In the shared-nothing architecture, each node is made of processor, main memory and disk and communicates with other nodes through the interconnection network. Each node is under the control of its own copy of the operating system and thus can be viewed as a local site (with its own database and software) in a distributed database system. Therefore, most solutions designed for distributed databases such as database fragmentation (called partitioning in parallel databases), distributed transaction management and distributed query processing may be reused.
As opposed to symmetric multiprocessor (SMP), shared-nothing is often called massively parallel processor (MPP). Many research prototypes and commercial products have adopted the shared-nothing architecture because it has the best scalability. The first major parallel DBMS product was Teradata which could accommodate a thousand processors in its early version in the 1980s. Other...