Guardians of Grail

cheapbag214s · Posted: Mon 3:03, 12 Aug 2013 Post subject: When to replace RDBMS-spun2

When you should replace RDBMS
A project can usually benefit from the use of a document database during the development phase from the project,[link widoczny dla zalogowanych], since you won't need to design complex entity-relation diagrams or write complex join queries. I've detailed other uses of document databases within this answer.
In case your application needs to handle very large amounts of data,[link widoczny dla zalogowanych], the event phase will probably be longer if you use a specialized NoSQL solution for example Cassandra. However, whenever your application goes into production, it will greatly benefit from the performance and scalability of Cassandra.
the applying may benefit by using a NoSQL solution that is aimed at storing data model X and perform Y operations about the data. runs like hell with a million concurrent users)?
Which kind of NoSQL database?
Key-value stores can only be queried by key in many instances. They're useful to store simple data, for example user sessions,[link widoczny dla zalogowanych], simple profile data or precomputed values and output. Even though it can be done to keep more complicated data in key-value pairs, it burdens the application with the responsibility of maintaining 'manual' indexes to be able to perform more complex queries.
Triplestores are for storing Resource Description Metadata. I don't anything about these stores, aside from what Wikipedia informs me, so you will need to do some research on that.
Column-family stores are made for storing and processing large levels of data. They are utilised by Google's internet search engine and . The information is queried by MapReduce functions. Although MapReduce functions might be hard to understand at first, the concept is very simple. Here's an analogy which (hopefully) explains the concept:
Imagine you've multiple shoe-boxes full of receipts,[link widoczny dla zalogowanych], and you want to calculate your total expenses. You invite a number of your pals over and assign a person to each shoe-box. Each person writes on the total of every receipt in the shoe-box. This method of selecting the required information is the Map part.
When a person has down on paper the totals of (a number of) his receipts, he can summarize these totals. This is the Reduce part and can be repeated multiple times until all receipts have been handled. Ultimately,[link widoczny dla zalogowanych], all your friends come together and sum up their total sums, giving you your total expenses. This is the final Reduce step.
The benefit of this approach is you can have any number of shoe-boxes and you can assign any number of individuals to a shoe-box and still end up getting the same result. Each shoe-box is visible like a server within the database's network. Each friend could be seem as a thread on the server. With MapReduce you could have your data distributed across many servers and also have each server handle part of the query, optimizing the performance of the database.
Document-oriented stores are explained in this question,[link widoczny dla zalogowanych], so I won't discuss them here.
Graph databases are for storing networks of highly connected objects,[link widoczny dla zalogowanych], such as the users on a social networking for instance. These databases are optimized for graph operations,[link widoczny dla zalogowanych], for example finding the shortest path between two nodes,[link widoczny dla zalogowanych], or finding all nodes within three hops from the current node. Such operations are very expensive on RDBMS systems or other NoSQL databases,[link widoczny dla zalogowanych], but very cheap on graph databases.
相关的主题文章：

[link widoczny dla zalogowanych]

[link widoczny dla zalogowanych]

[link widoczny dla zalogowanych]

[link widoczny dla zalogowanych]

[link widoczny dla zalogowanych]