{"id":5025,"date":"2020-09-23T11:23:25","date_gmt":"2020-09-23T11:23:25","guid":{"rendered":"https:\/\/blog.verbat.com\/?p=2467"},"modified":"2024-05-22T07:14:12","modified_gmt":"2024-05-22T07:14:12","slug":"demystifying-mongodb","status":"publish","type":"post","link":"https:\/\/www.verbat.com\/blog\/demystifying-mongodb\/","title":{"rendered":"Demystifying MongoDB"},"content":{"rendered":"\n

Database are at the core of almost all internet and enterprise applications<\/a><\/strong>. The need for scale and faster application development<\/a><\/strong> has brought in a new generation of databases loosely termed as NoSQL databases. These are in non-tabular format and save data in a different format compared to relational database whereas relational database saves data in rows and columns organized into tables or sheets.<\/p>\n\n\n\n\n\n\n\n

Often developers encounter certain conceptual and technical difficulties with relational database when it is served with an application or multiple applications written in an object-oriented programming language or style. This is because a typical application objects must be mapped onto tables. But a typical application object is rarely related to a single table. This misalignment of application layer objects to tables and rows is called an Impedance Mismatch.<\/p>\n\n\n\n

Impedance Mismatch<\/h4>\n\n\n\n

\/\/ your application code<\/strong>
class Foo {int x; string [] tags;} <\/strong> <\/p>\n\n\n\n

In the above example, you can see that the object Foo has an integer x and a string of tags. To match Foo to a relational database you need three tables, one for the main object, another for the tags and a third to relate them together.  <\/p>\n\n\n

\n
\"\"<\/figure>\n<\/div>\n\n\n

To translate between the object in memory and\nwhat is saved in the database as an application developer you are forced to\ndevelop a mapping layer or use an ORM (Object Relational Mapper). Most of the\ntime objects don\u2019t translate to tables and rows. To top it off we may have to\nuse polymorphism or inheritance. Mapping these on to a relational database can\nadd a lot of complexity to the application.<\/p>\n\n\n\n

MongoDB adopts an entirely different\napproach. There is no schema to define, no tables and no relationships between\ncollections of objects. MongoDB allows you to structure data in a way that is\nefficient for computers to process and natural and easy for humans to read.\nEach document you save can be simple or complex as your application demands.\nThis also means that they can adapt or add data when they need without worrying\nthat this minor change would break everything. \n<\/p>\n\n\n\n

Also, MongoDB differs fundamentally from legacy databases because it can coordinate multiple servers to store data.<\/p>\n\n\n

\n
\"\"<\/figure>\n<\/div>\n\n\n

Scalability Issues<\/h4>\n\n\n\n

Relational model requires a relational database engine for itself to manage writes in a special way. To ensure consistency as well as atomicity, it must lock rows and tables and permit only one writer access at a time.<\/p>\n\n\n

\n
\"\"<\/figure>\n<\/div>\n\n\n

Ensuring referential integrity across\nthe tables and rows increases the time the lock has to be in force. Higher\nlocking time implies less writes and updates per second resulting in higher\nlatency in transactions and thus slower application. Scaling out by replicating\nor sharding data to other server can make matters even complicated.<\/p>\n\n\n\n

If relational engines try to impose\nconsistency and extend these locks across networks, lock times can increase and\ntransactions latency even higher. This obviously results in a slower\napplication. <\/p>\n\n\n\n

Relational database addresses such\nissues by deformalizing tables or by relaxing consistency for\ninstance by allowing dirty reads. But this undermines the purpose behind a\nrelational database. These methods are often used by production systems for\nreporting or decision support system database where there is a tolerance for a\ncertain margin of error. <\/p>\n\n\n\n

\n

\u201cSo how does MongoDB approach these issues\u201d<\/strong><\/p>\n<\/blockquote>\n\n\n\n