nessDB Version 2.0 is nearing!

[BohuTANG]

Hi,

Version 2.0 is nearing done, the features are:

  =============

  a. Better performances on Random-Read/Write

  b. Using Small-Splittable Tree as storage engine

  c. Using fractional cascading to speed up search

  d. Cache-oblivious write-optimized structure

  e. Using fractional cascading to speed up search

Insert cost is O((log N)/B), and query is O(log N), NOW!

For more details, please to https://github.com/shuttler/nessDB

Best

BohuTANG

[han goo Kang]

Awesome~! : )

2.0 feature has a various and interested algorithm.

typing right "git fetch shuttler/v2.0 "  haha

I have a question.

you write Insert cost is (O((log N)/B)), what is 'B'?

2012년 11월 27일 화요일 오전 11시 14분 13초 UTC+9, BohuTANG 님의 말:

[Bohu TANG]

Hi,

Here, 'B' stands for one constant parameters, it's different from B-Trees Block-Size: B(4KB or 32KB, depended on systems). Since Small-Splittable Tree (SST) is cache oblivious, it's core idea is DON'T CARE the Block-Size(Parameters B, unknown to the algorithm), which mentioned here[pdf]: http://supertech.csail.mit.edu/papers/sbtree.pdf

Thanks,

BohuTANG

--

BohuTANG

"The great artist is the simplifier."--- Vincent Van Gogh

[Alex Wilson]

Is it at all possible that some kind of range or prefix query will be available in 2.0? Is it possible to do this with SSTs?

[Bohu TANG]

@Alex

I think it should be implemented, but has NOT yet put it on the agenda, sorry for I am being busy recently.

Really really welcome anyone to commit.

Thanks