Hi Zakelly, Thanks for replay, I have add some more information in the FLIP document. Please help review and let me know if further evidence is needed.
https://cwiki.apache.org/confluence/display/FLINK/%5BWIP%5DFLIP-310%3Ause+VARINT+and+ZIGZAG+to+encode+ROWDATA+in+state#id-[WIP]FLIP310:useVARINTandZIGZAGtoencodeROWDATAinstate-TestPlan [image: image.png] Zakelly Lan <zakelly....@gmail.com> 于2023年5月10日周三 02:03写道： > Hi Xiaogang Zhou, > > Thanks for driving this! > > I'm pretty interested in your verification test, could you please > provide more details? AFAIK, the performance is related to the format > of user data and the state size of the RocksDB, as well as the memory > setup (determines the proportion of IO required to process data). > > > Best regards, > Zakelly > > On Fri, May 5, 2023 at 3:42 PM Xiaogang Zhou > <xiaogang.zhou.fl...@gmail.com> wrote: > > > > Hi Guys, > > > > I have created a FLIP WIKI FLIP-310 > > < > https://cwiki.apache.org/confluence/display/FLINK/%5BWIP%5DFLIP-310%3Ause+VARINT+and+ZIGZAG+to+encode+ROWDATA+in+state > >, > > and documented my thinking about using the varint format in FLINK state > to > > improve the FLINK state performance. > > > > Would you please help review and let me know what you think? >
