Wednesday , April 24 2024

Sybase IQ Architecture


Single Node Architecture:

when we talk about Sybase IQ architecture , We should explain about the Three important IQ stores and three main files. which plays a vital role in Sybase IQ server.

1) Catalog Store
–> Stores Metadata in system tables that is needed to run the database.
–> Catalog Store is always recommended to create on a file system
–> Physical Name : dbname.db
–> Logical Name : SYSTEM
–> Catalog Store has a separate transaction log file. Physical file name of the transaction log is dbname.log

2) IQ store
–> Stores User data as compressed indexes on the disk
–> Logical Name : IQ_SYSTEM_MAIN for the first/default dbspace of the IQ store. After that we can add as many dbspaces as we want that depends on the space requirement..
–> Physical Name :
–>IQ_SYSTEM_MAIN contains necessary information to open the database , IQ checkpoint,info about rollback/rollforward , backup related metadata. So it is suggested not to create user tables on this dbspace , instead create new  IQ_USER_MAIN dbspace and use it for user data.

3) IQ temporary Store
–> acts like a tempdb in ASE . work area for sorting and grouping
–> Physical Name : dbname.iqtmp
–> Logical Name : IQ_SYSTEM_TEMP is the default dbspace .Depends on the space requirement we can add number of dbspaces.


Important Server Files

1) IQ Message Log
–>This is the IQ servers main log where all the important messages to monitor are logged
eg: – Insert & delete Messages , Error Messages, query plans , status messages.
–> physical file name : dbname.iqmsg

2) Catalog Store transaction log
–> this file is the transaction log for the catalog store .
–> physical file name : dbname.log

3) IQ server logs
–> They are created each time IQ servers is started and logs the startup server log information .

Dbspace :

–>dbspace is like a device in Sybase ASE . Here in IQ we create / extend  Store using dbspaces .
–>dbspace may be a file system file or Raw partition .
–>The IQ store , catalog store and IQ temporary store consists of Multiple dbspaces .

SybaseIQ  Multiplex Architecture

Sybase Multiplex , the name indicates that it has more than one node , may be 2 or 3, or 4 depends on the project requirements .

Types of servers in a multiplex  or role of each server in a Multiplex :

  1.  Coordinator
  2.  Writer
  3. Readers

Short Notes on Multiple server roles :

Coordinator Node
–>Each multiplex configuration has one and only one coordinator node that is capable of readonly and read-write operations and is also responsible for bookkeeping to manage other nodes.
–>When you convert an existing simplex server to multiplex, it becomes the coordinator node.

All nodes in a multiplex configuration are secondary nodes except the coordinator.
Secondary nodes:Can be either read-only nodes (reader nodes) or read-write nodes (writer nodes).

Writer nodes:
Can run read-only and read-write operations against shared IQ objects.
• Can manipulate local data in temporary and SA base tables.

Reader nodes:
• Can run read-only operations against shared IQ objects.
• Can manipulate local data in temporary and SA base tables.

–> Each server has its own catalog store, tempdb store and they all share IQ main Store .


About Bhaskara Vasamsetti

One comment

  1. IQ has excellent speed of data loading, batch style; for load large amount rows/data set. IQ table can be over 1000 columns wide, snowflake DB design.
    IQ load is even faster if the data file is visible by the IQ host, or we called IQ local loads, the file can be a compressed (.gz) data file.

    And IQ is very good at doing complex ANSI join, BO/Cognos/Micro Strategy Cube; such as over 20-30 table joins. For complex reporting SQL, IQ can be over 50-100 times faster than traditional Oracle / MSSQL.

Leave a Reply

Your email address will not be published. Required fields are marked *