Hello every body
My question is :
can we apply the optimization technique of " Horizontal fragmentation (partitioning)" in this new type of data stores (NoSql) just like in relational databases?
thanks at advance
in traditional (relational database) we try to see frequent queries and what they use as condition, so we partitione our tables in order to minimize response time of queries (queries scan only fragments that they need instead of full tables). Is this technique of partitioning exist in NoSql?
No, not really. I suggest reading the documentation here:
You may want to start with the link "Intelligent Driver" and also read the Getting Started Guide.
Oracle NoSQL database hash-distributes the records across the shards in the server. This is very different from the way data are organized in a relational database store. The data access patterns are also very different. I suspect that you won't have the hotspot problem with NoSQL Database.
You should take a look at the documentation as Mark suggested. Also, a quick experiment with NoSQL DB should be sufficient to verify this hypothesis. NoSQL DB is very easy to set up and use.
NoSQL DB is designed to provide very low latency access to data; due to the hash-distribution across multiple shards, there should be no hotspots except for the case where you're retrieving the same exact record repeatedly. Each shard in a NoSQL Database also has replica servers which can provide additional read capacity and scalability.
Hope this helps.