Managing digital assets and instance-level metadata is critical to many company's business. It affects everything from content availability to analysis of customer usage behavior to the ability to get insights to monetization potential, and drive business innovation.
In this session Jesús will explain how companies are leveraging the advantages of a graph platform like Neo4j over traditional relational databases and other types of data and metadata stores for DAM and discuss the success stories of Scripps Networks and Adobe Behance.
3. ➔ What is DAM + Areas where
graphs add value
➔ Two DAM success stories
➔ Q&A
4. What is DAM?
4
Digital asset management (DAM) solves the problem
of storing, organising, finding, retrieving, sharing and
distributing rich-media assets (text, graphics, images,
video, audio...)
5. 5
Key Capabilities in DAM solutions
Repository and library services
Metadata and taxonomy management
Search and navigation
Lifecycle and rights management capabilities
Identity and Access Management
Production analytics
Repository and library services
Metadata and taxonomy management
Search and navigation
Lifecycle and rights management capabilities
Identity and Access Management
Production analytics
7. “All organisations are now media organisations,
based on the amount of content they create,
the audiences they deliver it to, and the stories
they tell to connect with customers...” (*)
(*) https://www.forrester.com/go?objectid=RES138067
8. Neo4j’s Property Graph
8
Capture Complexity
Flexibility
High Performance
Scale
Rich, Dynamic Graph
Model on a Native
Graph Platform
DAM Needs...
Neo4j - The worlds leading Graph Platform
13. The Model: Multiple DAMS in one graph
Broadcast Video
Related Assets
(recipes)
Production
Deliverables
14. The Model: Flexibility is key
MATCH (a:Asset)-[:PART*]-(nlv:Asset)
WHERE a.scrid = “1134”
AND nlv.type = “NLVideo”
RETURN nlv.scrid
15. Scripps Networks Asset Registry
Model Flexibility requirements
New asset types, attributes and relationships
need to be easy to create/add
Need to be able to query related data
https://www.youtube.com/watch?v=4jlK88WXm8I
Graph Characteristics
10M to 47M nodes in 3 years (x5)
Read/Write ~ 400/1
Peak 8K requests/min (AVG of 3K)
35Bill – 55Bill weekly requests
19. Solution based on Cassandra
Cluster of 48 nodes
Dataset ~50tb
Action broadcast taking 15 min (!)
Activity Feed Old Approach
20. MATCH (u:User { id: “11”})-[:FOLLOWS]->(f:User)-[p:PUBLISH]->(pr:Project)
RETURN f.name, p.date, pr ORDER BY p.date DESC
FOLLOWS
APPRECIATE
PUBLISH
FOLLOWS
Activity Feed New Graph Data Model
21. Techniques for boosting read perf on dense nodes : Tiered Relationships
MATCH (u:User { id: “11”})-[:FOLLOWS]->(f:User)-[p:LAST_200]->(pr:Project)
RETURN f.name, p.date, p.type, pr ORDER BY p.date DESC
22. Techniques for boosting read perf on dense nodes : Qualified Relationships
MATCH (u:User)-[:FOLLOWS]->(f:User)-[p:APPRECIATE_32]->(pr:Project)
RETURN f.name, pr
ORDER BY p.date DESC
https://maxdemarzi.com/
23. #cats selected Column Store Neo4j
1 .55 .21
4 1.15 .29
13 4.07 .65
Avg processing time in seconds
Improved onboarding
experience
24. ● Human maintenance hours
down well over 300%
● 1/1000 storage required
● Lower Cost
● Improved user-facing
functionality
● Easy extensibility
Behance Activity Feed
https://www.youtube.com/watch?v=bPM9hVorPSM
25. Takeaways
➔ Graph → efficient / flexible / expressive
representation of asset data
➔ Neo4j platform gives you a head start
➔ Are you ready to build/enhance your DAM
solution with Neo4j?
26. What next?
1. Go to neo4j.com and download our graph
platform.
2. Watch/Read the resources mentioned in this
webinar. Don’t take my word for it.
3. Join our online community site to learn from
others and to share your experience
https://community.neo4j.com/
4. Reach out to us for licensing info and best
practices on building DAM solutions with Neo4j