If you have been reading business or technical journals recently, you have probably read about data warehousing or using very large databases for relationship marketing or decision support systems. One of the terms you probably have seen is Massive Parallel Processing (MPP). In this article we will look at this technology to understand how it works and why it is becoming a strategic weapon for many companies with their data warehouse.
MPP machines are designed specifically to handle very large data warehouses or very large databases (VLDBs). We are talking about hundreds of gigabytes or many terabytes of data. Several of the largest retailers, railroads and airlines in the world use MPP technology to store more than several years' worth of detailed sales data. The largest single table on a relational database is on an MPP machine with over 4 BILLION rows! MPP machines use hundreds or thousands of processors instead of one or a few processors to do their work. Each of these many processors has it own set of disk storage units (DSUs) or dedicated storage area to hold the data. In order to access these massive data stores, MPP machines often use shared nothing or shared almost nothing designs. The difference between these would be with true shared nothing design, the processing engines that have their own copy of the operating system, memory and DSUs. They communicate with each other over an I/O bus. They do not share memory or DSUs like SMP (symmetrical multi- processing) machines do. Some of the newest releases of MPP DBMSs have moved to a shared almost nothing architecture, in which clustered SMP nodes are used. A node in the clustered node design looks very much like a standard SMP node. The major difference is that these nodes communicate with each other over an I/O bus and appear as one large logical machine.
Register or login for access to this item and much more
All Information Management content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access