Roadmap 3.0 - Githubissues

tglman commented 8 years ago

Core

[ ] New Profiler and Session recording
[ ] Page compression https://github.com/orientechnologies/orientdb/issues/5383
[x] Refactor Add Database Factory API https://github.com/orientechnologies/orientdb/issues/6011
[ ] Remove Global Transaction Lock https://github.com/orientechnologies/orientdb/issues/6013
[ ] 2QCache Implementation Optimization
[x] Clean up Api and implementation for both remote+local https://github.com/orientechnologies/orientdb/issues/6011
[x] Speed up post processing of records/vertexes https://github.com/orientechnologies/orientdb/issues/6035
[x] NUMA-Aware Reader-Writer Locks https://github.com/orientechnologies/orientdb/issues/6037
[ ] Synchronous WAL writes only https://github.com/orientechnologies/orientdb/issues/5956
[x] Parallel 2-phase commit TX in the core are the basis for a more reliable and easy replication protocol
[ ] Removing of the keys in the index. We need also to change the HASH algorithm (maybe) to have less conflicts. And the index has always to check in the record for match. But this should save a lot of space = more pages in RAM = > speed
[ ] "Small Pages", that is much more than just smaller pages than 64Kb
[ ] Change in the index API to help the SQL planner to take decisions. I guess it's about stats
[ ] Incremental backup should use the WAL when present, so it would be blazing fast
[ ] Delta sync should act like the incremental backup (=scanning all the pages) when the WAL is not present, making ALWAYS possible the delta sync between nodes
[ ] The delta synch now fails if you changed the db structure, like you added an index/cluster/class/etc. We need to make this working, so (like above) it must be ALWAYS possible the delta sync between nodes
[ ] Improve index scalability

Query

[x] New Query Engine
[x] Simplify query API
Binary & Remote protocol
[ ] Serialization Optimizations https://github.com/orientechnologies/orientdb/issues/3742
[x] Remote Push + Refactor + Unit Tests https://github.com/orientechnologies/orientdb/issues/3496
[ ] Merge Embedded Ridbag (optional)
[x] Server side transaction https://github.com/orientechnologies/orientdb/issues/181
[ ] Remote serializer negative fetch plan https://github.com/orientechnologies/orientdb/issues/3820
[x] Paginated/Streaming of query result set https://github.com/orientechnologies/orientdb/issues/6012
Lucene
[x] upgrade to lucene 5.5.x or 6.x (no more filters and other things) #6069
[x] database level search index #5069
[ ] index metrics for management
[x] create functions instead of lucene operator
Plugins & Drivers
[x] TinkerPop 3 https://github.com/orientechnologies/orientdb/issues/2441
[x] JDBC: check compatibility with BI tools, implement features on demand
ETL
[x] refactor of parallel running #5381
Studio
[x] Support for Teleporter advance mapping
[x] Database level search index #5069

cwaldron commented 8 years ago

Will the extension of the clusterId to 8 bytes become part of the 3.0 roadmap? I presume this is a major feature upgrade of the Core.

saeedtabrizi commented 7 years ago

@lvca , @tglman , I thinks after adding the ICU4J to core of OrientDB , we can add a powerful supporting for INTL (globalization) features to core api and it can be add to this epic too . We can reach more performance for sorting , collation , calendar support , message formatting, plural word support . I was wondered when i saw the chart of compare ICU collation vs Java performance . it can be used in orientdb version 3 family to achieve more performance .

I hope upgrade the OrientDB with the new feature and have shiny days .