                 Visual Warehouse Live Demo

This document is a guide to run a demonstration using the sample database.
It contains the following sections:
 1. The business scenario
 2. Demo environment and startup
 3. Showing Visual Warehouse

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-

1. The Business Scenario

The scenario used in the demo is centered around the analysis of the
financial status of a manufacturing corporation. The corporation
manufactures goods (machines) at several locations and obviously
sells the machines that have been manufactured. Although all the machines
are manufactured in the same country, the sales network extends
internationally. From a finance point of view, our executives want to
get accurate and rapid answers to the following questions:
 - What machine types do sell best and where?
 - For which machine types is the profitability the largest?
 - Are they some machine types whose sales constantly outperform others?
 - Should we discontinue the production of certain types of machines?
 - Should some sales branches be closed or discontinue the sale of
   certain types of machines?
 - Who should I reward for outstanding performance?
 - How can I express the conditions under which awards should be
   delivered, and how could that process be automated?

Why is this information hard to be exploited by the end users?
 1. The information is scattered in a variety of operational systems.
 2. The information may reside on different kinds of databases
    In the demo, all the data is on the workstation in DB2/NT or in
    flat files, but the data could also have been placed in remote DB2
    systems, or Oracle, Sybase, Informix, Microsoft SQL Server, or even
    in IMS, VSAM or other flat files.
 3. The operational data may require cleansing: in our example, all the
    sales data in the operational databases is expressed in local
    currency, with a field that identifies the currency used (e.g. DM,
    FB, Lira, Dollar, German Mark). Obviously, the sales data will need
    to be aggregated using a common currency! So, the cleansing operation
    consists in identifying the name of the country that corresponds to
    the currency field. Having now the name of the country, we can look
    up into a table of exchange rates - data that we bought externally -
    to convert all pricing information into a single currency.
 4. The operational data may need to be reorganized to be useful
    for informational purposes: it may require different levels
    of summarization and aggregation. At the highest level, the net
    income, we need to aggregate all the sales data from each sales
    branch, and all the manufacturing data from all the plant sites
    before the monthly net income can be computed for each machine
    type.
 5. Records of past activity are usually not kept in the operational
    environment: so trend analysis is difficult, if not impossible to do.
 6. Lots of manual and non-synchronized extracts are usually required.
    What happens if I cannot get access to the operational data in one
    of the branch offices? Will the aggregation yield incomplete results?
    How do I keep track that some data is currently missing?
 7. End users are typically not being given notification when a crucial
    situation has occurred: I would like to know when the sales are below
    a certain threshold; or, as CEO of the company, I would like to
    have a process that detects automatically what are the sales branches
    that have performed outstandingly, and reward these branches.
 8. End users need to have a clear understanding of the meaning of the
    data they look at, the currency of the data, where the data came
    from. All this description is rarely available today.
 9. End users must have great flexibility in accessing the data. Most
    end-user tools that use the ODBC or DB2 for Windows NT CAE interface
    can directly access Visual Warehouse data. Or users may prefer to
    browse first through DataGuide to see what kind of data exists, and
    from there launch the application of choice. Still, another way to
    access the data is to use internet or intranet browsers.

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-

2. The demo sample

The demo is composed of several sample operational databases and a
Visual Warehouse datamart.

              ------------------------------------
              Description of operational databases
              ------------------------------------
DB2/NT Source database "REVENUE"
   The following tables are in the database:

   Creator Name        Description
   --------------------------------------------
   PAUL    REVENUEL01  Revenue data for Atlanta
   PAUL    REVENUEL02  Revenue data for Seattle
   PAUL    REVENUEL03  Revenue data for Mexico
   PAUL    REVENUEL04  Revenue data for Tokyo
   PAUL    REVENUEL05  Revenue data for Singapore
   PAUL    REVENUEL07  Revenue data for Europe

DB2/NT Source database "COST"
   The following tables are in the database:

