaptTo() 2014 8

9. Garbage collection
garbage
adaptTo() 2014 9

10. Concurrency and Conflicts
adaptTo() 2014 10

11. Concurrent updates
r2a
r1 r2b
adaptTo() 2014 11

12. Merging
merge
r2a
upates
r1 r3
r2b
adaptTo() 2014 12

13. Conflict handling: serialisation
 Fully serialised
 Fail, no concurrent update
 Partially serialised
 Concurrent conflict free updates
adaptTo() 2014 13

14. Conflict handling strategies: merging
 Partial merge
 Conflict markers, deferred resolution
 Full merge
 Need to choose victim
adaptTo() 2014 14

15. Replicas and Sharding
adaptTo() 2014 15

16. Replica and caches
master copy full replica cache
adaptTo() 2014 16

17. Sharding strategies
by path by level by hash with caching
adaptTo() 2014 17

18. Implementations
adaptTo() 2014 18

19. MicroKernel / NodeStore
 Tree / Revision model implementation
Responsible for
Clustering
Sharding
Caching
Conflict handling
Not responsible for
Validation
Access control
Search
Versioning
adaptTo() 2014 19

20. Current implementations
DocumentMK TarMK (SegmentMK)
Persistence MongoDB, JDBC Local FS
Conflict handling Partial serialisation Full serialisation
Clustering MongoDB clustering Simple failover
Sharding MongoDB sharding N/A
Node Performance Moderate High
Key use cases Large deployments (>1TB), concurrent Small/medium deployments, mostly
deployments, mostly read
adaptTo() 2014 20

21. Access Control
adaptTo() 2014 21

22. Accessible paths
a d
b c
adaptTo() 2014 22

23. xistentialism
 All paths traversable
 Node may not exist
 Decorator on NodeStore
root.getChildNode("a").exists();
root.getChildNode("a")
.getChildNode("b").exists();
⟹ false
⟹ true
adaptTo() 2014 23

24. Comparing Revisions
adaptTo() 2014 24

25. Content diff
 What changed between trees
 Cornerstone for
 Validation
 Indexing
 Observation
 …
adaptTo() 2014 25

26. What changed?
Δ
adaptTo() 2014 26

27. Example: merging
r1
r2a
r2b
r3
Δ
r1 ➞ r2a
“a” modified
“b” removed
Δ
r1 ➞ r2b
“d” modified
“x” added
adaptTo() 2014 27

28. Commit Hooks
adaptTo() 2014 28

29. Commit hooks
 Key plugin mechanism
 Higher level functionality
 Validation (node type, access control, …)
 Trigger (auto create, defaults, …)
 Updates (index, …)
adaptTo() 2014 29

30. Editing a commit
Δ Δ + x
adaptTo() 2014 30

31. Commit hooks
 Based on content diff
 pass a commit
 fail a commit
 edit a commit
 Applied in sequence
adaptTo() 2014 31

32. Type of hooks
CommitHook Editor Validator
Content diff Optional Always Always
Can modify Yes Yes No
Programming model Simple Callbacks Callbacks
Performance impact High Medium Low
adaptTo() 2014 32

33. Observers
adaptTo() 2014 33

34. Observers
 Observe changes
 After commit
 Often does a content diff
 Asynchronous
 Optionally synchronous
 Local cluster node only
adaptTo() 2014 34

35. Examples
 JCR observation
 External index update
 Cache invalidation
 Logging
adaptTo() 2014 35

36. Search
adaptTo() 2014 36

37. Query Engine
SELECT
WHERE x=y
/a//*
Parse
r
parse execute post process
Parse
r
Parser
Parser
Index
Index
Index
Travers
e
adaptTo() 2014 37

38. Index Implementations
 Property (ordered)
 Reference
 Lucene
 In-content or file system
 Solr
 Embedded or external
adaptTo() 2014 38

39. Big Picture
adaptTo() 2014 39

40. Big picture
JCR API
Oak JCR
Oak API
Oak Core
NodeStore API
MicroKernel
Plugins
adaptTo() 2014 40

41. Resources
http://jackrabbit.apache.org/oak/
adaptTo() 2014 41

42. Appendix
adaptTo() 2014 42

43. Resources
http://jackrabbit.apache.org/oak/
http://jackrabbit.apache.org/oak/docs/
https://svn.apache.org/repos/asf/jackrabbit/oak/trunk/
adaptTo() 2014 43