•A data warehouse by nature is an open, accessible system.

•Aim : Make large amounts of data easily accessible to the
users

•Any Security restrictions seen as obstacles to that goal,
become constrains on the design of data warehouse.

•This is not to say that security is not important; on the
contrary ,security is paramount to ensuring that the data
itself remains clean,consistent and integral.
•It is important to establish early any security and audit
requirements that will be placed on the data warehouse.

•Clearly, adding security will affect performance and design
of data warehouse.

Security can affect many different parts of the data
warehouse such as
                    User Access
                     Data Load
                     Data Movement
                     Query Generation
Data Classification
            Based on sensitivity
            Based by role or job function

User Classification
             Based on Department,Section,Group etc..
                 (User access hierarchy )
             Based on their role
                  (Role access hierarchy )
Data warehouse Inc.




             Sales                              Marketing



                                             Snr Analyst




                                    Analyst                 Analyst
      Administrator               Aggregation               Detailed




 Database             Reference             Summarized             Detailed
Admin Data              Data                 Sales data           Sales data
Data
            Ware House




  Sales                  Marketing
Data Mart                Data Mart




 Users                     Users
Data warehouse Inc.




           Sales                               Marketing


Analyst        Administrator     Snr Analyst        Analyst         Administrator


Analyst                                             Analyst         Administrator


Analyst                                             Analyst


Analyst

           Detailed        Reference         Summarized          Detailed
          Sales data         Data             Sales data      Customer Data
Select
             customer,account_number,sum(value),count(transaction_id)
From
             txn_last_quarter
Where
              transaction_date
                                        between ‘01-jun-96’ and ‘30-jun-96’
Group by
             customer account _number

---------------------------------------Restricting users by using views as--------------------------------------------

Create view sales_lq as
Select
           customer,account_number,sum(value),count(transaction_id)
From
           txn_last_quarter
Where
            transaction_date
                               between ‘01-jun-96’ and ‘30-jun-96’
                      and
                               account_id<>123456789
Group by
           customer account _number
Create view sales_lq as

Select
           customer,account_number,sum(value),count(transaction_id)
From
           txn_last_quarter
Where
           transaction_date
                              between ‘01-jun-96’ and ‘30-jun-96’
                    and
                              account_id<>123456789
                    or
                              account_id<>234567891
Group by
           customer account _number
•Where possible, avoid the use of views to enforce data access restrictions. They can
rapidly become a nightmare to maintain.

•Use Multiple Tables instead of views also creates duplication and overheads.

•Create a dummy field in position or nullify it from the user.
•Legal Requirements
The design team will require some analysts with   knowledge and experience of
business area.

•Audit Requirements
        connections
        •disconnections
        •data access
        •data change


Understand the reasons for each audit requirement.
        Only implement those that are genuinely required for local, company and
    security reasons.
•Network Requirements
When doing the security requirements capture it is important not to overlook issues
      such as network security.

         encryption of data needed?
         which network routes the data can take?

•Data Movement

         Where is the flat file is stored?
         who ha access to that disk space?

         do you backup encrypted or decrypted versions?
         do these backups need special tapes to store ?
         who has access to these tapes ?

         Where that temporary table to be held ?
         how do we make such tables visible ?
•Documentation
It is probably better to document all the restrictions as part of a seperate data
warehouse security policy document.

          Data Classifications
          User Classifications
          Network Requirements
          Data Movement and storage requirements
          All audible actions

•High Security Environments
         Trusted RDBMS
                  ! Trusted RDBMS will generally run on trusted operating systems.
          Covert channels
         ! Avoid creating covert channels that Inadvertently make information
         about data available.

         ! Covert channels are not typically a problem, as the majority of data
         warehouses do not require such high level of security.
•Views
Some of the common restrictions that may apply to the handling of views are
          restricted Data Manipulation Language(DML) operations,
          lost query optimization paths.
          restriction on parallel processing of view projections.

•Data Movement
Different ways in which bulk data movement can occur
                  data loads
                  aggregation creation
                  results temporary tables
                  data extracts

•AUDITING
•APPLICATION DEVELOPMENT

Extra Security code may be needed for each of the process managers :
          load manager
          warehouse manager
          query manager

•Data Base Design

         If a table has three indexes ,three constraints, and five views on it, each copy
of the table will probably add not just the copy but 11 other objects to the database as
well.

