Master Large Data0Streams with Microsoft Stream Insight

摘要

Discovering that the production line has gone down, users' media streams are skipping or one of your products has become a "must have" is easy once it has already happened. The real trick is identifying these scenarios as they happen or even predicting them based on past trends. Successfully predicting scenarios like these requires a near-realtime approach. By the time relevant data is extracted, transformed and loaded into a traditional business intelligence (BI) solution like SQL Server Analysis Services (SSAS), the situation has long since changed. Similarly, any system that relies on a request-response pattern to request updated data from a transactional data store (such as a SQL Server Reporting Services, or SSRS, report) is always operating on stale data near the end of its request-polling interval.