- EMF-based tools, for instance, reverse engineering ones have reached their limits in terms of scalability
- Models up to 10^6 elements
- Robust persistence backend is a key success to model management operations (e.g. query and transformations)
- Existing backends focus more on time performance
- Provides a persistence layer for EMF-based models on top of Graph-DB (neo4j)
- Balances time performance and memory optimization
- Enables Light-weight first time loading and change tracking
- Enables dirty saving and and unloading mechanisms
- Provides adapted java code generation
Traversal: Traversing the whole resource and changing name values
Save: Saving the resource
Model size : 400 MB
Thanks to dirty save, we can maintain a good memory footprint, akin to XMI.
In time performance we notice that neo4EMF is consuming more time during traversal, this is due to the periodical dirty save to the resource. During save, Neo4EMF is consuming less time, since it only performs a cleaning of the of database – the save is done --
Traversal: Traversing the whole resource and changing name values
Save: Saving the resource
Model size : 400 MB
Thanks to dirty save, we can maintain a good memory footprint, akin to XMI.
In time performance we notice that neo4EMF is consuming more time during traversal, this is due to the periodical dirty save to the resource. During save, Neo4EMF is consuming less time, since it only performs a cleaning of the of database – the save is done --
Traversal: Traversing the whole resource and changing name values
Save: Saving the resource
Model size : 400 MB
Thanks to dirty save, we can maintain a good memory footprint, akin to XMI.
In time performance we notice that neo4EMF is consuming more time during traversal, this is due to the periodical dirty save to the resource. During save, Neo4EMF is consuming less time, since it only performs a cleaning of the of database – the save is done --
Traversal: Traversing the whole resource and changing name values
Save: Saving the resource
Model size : 400 MB
Thanks to dirty save, we can maintain a good memory footprint, akin to XMI.
In time performance we notice that neo4EMF is consuming more time during traversal, this is due to the periodical dirty save to the resource. During save, Neo4EMF is consuming less time, since it only performs a cleaning of the of database – the save is done --
Traversal: Traversing the whole resource and changing name values
Save: Saving the resource
Model size : 400 MB
Thanks to dirty save, we can maintain a good memory footprint, akin to XMI.
In time performance we notice that neo4EMF is consuming more time during traversal, this is due to the periodical dirty save to the resource. During save, Neo4EMF is consuming less time, since it only performs a cleaning of the of database – the save is done --
Traversal: Traversing the whole resource and changing name values
Save: Saving the resource
Model size : 400 MB
Thanks to dirty save, we can maintain a good memory footprint, akin to XMI.
In time performance we notice that neo4EMF is consuming more time during traversal, this is due to the periodical dirty save to the resource. During save, Neo4EMF is consuming less time, since it only performs a cleaning of the of database – the save is done --
Traversal: Traversing the whole resource and changing name values
Save: Saving the resource
Model size : 400 MB
Thanks to dirty save, we can maintain a good memory footprint, akin to XMI.
In time performance we notice that neo4EMF is consuming more time during traversal, this is due to the periodical dirty save to the resource. During save, Neo4EMF is consuming less time, since it only performs a cleaning of the of database – the save is done --
Traversal: Traversing the whole resource and changing name values
Save: Saving the resource
Model size : 400 MB
Thanks to dirty save, we can maintain a good memory footprint, akin to XMI.
In time performance we notice that neo4EMF is consuming more time during traversal, this is due to the periodical dirty save to the resource. During save, Neo4EMF is consuming less time, since it only performs a cleaning of the of database – the save is done --