OLAP (Online Analysis Processing) This noun is raised by EFCODD in 1993. However, the mainstream products in the market have been almost all in 1993, and some have even more than 30 years. . There are a lot of OLAP products, which will mainly involve Cognos (PowerPlay), Hyperion (EssBase), Microskeis Service, as well as products of MicroStrategy several major manufacturers. The fast gunman VS multi-faceted hand is simply considered from a cost, and Microsoft's product is the most cost-effective, Cognos and MicroStrategy are in the same horizontal line, which is more expensive than Microsoft. Hyperion (EssBase) is independent, and has also occupied the largest share of the US OLAP market, and its product prices have to be higher. From the market share, the foreign market report analysis, Microsoft, Cognos, Hyerion occupies mainstream. In China, there is currently no authoritative market report. If you only look at the projects that come into contact, there are many Cognos, and I buy EssBase is also a lot. These years are some large enterprises to build a BI project, there is enough budget, more choice of Cognos, EssBase; and MicroStrategy, entering China is not early, these years in the government, the financial industry is also built. If the development is developed, Microsoft's products are commenced with a friendly user interface and quickly get started. On the OLAP product, Microsoft still carries out this fine tradition and has a further standardization trend, developed OLE DB for OLAP and MDX (Multi-Dimensional Express multi-dimensional expression); participating in the XMLA (XML for Analysis) specification, too I want to be a data transmission standard for an OLAP server and front-end analysis. Cognos starts with a desktop OLAP, has been known for light and fast. The so-called desktop OLAP can be accessed locally with the client. Although Poweplay has already evolved into a C / S structure OLAP server, its lightweight features still continued, and provide a PowerPlay Web Explorer interface that can be concisely deployed and interactive. From the Internet, we can quickly search for many PowerPlay Web-based analytics applications. Essbase is a more complex product as an old name. The so-called complex, two layers, one is to provide a rich API, so you can fully customize development; second, it is difficult to develop, it is not easy to deploy. This is also a major cause of many users in China. Compare EssBase and PowerPlay, it will find the opposite characteristics: EssBase is complex and the PowerPlay is concise. For both, it is better to be more objective, because when you complain when you complain that EssBase is complicated, there are also some people who complain how to customize the customization of PowerPlay. This situation is actually related to the positioning of these two products, and EssBase is more focused on high-performance multi-dimensional storage services, while PowerPlay is more focused on fast multidimensional access. In other words, EssBase is like a professional camera in a professional camera. Which of the better issues is definitely there is a different answer. Of course, if you want to find a middleman between finding complex and concise, I think Microsoft's Analysis Service is such a product.
However, it is important to note that this product and SQL Server binding are tight, which is Microsoft's consistent strategy.
Hundreds of flowers vs. Unique vs. Decreasing, OLAP is generally divided into two kinds of Molp (Multi-Dimensional OLAP) and ROLAP (RELATIONAL OLAP). In addition, there is a mixed HOLAP (Hybrid OLAP). Among them, Cognos PowerPlay, Hyperion EssBase and Microsoft's Analysis Service These products are MOLAP products. This type of product extracts data from a relational database (even text file, Excel file), stored in its own database. This database is different from us to the Oracle we have seen, and DB2 is different, it is a proprietary format, and there is no standard access interface. Therefore, how these products do not have the same multidimensional storage, and the general principle is to store data in a multi-dimensional array in programming languages. The metrics are stored in the cell of the array, and each of the arrays corresponds to one dimension, where the coordinates of the dimension of the dimension. You can imagine that the cells of the multidimensional database with dimensions, the dimensions have a great relationship, and the database is also expanded as the dimension increases. Therefore, for mloAP products, multi-dimensional storage storage space, performance is naturally a key. EssBase provides a lot of optimization in this regard, but sometimes it will look too complicated. PowerPlay also provides some options such as Cube partitioning, etc., which is a relatively simple optimization method. The core function of the OLAP product is to provide multi-dimensional storage, and it is possible to convert an OLAP access operation to a request for data and returns, which most of which is sent by the user through the front end, so it is necessary to consider which front-end tools. Cognos PowerPlay is a relatively closed product, which has its own client and web explorer, you also want to use other front end to access it. Hyperion and Microsoft use an open interface to provide a rich access API, and third parties can use these APIs to access their databases. The above mentioned Microsoft developed MDX and participated in XMLA (XML For Analysis) specification, in fact, some third-party front-end tools are based on such standards and OLAP products, such as using BO Webi to connect EssBase. What's more, Microsoft's server also provides MDX to query multidimensional data, just like SQL to access the relational database. It is true that this seems to be more cool, but it is clear that although there are XMLA and MDX standards, it is not very mature, and is not the only standard. So even if there is a third-party front-end tool to access these OLAP servers, it can only be said that there are some options, truly on the front-end function, and cannot guarantee more enriched than the closed structure. If the OLAP product market is almost all occupied by MOLAP, then a company will definitely disagree, that is, MicroStrategy, it is almost the only ROLAP product that is still a certain market share. This is a very strange thing, from the first ROLAP product Metaphor to Metacube, Whitelight, MicroStrategy, these independent ROLAP vendors seem to be difficult to survive, only MicroStrategy is up to now. Is it a powerful product, or a market? It is still unknown. In principle, ROLAP stores data in relational databases, and certainly requires that the relationship model is very strict, such as complying with a star mode or snowflake mode, can define metadata, metrics, facts, and aggregation tables.
