Conversation
HyPer is a hybrid main memory database developed by database chair of TUM. Many novel techniques such as serializable MVCC, data-centric query compilation with LLVM, NUMA-aware query execution and ART index are developed, making it very efficient. All the publications about HyPer can be found in this website.
|
Thanks. What I'd do is probably to include a paper from HyPer and then link to the page from there. The links at the end are not for linking to publications by a specific research group, but links to general reading lists. |
|
I think "How Good Are Query Optimizers, Really?"[1] by Viktor Leis is a good candidate. It explores how cardinality estimation, cost model and enumeration space can influence query optimizers. And it introduces a real dataset with real queries which have many correlations, making it much more suitable for doing experiments for query optimizer than some existing datasets such as TPC-H whose data is independent and uniformly distributed. |
|
I like that paper. We should perhaps create a section on query optimizer and put that paper there (and also move the Selinger paper over). Would you be interested in doing that? |
|
Sorry it took me forever to get back on this. |
|
Don't worry. Yes, I'm glad to do that. I will make a new pull request later. |
2 participants
HyPer is a hybrid OLTP&OLAP main memory database developed by database chair of TUM. Many novel techniques such as serializable MVCC, data-centric query compilation with LLVM, NUMA-aware query execution and ART index are developed, making it very efficient. All the publications about HyPer can be found in this website.