Teardata - II

Teradata Storage Architecture

The Parsing Engine interprets the SQL command and converts the data record from the host into an AMP message.

The Message Passing Layer distributes the row to the appropriate Access Module Processor (AMP).

The AMP formats the row and writes it to its associated disks

The disk holds the row for subsequent access.

The Host or Client system supplies the records. These records are the raw data from which the database will be constructed.

The Parsing Engine is a component that interprets SQL requests, receives input records and passes data. To do that it sends the messages through the Message Passing Layer to the AMPs.

The Message Passing Layer is implemented variously as hardware or software, depending on the platform used. It determines which unit or units should receive a message.

Think of the AMP (Access Module Processor) as a computer designed for and dedicated to managing a portion of the entire database. It performs all the database management functions – such as sorting, aggregating, and formatting the data. It receives data from the PE, formats the rows, and distributes the rows to the disk storage units it controls. It also retrieves the rows requested by the parsing engine.

Disks are simply disk drives associated with an AMP. They store the data rows. On current systems, they are usually implemented using a disk array (to be discussed later).

Teradata Storage Architecture

Parsing Engine dispatches request to insert a row.

Message Passing Layer insures that a row gets to the appropriate AMP (Access Module Processor).

AMP stores row on its associated disk.

Each AMP can have multiple physical disks associated with it.

Teradata Retrieval Architecture

Retrieving data from the Teradata RDBMS simply reverses the process of the storage model. A request is made for data and is passed on to a Parsing Engine (PE). The PE optimizes the request for efficient processing and creates tasks for the AMPs to perform, which will result in the request being satisfied. These tasks are then dispatched to the AMPs via the Message Passing Layer. Often times all AMPs must participate in creating the answer set, such as in returning all rows of a table. Other times, only one or a few AMPs need participate, depending on the nature of the request. The PE will insure that only the AMPs that are needed will be assigned tasks on behalf of this request.

Once the AMPs have been given their assignments, they will retrieve the desired rows from their respective disks. If sorting, aggregating or formatting of any kind is needed, the AMPs will also take care of that. The rows are then returned to the requesting PE via the Message Passing Layer. The PE takes the returned answer set and returns it to the requesting client application.