•Testing
Further security additions will increase the complexity of the programme cause
increase in errors during testing phase and also additional added functionality to be
Data Warehosing -Security

Data Warehosing -Security

  • 2.
    •A data warehouseby nature is an open, accessible system. •Aim : Make large amounts of data easily accessible to the users •Any Security restrictions seen as obstacles to that goal, become constrains on the design of data warehouse. •This is not to say that security is not important; on the contrary ,security is paramount to ensuring that the data itself remains clean,consistent and integral.
  • 3.
    •It is importantto establish early any security and audit requirements that will be placed on the data warehouse. •Clearly, adding security will affect performance and design of data warehouse. Security can affect many different parts of the data warehouse such as User Access  Data Load  Data Movement  Query Generation
  • 4.
    Data Classification  Based on sensitivity  Based by role or job function User Classification  Based on Department,Section,Group etc.. (User access hierarchy )  Based on their role (Role access hierarchy )
  • 5.
    Data warehouse Inc. Sales Marketing Snr Analyst Analyst Analyst Administrator Aggregation Detailed Database Reference Summarized Detailed Admin Data Data Sales data Sales data
  • 6.
    Data Ware House Sales Marketing Data Mart Data Mart Users Users
  • 7.
    Data warehouse Inc. Sales Marketing Analyst Administrator Snr Analyst Analyst Administrator Analyst Analyst Administrator Analyst Analyst Analyst Detailed Reference Summarized Detailed Sales data Data Sales data Customer Data
  • 8.
    Select customer,account_number,sum(value),count(transaction_id) From txn_last_quarter Where transaction_date between ‘01-jun-96’ and ‘30-jun-96’ Group by customer account _number ---------------------------------------Restricting users by using views as-------------------------------------------- Create view sales_lq as Select customer,account_number,sum(value),count(transaction_id) From txn_last_quarter Where transaction_date between ‘01-jun-96’ and ‘30-jun-96’ and account_id<>123456789 Group by customer account _number
  • 9.
    Create view sales_lqas Select customer,account_number,sum(value),count(transaction_id) From txn_last_quarter Where transaction_date between ‘01-jun-96’ and ‘30-jun-96’ and account_id<>123456789 or account_id<>234567891 Group by customer account _number •Where possible, avoid the use of views to enforce data access restrictions. They can rapidly become a nightmare to maintain. •Use Multiple Tables instead of views also creates duplication and overheads. •Create a dummy field in position or nullify it from the user.
  • 10.
    •Legal Requirements The designteam will require some analysts with knowledge and experience of business area. •Audit Requirements connections •disconnections •data access •data change Understand the reasons for each audit requirement. Only implement those that are genuinely required for local, company and security reasons.
  • 11.
    •Network Requirements When doingthe security requirements capture it is important not to overlook issues such as network security.  encryption of data needed?  which network routes the data can take? •Data Movement  Where is the flat file is stored?  who ha access to that disk space?  do you backup encrypted or decrypted versions?  do these backups need special tapes to store ?  who has access to these tapes ?  Where that temporary table to be held ?  how do we make such tables visible ?
  • 12.
    •Documentation It is probablybetter to document all the restrictions as part of a seperate data warehouse security policy document.  Data Classifications  User Classifications  Network Requirements  Data Movement and storage requirements  All audible actions •High Security Environments Trusted RDBMS ! Trusted RDBMS will generally run on trusted operating systems.  Covert channels ! Avoid creating covert channels that Inadvertently make information about data available. ! Covert channels are not typically a problem, as the majority of data warehouses do not require such high level of security.
  • 13.
    •Views Some of thecommon restrictions that may apply to the handling of views are  restricted Data Manipulation Language(DML) operations,  lost query optimization paths.  restriction on parallel processing of view projections. •Data Movement Different ways in which bulk data movement can occur  data loads  aggregation creation  results temporary tables  data extracts •AUDITING
  • 14.
    •APPLICATION DEVELOPMENT Extra Securitycode may be needed for each of the process managers :  load manager  warehouse manager  query manager •Data Base Design If a table has three indexes ,three constraints, and five views on it, each copy of the table will probably add not just the copy but 11 other objects to the database as well. •Testing Further security additions will increase the complexity of the programme cause increase in errors during testing phase and also additional added functionality to be