Everywhere they write abstract concepts. What does this term entail?

With this technology is connected, sort of, OLAP. But here everything is clear - these are multidimensional cubes in which data is stored. Those. Such a multidimensional array, and only.

Is it possible to provide OLTP, how can I present an OLAP cube?

  • This is what the wiki says: OLTP applications cover a wide range of tasks in many industries - automated banking systems, ERP systems (enterprise resource planning systems), banking and stock exchange operations, and in industry - registering the passage of parts on the conveyor, fixing in the statistics of visitors website visitor, automating accounting, inventory and document accounting, etc. OLTP applications typically automate structured, repetitive data processing tasks, such as inputting kazov and banking transactions. - Konst
  • Okay. This is what I read. But, let's say, if I myself want to implement it, then how do I get started? - hedgehogues
  • first decide on the purpose of your future system. As a rule, OLAP systems are built on the basis of a star schema, and such a solution is rarely found in OLTP. Also in OLTP databases, as a rule, good data normalization and tables are designed for quick recording in a transaction in multi-user mode. - Konst
  • Good. Suppose I work in a bank, I need to fill in information on customers. - hedgehogues
  • select all entities in separate tables. add / change / delete make as small a data set as possible in one short transaction - Konst

2 answers 2

Well, for example, selling tickets to the train. Every day, many people do the same operation. Main characteristics: fixed operations (the ticket looks almost the same for everyone) and a large flow of input data. It is clear that if you use the traditional (as with sites, for example) DBMS approach without doping, this will be extremely inefficient, since Most of the functionality will generally stand idle, but it also does not interfere where it is asked. The OLTP system still has, or may not, have a traditional DBMS, but already focused on speed. How can speed be increased — predefined sizes of requests, an enumerable set of fields (zone, direction, final, etc.), a relatively simple structure of the storage itself (so that it does not have to perform three-story requests), simple integrity control mechanisms (transactionality). From here and online - on the fly, transaction - all actions should be brought to the end, processing - processing :). In general, OLTP is a concept that can define, help, and suggest how best to implement fast processing of all large data streams. Something like this

  • Those. Is OLTP implemented on a regular DBMS? - hedgehogues
  • maybe, and you can offer your own solution, at least based on the file system. Understand, OLTP is a concept rather than a specific product or technology. Yes, there are solutions and recommendations for implementation, but on the whole you are not limited in terms of implementation (but it is recommended to peep how they do if :)) OLTP is like a conveyor, people will be engaged in the assembly, robots are a question of the factory owner. - user3195373 2:55 pm
  • Can advise where it is possible to look at an example of implementation on the basis of a DBMS? - hedgehogues
  • one
    Microsoft's AdventureWorks database: technet.microsoft.com/ru-ru/library/ms124659 ( v = sql.100) .aspx - Konst

OLTP is a transactional system. Those. systems focused on the rapid addition of transactions (operations) and, possibly, their changes. Those. add a document, change a document, delete a document.

The main priority of OLTP is the transaction commit rate. Reporting and analysis in OLTP is already the second priority. The application architecture is already a consequence of the specifics of this task. "Yes" - РСУБД better suited for OLTP systems. Caches and servers are not required in this case.

In OLAP , on the contrary, the main priority is to quickly obtain various analytical samples, with filters and aggregates. OLAP can say this is interactive reporting.