Apache Jackrabbit Oak - Scale your content repository to the cloud

http://robert.muntea.nu @rombert

Apache Jackrabbit Oak – Scale your content repository to the cloud

Apache Jackrabbit OakScale your content repository to the cloud

Robert Munteanuhttp://robert.muntea.nu

http://robert.muntea.nu @rombert

Who I am

$DAYJOB Adobe Experience

Manager− Apache Sling− Apache Jackrabbit− Apache Felix

FOSS MantisBT Mylyn Connector for

MantisBT Mylyn Connector for Review

Board Apache Sling

http://robert.muntea.nu @rombert

Before we begin - some key terms

● JCR● Apache Jackrabbit● Apache Jackrabbit Oak

http://robert.muntea.nu @rombert

Agenda

● Content● Repository● Scale● Cloud

http://robert.muntea.nu @rombert

Content – What is content?

http://robert.muntea.nu @rombert

Content – Node type defintions

[nt:resource] > mix:referenceable- jcr:encoding- jcr:mimeType mandatory- jcr:data (binary) mandatory- jcr:lastModified (date) mandatory ignore

[nt:unstructured] orderable- * (UNDEFINED)- * (UNDEFINED) multiple+ * = nt:unstructured multiple VERSION

http://robert.muntea.nu @rombert

Content - Hierarchy

blog

hello-world

images

jcr:content

some-cat.jpg

other-cat.jpg

media

cats-singing.ogg

http://robert.muntea.nu @rombert

Content - Properties

some-cat.jpg

- jcr:primaryType = app:asset- jcr:title = Some Cat- jcr:description = A longer description of this picture of a cat- jcr:created = 2014-06-03T00:00:00.000+02:00- jcr:lastUpdated = 2014-06-03T11:00:00.000+02:00- tags = [Animal, Cat, Color]- width = 400- height = 600

http://robert.muntea.nu @rombert

Repository – the Content Repository Api for Java

http://robert.muntea.nu @rombert

Repository – who implements the JCR standard

http://robert.muntea.nu @rombert

Repository – who interfaces with JCR

http://robert.muntea.nu @rombert

Repository – Apache Jackrabbit Oak

● Open Source implementation of JCR● Developed under the Jackrabbit umbrella● Not “only” fast, but built to scale from the ground up● Aims to be a successor of Jackrabbit in a later version● #7 most active project in the ASF, according to Ohloh

http://robert.muntea.nu @rombert

Repository – Digital Asset Management Sample

images

some-cat.jpg

renditions

small.jpg

annotations

initial-review

Uploaded by author

Event-based generation

Restricted by ACLs

ripple.jpg

http://robert.muntea.nu @rombert

Repository – storing configurations

etc

rendition

ripple

- orientation = /etc/rendition/ripple/options ↵/vertical- antialiasing = true- edges = /etc/rendition/ripple/options/wrap- wave type = /etc/rendition/ripple/options/ ↵wave_type- period = 20- amplitude = 5- phase shift = 0

http://robert.muntea.nu @rombert

Repository – Adding Apache Sling to the mix

images

some-cat.jpg

renditions

small.jpg

annotations

initial-review

http://oak/images/some-cat.jpghttp://oak/images/some-cat.pdf

http://oak/images/some-cat/renditions.zip

http://oak/images/some-cat/annotations.odt

ripple.jpg

http://robert.muntea.nu @rombert

Repository – Invoice Management Sample

invoices

$BIGCLIENT

2014

ABC-00014

original.pdf

renditions

jcr:content

Access restricted by ACLs

Used for previews

Changes denied by ACLs

http://robert.muntea.nu @rombert

Repository – Content Management Sample

par_2_with_image

- jcr:description = “Lorem ipsum...”

- image = /content/images/cats/tabby.jpg

blog

hello-world

jcr:content

heading_1

par_1

par_2_with_image

par_3

http://robert.muntea.nu @rombert

Scaling – Oak topologies

TarMK● Standalone● Fast● Easy to maintain● Supports failover● No horizontal scalability

MongoMK● Horizontally scales for read and write access● Supports failover● Not as fast as TarMK

http://robert.muntea.nu @rombert

Scaling – MongoMK standalone

Oak Server 1

Mongo Primary

Oak Server 2 Oak Server 3

Read and write

Read and write

Read and write

http://robert.muntea.nu @rombert

Scaling – MongoMK replica sets

Mongo Primary

Oak Server 1

Read and write

Oak Server 2

Mongo Secondary

Oak Server 3

Read

Read

Write

Mongo Primary

Read and write

Write

Mongo Secondary

http://robert.muntea.nu @rombert

Scaling – MongoMK Sharding

Oak Server 1+ mongos

Mongo Primary

Oak Server 2+ mongos

Read and write

Read and write

Oak Server 3+ mongos

Mongo Shard Primary

Oak Server 4+ mongos

Read and write

Read and write

Mongo Shard Primary

Read and write

Read and write

http://robert.muntea.nu @rombert

Scaling – BlobStore implementations

● FileBlobStore● MongoBlobStore● S3DataStore● FileDataStore● CloudBlobStore ( experimental )● MongoGridFSBlobStore ( experimental )

http://robert.muntea.nu @rombert

Cloud

● Blob storage● S3DataStore● CloudBlobStore ( experimental )

● Node storage● MongoMK

● DevOps friendliness● Easy deployment with oak-run or Apache Sling

http://robert.muntea.nu @rombert

Wrapping up

● Content● Repository● Scale● Cloud

http://robert.muntea.nu @rombert

Resources

● https://jackrabbit.apache.org/oak ● https://jackrabbit.apache.org/ ● https://sling.apache.org/

http://robert.muntea.nu @rombert

Credits

● Icons courtesy of the Tango Icon Library - http://tango.freedesktop.org/Tango_Icon_Library● Images courtesy of http://www.freedigitalphotos.net