US 12,169,487 B1
Direct cold tier upserts and partition splits enabling concurrent streaming data ingestion for time series databases
Kanishka Chaturvedi, Seattle, WA (US); Timothy A. Rath, Olympia, WA (US); Rolf Petter Skog, Seattle, WA (US); Zhihao Li, Bothell, WA (US); and Zhong Ren, Seattle, WA (US)
Assigned to Amazon Technologies, Inc., Seattle, WA (US)
Filed by Amazon Technologies, Inc., Seattle, WA (US)
Filed on Dec. 10, 2021, as Appl. No. 17/548,293.
Claims priority of provisional application 63/283,369, filed on Nov. 26, 2021.
Int. Cl. G06F 16/23 (2019.01); G06F 7/14 (2006.01); G06F 16/22 (2019.01)
CPC G06F 16/2379 (2019.01) [G06F 7/14 (2013.01); G06F 16/221 (2019.01); G06F 16/2365 (2019.01)] 20 Claims
OG exemplary drawing
 
1. A computer-implemented method comprising:
receive one or more requests to directly insert one or more data points into a cold tier of a time series database, wherein the time series database includes a first portion of data stored within a hot tier and a second portion of data stored in the cold tier in a column-oriented data format, wherein the first portion of data and the second portion of data belong to a same partition of the time series database;
inserting the one or more data points into a scratch space storage that is not part of the hot tier or the cold tier;
determining to merge the one or more data points into the cold tier;
downloading the second portion of data from a first storage location of the cold tier;
storing the second portion of data in one or more non-transitory computer-accessible storage media in a row-oriented format;
generating a third portion of data based on merging the second portion of data in the row-oriented format and the one or more data points of the scratch space storage;
storing the third portion of data, in column-oriented format, at a second storage location of the cold tier; and
updating mapping metadata for the partition to remove a reference to the first storage location and to include a reference to the second storage location.