In ShannonBase, we proivde two types sync ways:(1) strong consitence; (2) async consitence; Different consistent type has different implementation.
1: strong consistence：To monitor DML operations in innodb layer, and apply the changes in sync. This makes performance impact.
2: Async consistence: To use redo log a immediate media to shipp all changes to rapid engine, and parse these redos, then apply those changes. This's friendly to distributed framework, especially on cloud evn.
In single node mode, strong is recommended, but, on cloud, async is a good choice.
Oracle In memory column store, about re-population and query
During automatic repopulation, traditional access mechanisms are available. Data is always accessible from the buffer cache or disk. Additionally, the IM column store is always transactionally consistent with the data on disk. No matter where the query accesses the data, the database always returns consistent results.
Summary
In ShannonBase, we proivde two types sync ways:(1) strong consitence; (2) async consitence; Different consistent type has different implementation.
1: strong consistence：To monitor DML operations in innodb layer, and apply the changes in sync. This makes performance impact. 2: Async consistence: To use redo log a immediate media to shipp all changes to rapid engine, and parse these redos, then apply those changes. This's friendly to distributed framework, especially on cloud evn.
In single node mode, strong is recommended, but, on cloud, async is a good choice.
Oracle In memory column store, about re-population and query