Cognos_10_介绍.ppt

Module 7 – IBM Cognos Real Time Monitoring IBM CONFIDENTIAL – DO NOT DISTRUBUTE 7-* Stream Query Processing Real-time monitoring does not follow the same query processing paradigm as traditional business intelligence. Traditional business intelligence uses structured query language (SQL) in order to query a database, then wait for and return the results. Real time monitoring instead uses stream processing. The query language is CSQL or continuous structured query language instead of SQL. The engine “listens for” and receives a continuous stream of data, monitoring for changes or exceptions in that data. The information is loaded into memory for immediate display. There is practically no data latency (latency is in the milliseconds) and cases have shown stream processing to process more than 700 million rows per second. Example: Stock Exchange A good example to use here is the stock exchange, or trading floor example: On the stock trading floor – trades are placed Event engines are looking for certain patterns to initiate either a sell or a buy transaction The event engine will examine the streams of trade data, most of which is not relevant to the engine seeking the pattern When a pattern occurs that matches the designated pattern, the engine will either initiate the transaction automatically and send a confirmation to the person overseeing the transaction, or send an alert to the person with a work flow to approve the transaction as it is taking place. In either action, time is of the essence, since the trade window is so small Module 8 – External Data Access IBM CONFIDENTIAL – DO NOT DISTRUBUTE 8-* The user must be granted the Allow External Data capability by the Cognos administrator. There are governors set up on new external data packages for maximum external data file size and maximum number of rows. The default values can be changed by modifying the settings in Framework Manager. Module 9 – IBM Cognos Statistics IBM CONFIDENTIAL – DO NOT DISTRUBUTE