   Creator Name        Description
   --------------------------------------------
   JOHN    COSTL10     Cost data for Boston
   JOHN    COSTL11     Cost data for San Jose
   JOHN    COSTL12     Cost data for Pittsburgh
   JOHN    COSTL13     Cost data for New York

DB2/NT Source database "EXCHANGE"
   The following tables are in the database:

   Creator Name        Description
   --------------------------------------------
   BANKER  RATE        Currency exchange rate for various countries


       ----------------------------------------------
       Description of sample data in Visual Warehouse
       ----------------------------------------------

Information Resources:
  MANUFACTURING Manufacturing data - source database
  SALES         Corporate sales data - source database
  SALESEMP      Employee records data - source file
  CURRENCY      Currency exchange rate data - external data
  Finance       Corporate finance data warehouse


User Group
  BUDGET        Provides access to Information Resources

Business Views
  Manufacturing cost in Boston      Production cost data for Boston
  Manufacturing cost in New York    Production cost data for New York
  Manufacturing cost in Pittsburgh  Production cost data for Pittsburgh
  Manufacturing cost in San Jose    Production cost data for San Jose
  Revenue - Atlanta                 Revenue data for Atlanta
  Revenue - Seattle                 Revenue data for Seattle
  Revenue - Mexico                  Revenue data for Mexico
  Revenue - Tokyo                   Revenue data for Tokyo
  Revenue - Singapore               Revenue data for Singapore
  Revenue - Europe (local currency) Revenue data for European countries
                                        (transient data)
  Rate                              Monthly average exchange rate
  Sales Managers                    Name and address of area sales managers

  Sum of cost by location          Production cost for each machine
                                   at each location
                                   -- built from the 4 "COST" BVs
  Sum of cost by machine           Total production cost for each machine
                                   -- built from "sum of cost by location"
  Revenue - Europe                 Revenue data for European countries
                                   -- built from "Revenue - Europe (local currency
                                   -- data cleansing performed by a
                                      user defined program UDPCLEAN in
                                      directory vwdemo
  Sum of revenue by location       Revenue for each machine
                                   at each location
                                   -- built from 6 "Revenue" BV, and rate
  Sum of revenue by machine        Total revenue for each machine
                                   -- built from "Sum of revenue by location"
  Corporate Profit                 Profitability for each machine
                                   -- built from "Sum of cost by machine"
                                      and "Sum of revenue by machine"
  Corporate Income                 Monthly net income for each machine type
                                   -- built from "Sum of cost by machine"
                                      and "Sum of revenue by machine"
  Corporate Awards                 Awards to best performing regions
                                   -- built from "Corporate Profit",
                                      "Sum of revenue by machine",
                                      "Sum of revenue by location"
                                   -- triggered by "Corporate Income"
                                      using program TRGAWARD in vwdemo
  Award Recipients                 List of award recipients
                                   -- built from "Corporate Awards"
                                      and "Sales Managers"


-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-

2. Demo environment and startup

Startup procedure:
 1. Power on the computer and wait until it quiesces
 2. Double click on "Visual Warehouse" icon in the Program Manager window
 3. In the "Visual Warehouse (Common)" window, double click on the
    Visual Warehouse Desktop" icon.
 4. When prompted in the IBM Visual Warehouse Logon window, enter VWADMIN
    in the Password box.
 5. Click on "OK", or press "Enter"
 6. IBM Visual Warehouse Desktop window appears after a short while.

