Hi, Of cause Ignite cache is more powerful than just one relational table. Cache is object oriented storage and can store a lot of data tables in one cache record.
Selecting how many caches to use is design question: - one per each table, - one per some business object (root for several tables), - some mix of business object types to be placed in one cache – all objects to be in one cache. There is one more aspect to be considered at system design. It is data collocation. Ignite is distributed system. And important thing is to reduce number of requests to different nodes for most used requests/updates. There is opportunity to collocate data at one node. Sometimes it is also possible to use replicated caches for rarely updated parameters (dictionaries). Important thing need to be remarked: Ignite uses H2 for query parsing, but not for data in memory. Cache data is stored in faster and scalable manner than in relational table. Best Regards, Dmitriy Pavlov пн, 29 мая 2017 г. в 12:06, rick_tem <[email protected]>: > As well, for companies that have a large number of tables (as does ours), > using a cache per table would probably not be ideal. > > Best, > Rick > > > > -- > View this message in context: > http://apache-ignite-users.70518.x6.nabble.com/Correct-Way-to-Store-Data-tp13163p13187.html > Sent from the Apache Ignite Users mailing list archive at Nabble.com. >
