One
1, based on distributed file storage
2, add more nodes under high load, can guarantee the performance of the server
3. Store data as a document
Second, Comparison
1. Stability
2, index, index in memory, can improve the performance of random read and write. If the index is not fully in memory, once the random read and write is high, it will be frequently disk exchange, MongoDB performance will drop sharply
3, occupy the space is very big, because it belongs to the typical space to change the time principle the type. Then it's a bit more disk space than the normal database, and so far it has not implemented the online compression function,
In MongoDB frequently in the data additions and deletions, if the record changes, such as the size of the data has changed, it is easy to produce some data fragments, the result of fragmentation,
One is that the index can have performance problems,
The other is in a certain amount of time after the occupied space will be inexplicably increased, so to regularly repair the database, periodically re-index, which will improve the stability and efficiency of MongoDB.
In the latest version, it has been implemented on-line compression, should be in the 2.0 version around, should be able to achieve online compression, can be performed in the background now repair database of some operations. If that's the case, it solves the current obsession
Our big problem.
4, MongoDB to the data between the transaction relationship support is relatively weak
5, operation and maintenance is not convenient
The advantages of MongoDB
1. It is suitable for the requirement model which is not clear to the database specific data format or the database data format changes frequently, and is very friendly to the developer.
2. With a distributed file system, it is easy to deploy to the server cluster.
MongoDB has a shard concept, is convenient for the server shard use. The insertion performance of each additional shard,mongodb will also grow in close multiples, and the disk capacity can be easily expanded.
3. With the support of the Map-reduce computing framework, it is also convenient for data statistics. Similar to GROUP by
MongoDB vs. mysql command the traditional relational database is usually composed of three hierarchical concepts of database, table, record,
MongoDB is made up of three levels of databases (database), Collections (collection), document objects (documents).
MongoDB is for tables in relational databases, but there is no concept of columns, rows, and relationships in the collection, which embodies the nature of pattern freedom.
Excerpt from: http://www.cnblogs.com/caihuafeng/p/5494336.html
The difference between MongoDB and MySQL