There are all kinds of BI tools on the market, and their functional performance is very beautifully packaged, but in practical applications, we often pay more attention to simple technical features and solutions. For big data, the application trend in the future is irresistible, and many enterprises are also in need of big data analysis and processing. Below we list three mainstream BI systems on the market, and discuss the characteristics of "big data", mainly with Hadoop, The connection and performance of Spark and multi-dimensional analytical databases.
Tableau 's Big Data Strategy
1. Currently, Tableau's applicable big data ecosystem connections include:
-
Hadoop: Cloudera Impala and Hive, Hortonworks Hive, MapR Hive, Amazon EMR with Impala and Hive, Pivotal HAWQ, IBM BigInsights
-
NoSQL：MarkLogic、Datastax
-
Spark：Apache Spark SQL
-
Multidimensional Analytical Databases: Teradata Aster, HP Vertica, SAP Hana, SAP Sybase, Pivotal Greenplum
2. Tableau can connect to data sources in real-time, or bring them into memory. Live connections are great when it comes to quick and interactive query analysis. However, due to the in-memory analysis, the hardware requirements are relatively high, and the efficiency will be relatively low when the amount of data is large.
3. Big data self-service visualization for business users. Business users can visualize their data using drag-and-drop operations without writing complex SQL, Java code or MapReduce jobs. Tableau simplifies the task of analyzing data, allowing users to uncover visible insights from data faster than ever before.
FineBI 's dual sets of big data solutions
1. The big data ecosystem connections supported by FineBI include
-
Big data platforms: Vertica, GreenPlum, Apache Kylin, Impala/hive, hive, Starlink,
-
NO SQL database: MongoDB
-
Multidimensional analytical databases: Teradata, SAP Hana, Greenplum, SAP Sybase, Essbase, SSAS
2. FineBI supports two data connection and processing modes, FineIndex and FineDirect, to deal with different data schemes.
-
FineDirect is dealing with some results that need to be displayed, such as flow analysis of bank transaction risks, and directly connected to big data platforms such as hadoop, kylin, and greenplum.
-
FineIndex uses cube connection, which means database-FineIndex-front-end analysis. FineIndex is equivalent to an intermediate library, which is used to store data tables, associate escape indexes, and so on. All of these greatly improve the efficiency of subsequent front-end analysis and processing of data (because direct sql fetching, the efficiency is limited by the database itself, when the amount of data is large, the general analysis tool is easy to get stuck and promoted, and the memory overflow causes the system to become unresponsive)
3. Data blending for performing analysis across data sources. FineBI can mix big data with other data sources (such as MySQL, Excel files, etc.), allowing users to access data across various data sources and integrate them into one library (FineIndex).
Qlikview
1. Qlikview is also a memory-based BI, which can respond to users' analysis needs in real time. Similar to Tableau, the performance is mostly determined by the database. It can directly collect data from the user's business system and perform multi-dimensional analysis, eliminating the drawbacks that traditional BI can only be implemented based on data warehouse (marketplace). (The author did not find too much elaboration on big data from the Internet and its database, and there is no official publicity in this regard).
2. At the core of QlikView is a patented AQL framework. There is no need for traditional OLAP cubes or databases for analysis, and the basic hardware platform is fully utilized to conduct business analysis on hundreds of millions of data. QlikView applications enable a variety of end users to interactively analyze important business information in a highly visual, powerful and creative way.