It’s enough to divide 100 million data into about 30 tables. The data volume of a single table should not exceed 500W, otherwise the indexing efficiency will decrease. Create an index for the necessary query conditions. There is no problem with a single 100ms query. Cobar is just A framework of sub-tables and sub-databases has nothing to do with query efficiency. Hbase is a storage system, and the two cannot be confused
Improve query performance (optimized reading), generally distributed + multiple copies. 100 million records is a large amount for a single MySQL, but it is too easy for HBase. However, HBase belongs to NOSQL and is very different from MySQL in application. This should be considered carefully. There are also ElasticSearch (itself is retrieval, but it is not bad as a database) and MongoDB, these can all be considered.
100 million items, it has not yet reached the level of sub-database, the dimension of sub-table is the key, just one master and N stations can read it, and do a good job of indexing
It’s enough to divide 100 million data into about 30 tables. The data volume of a single table should not exceed 500W, otherwise the indexing efficiency will decrease. Create an index for the necessary query conditions. There is no problem with a single 100ms query. Cobar is just A framework of sub-tables and sub-databases has nothing to do with query efficiency. Hbase is a storage system, and the two cannot be confused
cobar good? Is Hbase better?
You can try mycat
Try es
Improve query performance (optimized reading), generally distributed + multiple copies. 100 million records is a large amount for a single MySQL, but it is too easy for HBase. However, HBase belongs to NOSQL and is very different from MySQL in application. This should be considered carefully. There are also ElasticSearch (itself is retrieval, but it is not bad as a database) and MongoDB, these can all be considered.
100 million items, it has not yet reached the level of sub-database, the dimension of sub-table is the key, just one master and N stations can read it, and do a good job of indexing