Top level:
-ish Level Designs:- Secondary Indexes
- Transactions
Secondary Indexes:
- Ankita's Notes: Main design doc (see above)level design [Not up-to-date]
- /wiki/spaces/RAM/pages/6848751
- /wiki/spaces/RAM/pages/6848613
- Benchmarking Ideas
Notes for various parts:
- Ashish's Notes ( on some design aspects)
- Arjun's Notes ( on design for multi-key objects)Tasks for SLIK 0.5
- Index split and migration algo
- /wiki/spaces/RAM/pages/6848613
- Benchmarking Ideas
- /wiki/spaces/RAM/pages/6848751
- /wiki/spaces/RAM/pages/6848592
- Indexing on RamCloud experiment results (from OSDI '14 submission)
- 13205520 (readRange while concurrent update)
Btree notes:
- Survey of potential trees for implementing indexes
- Index tree issues as of Nov 2014
- SLIK in-house btree notes by Stephen
Related work:
...
- Data stores reference for indexing [Not up-to-date][Not updated
- ] General (higher level data models in other systemsNotes on some interesting related work [Not up-to-date]
Old notes (not relevant anymore):
- Tasks for SLIK 0.5
- /wiki/spaces/RAM/pages/6848592 (before writing first draft of paper)
History:
- Ankita's original list of various data model related aspects that could be worked on. From here, we decided to first focus on secondary indexes and transactions.
- Older discussions done in the group (old = 2009/2010). Note: many of the points there are not useful / relevant anymore since design of RAMCloud has changed a lot since then.