07 May Teradata Architecture
The principal strength of Teradata is its parallelism. The architecture of the Teradata is designed in such a way to keep the strength in mind. Teradata is exclusive from any other databases due to its unique architecture only.
The main components of the Teradata architecture are mentioned below–
1. Parsing Engine (PE),
2. Access Module Processor (AMP) and
Parsing Engine (PE): When a user login to Teradata, it basically connect to Parsing Engine. When user submits a query, then PE takes action. PE creates a plan and then instructs AMPs what to do to get the result from the mentioned query. The parsing engine knows all like how many AMPs are connected to the Teradata system, how many rows are there in the table and what is the best possible arrangement to execute the query. Parsing engine is also called as Optimizer. It also checks the access right for the user weather he has the privilege for executing the query or not.
Access Module Processor (AMP): Each access module processor attached to the Teradata system works according to the parsing engine via the BYNET for the instructions. Each Access Module Processor is connected to its own disk and has the freedom to read or write the data to the disk. It can be considered as the computer processor. It has its own disk attached to it. Whenever AMP receives the instructions from the parsing engine, it fetches the data from disk and sends it to back to the parsing engine through the BYNET. Each of these AMP is permitted to read and write in its own disk only. This is also known as Shared Nothing Architecture.
BYNET: It is the communication channel between parsing engine and the Access Module Processor, which ensures that communication between AMP and PE is on right track. In the Teradata system there are always two BYNET systems. These are: BYNET 0 and BYNET 1. The reason two BYNET exist is –
1) If one of the BYNET fails, the second BYNET can take its place.
2) Two BYNETs can improve the performance of the system. The parsing engine and access module processor can talk to each other over both BYNET which speed up the communication.