First, compare the relational database and the database NoSql
1, concurrent support for
relational databases: to support concurrent transactions and through locks, high concurrency, low efficiency.
NoSql: breaking the constraints and transactional consistency traditional relational database paradigm, and therefore high concurrent performance.
2, storage and query
relational databases: the use of relational tables stored data, through SQL query
NoSql: distributed database, support massive data, mostly key store data, as well as tables / columns, documents, images, objects and more than value and other storage; query are also diverse, high query efficiency.
3, the extended embodiment
relational databases: the main hardware configuration by increasing the scale-up way to improve performance.
NoSql: database node to increase spread.
4, indexing
relational databases: B tree, hash and other
NoSql: the key index
5, Application
relational database: General art
NoSql: specific application areas, such as requiring huge amounts of data, high concurrency and high performance, scalability, and can be tolerated eventual consistency systems.
Second, the type of database NoSql
Storage type divided by
a key-value pair
representatives are Redis
2, the table / column stores
relational data is stored by row. Stored in columns is particularly suitable for a very large number of fields in the table, each inquiry from which we pick only some of the fields. Why, because the storage row, then even though we only get a small part of the field, but the system has to read out the entire line first, and then we pick the fields you want returned. Returns the content is indeed small, but has a lot of behind the consumption of performance. Because the data is stored in the inside pages, and page size is fixed, for example, is usually 8K, 16K, and so, if many fields in a row, the number of records contained in a less, in order for us to find the specified data, many will return to the page, if these pages are not in memory, you have to dispatch from the disk in, Oh, this time on the hard disk light flash in the wild, how can we not slow!
Stored in columns, then this situation is much more comfortable.
What table column stores use it? For example, the table used to store some of what environmental water quality, hydrological and meteorological indicators, I go, those indicators, and more to your cynicism.
South China is the common GBase column stores NOSQL database, like Tianjin.
3, the document
record is a document. So this says is key to storage and no not.
On behalf of mango DB (MongoDb). It is most like a relational database. Use javascript to query, cool it.
4, FIG.
Representative Neo4j.
Commonly used in mapping knowledge, artificial intelligence category bar.
5, objects, and multi-value storage
is unknown
Three, NoSql shortcomings
1, maturity is not enough, a number of key characteristics to be achieved
2, the open source database products support a limited
3, data mining and BI lack of support, many of the existing system can not be used directly NoSql
4, NoSql is a new thing, and less expert personnel
