Teradata Architecture:
Teradata mainly contains the following architectural components.
1. Parsing Engine (PE)
2. BYNET
3. AMP (Access module processor)
4. Disks
Parsing Engine:
Parsing Engine is a virtual processor (vproc).It has the following software components.
1. Session control
2. Parser
3. Optimiser
4. Dispatcher
Whenever a SQL request is given to the parsing engine the session control verifies for the session authorisation (user name and password) and based on that processes or will reject the request.
Parser verifies the sql request for proper syntax and evaluates them. Checks the data dictionary if all the objects exist and if the user has authority to access them.
Optimiser develops the least expensive (time) plan to execute the request .optimiser must know about system configuration, available AMP’S and PE’S. Optimiser enables Teradata to handle complex queries efficiently.
Dispatcher controls the sequence in which steps are executed an passes the steps to the BYNET. BYNET is a messaging layer in between parsing engine and access module processor. After the AMPs process the steps, the PE receives their responses over the BYNET. The Dispatcher builds a response message and sends the message back to the user.
Teradata uses hashing algorithm to distribute rows evenly across the amps .Based on the index column Teradata generates the hash value and based on that value data is sent to different amps. Data having same hash value will be sent to one amp (duplicate data).so depending on hash value data is evenly distributed across all the amps. when index is not selected properly data will not be distributed equally and hence leads to more data in one amp and less in one amp which is called skewness .
Teradata mainly contains the following architectural components.
1. Parsing Engine (PE)
2. BYNET
3. AMP (Access module processor)
4. Disks
Parsing Engine:
Parsing Engine is a virtual processor (vproc).It has the following software components.
1. Session control
2. Parser
3. Optimiser
4. Dispatcher
Whenever a SQL request is given to the parsing engine the session control verifies for the session authorisation (user name and password) and based on that processes or will reject the request.
Parser verifies the sql request for proper syntax and evaluates them. Checks the data dictionary if all the objects exist and if the user has authority to access them.
Optimiser develops the least expensive (time) plan to execute the request .optimiser must know about system configuration, available AMP’S and PE’S. Optimiser enables Teradata to handle complex queries efficiently.
Dispatcher controls the sequence in which steps are executed an passes the steps to the BYNET. BYNET is a messaging layer in between parsing engine and access module processor. After the AMPs process the steps, the PE receives their responses over the BYNET. The Dispatcher builds a response message and sends the message back to the user.
Teradata uses hashing algorithm to distribute rows evenly across the amps .Based on the index column Teradata generates the hash value and based on that value data is sent to different amps. Data having same hash value will be sent to one amp (duplicate data).so depending on hash value data is evenly distributed across all the amps. when index is not selected properly data will not be distributed equally and hence leads to more data in one amp and less in one amp which is called skewness .
No comments:
Post a Comment