Skip to main content

Teradata Architecture

By
We have many relational databases available in market, teradata is one such RDBMS it is usually preferred when we are dealing data with tera bytes in size and when we need a very fast quick response for retrieval data.

Teradata main advantage is its parallelism architecture and utilities.Lets get to know more about it in a deeper way



we have Parsing engine which has 4 components
a)Session control- It checks for user/login credentials before processing the query
b)Parser-It checks for the SQL syntax and user rights to access various database objects referred in the SQL query submitted by user
c)Optimizer-It generates the query execution plan
d)Dispatcher-It passes execution plan to bynet and receives all the responses and sends back to user

Next we have Bynet it is the communication layer between PE (parsing engine) and AMP's ( access module processors).There are two bynets available Bynet-0 and Bynet-1 which help in continuous communication between AMP's and PE

The final and the most important component is AMP. It is considered as heart of teradata. AMP's are the one which do retrieval and processing the data. Each AMP is connected to its virtual disk where data is stored.They work on their vdisks and do not have to access to other amps vdisks
More the number of AMP's the better will be performance

Hope this article helps you in understanding teradata architecture in the coming posts we will see about teradata utilities

Comments

Post a Comment

Popular posts from this blog

Comparing Objects in Informatica

By
We might face a scenario where there may be difference between PRODUCTION v/s SIT version of code or any environment or between different folders in same environment. In here we go for comparison of objects we can compare between mappings,sessions,workflows In Designer it would be present under "Mappings" tab we can find "Compare" option. In workflow manger under "Tasks & Workfows" tab we can find "Compare" option for tasks and workflows comparison respectively. However the easiest and probably the best practice would be by doing using Repository Manager.In Repository Manager under "Edit" tab we can find "Compare" option. The advantage of using Repository manager it compares all the objects at one go i.e. workflow,session and mapping. Hence reducing the effort of individually checking the mapping and session separately. Once we select the folder and corresponding workflow we Can click compare for checking out ...
11 comments
Read more

Target Load Type - Normal or Bulk in Session Properties

By
We can see the Target load type ( Normal or Bulk) property in session under Mapping tab and we will go for Bulk to improve the performance of session to load large amount of data. SQL loader utility will be used for Bulk load and it will not create any database logs(redolog and undolog), it directly writes to data file.Transaction can not be rolled back as we don't have database logs.However,Bulk loading is very as compared to Normal loading. In target if you are using Primary Key or Primary Index or any constraints you can't use Bulk mode. We can see this property in the below snap shot.
Post a Comment
Read more

SCD Type 2 Implementation in Informatica using dynamic lookup

By
We have already covered how a dynamic look up works the same can be found out from the following link http://dwbuddy.blogspot.com/2014/06/dynamic-look-up-in-informatica.html  So let's see how we can use it in a regular SCD type 2 mapping. The logic used in SCD type 2 is same as discussed in earlier blog i.e.  http://dwbuddy.blogspot.com/2015/01/scd-type-ii-implementation-in.html . The only difference being here is we have used dynamic look up instead of a static one. We go for dynamic look up when there are duplicate entries for a PK value and only the latest record present in the file should be termed as active. Let us consider the following sample data So  for EMP_ID 101 we have two records in the same day flow so the latest record i.e. with salary as 1500 should be termed as Active record. The mapping will look in the following way So in here we have a dynamic lookup couple of expression transformations,a transaction control transformation which is use...
Post a Comment
Read more