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-

3. A demonstration scenario of Visual Warehouse

The demo is intended to show how Visual Warehouse can automate the
process required to solve the business problem outlined in section 1.
Visual Warehouse provides a very powerful and yet simple mechanism:
 - to collect the data from the operational environments (via the
   definition of business views (BV), which are the templates that
   describe what needs to be extracted, and when).
 - to get operational data from relational as well as non relational sources.
 - to provide various views on the data in the datamart by defining
   business views from other already defined business views
 - to aggregate and summarize data
 - to cleanse the data (via the use of user-defined programs specified
   as part of the definition of the business view).
 - to express business processes (via triggers and the cascade function):
   for example, the president of our corporation may decide to put in
   place the following policy: after getting the monthly corporate
   financial results, awards will be distributed if the total net income
   exceeds a certain threshold. The awards will be delivered to the
   sales managers of each country that had contributed at least 25% of
   the total revenue for a particular type of machine. And there is
   an additional condition: the sales price has to be at least 3 times
   the manufacturing cost for that type of machine. If all these
   conditions are true, then award will be the equivalent of 10% of the
   revenue generated by the country. A rather complex business process,
   isn't it? Visual Warehouse can automate all this and more!
 - to do trend analysis (by keeping multiple editions of a business view)
 - to monitor the activity on the warehouse.
 - to use an end-user tool of choice to display and analyze the data.

In the description sections below, whenever there is a number in
parenthesis, we suggest you perform the step with the same number
described in the Operation section that precedes.
When running the demo, you can always activate the Help menu to get
extensive help on all the Visual Warehouse features.

          ....................................................

3.1. Visual Warehouse desktop

Section summary and objective
  This section and all the subsequent ones in chapter 3 highlight the
  functions performed by the data warehouse administrator to build and
  maintain the data warehouse.

Operation:
 1. double click on the IBM Visual Warehouse VWSAMP2.IWH icon.

Description:
  - When the VW administrator opens the desktop (1), the list of
    available functions is displayed. By activating these functions, the
    administrator is able to:
    - define the data to be put in the data warehouse (choose File menu
      on the menu bar, and highlight Business View Warehouse resource)
    - specify the operational data sources from which to extract the data
      (choose File menu on the menu bar, and highlight a data source)
    - associate users with functions they can activate and data they can
      access (choose Security on the menu bar)
    - manage the activity in the data warehouse (choose Operations on the
      menu bar),
    - define settings for configuration, trace and agents (choose the
      Settings menu on the menu bar).

    We invite you now to explore many of these functions, by following
    the scenario we described at the start.

          ....................................................

3.2. Collecting, consolidating, cleaning and aggregating the data

Section summary and objective
  Building a data warehouse requires the data warehouse administrator
  to specify what data needs to be extracted from the operational sources,
  what filter to apply, and when to extract the data. All these operations
  are defined as part of the Business View definition.

Operation: opening the business view list
 1. double click on "Finance" in the Resource Name column;
    this can also be done by single clicking on "Finance", clicking
    on File on the menu bar and choosing "Open" as a menu item.
 2. maximize the "Business View List" window

Description:
  - The data warehouse administrator maintains all the information about
    what data needs to be extracted, how to enhance the data, and when to
    refresh the data,through the business view definition process. (1) (2)
  - Here, we see a list of business views (BV) that the administrator has
    already defined. Business views can be in one of three statuses:
    development mode, test mode or production; in development mode (i.e.
    the definition is not complete yet), some business views are in
    test mode (i.e. the business view definition has been validated and
    the underlying table is created in the database), and some are in
    production mode (in this mode, a business view can be populated
    automatically according to a specific schedule).
  - Business views can be built from data stored in operational data
    sources, or from data stored in the data warehouse. In the latter
    case, a business view receives data from other business views.
  - We see here a list of BV that represent various aggregation levels
    of cost and revenue related to our corporation. Let us look at the
    steps involved in the definition of a BV.
  - To be able to change the definition of a business view, the BV
    must be in development mode; some minor changes that would not affect
    the result set can be performed while in test mode also.
  - So, let us open a BV in development mode; if no such BV exists yet,
    put one in development mode by following the next operation set.

Operation: creating a business view in development mode
 1. perform these steps only if Manufacturing cost in Boston0002 does not
    exist or is not in development mode status.
 2. click on the business view called "Manufacturing cost in Boston"
 3. click on Edit on the menu bar
 4. click on Copy
 5. A Copy Business View window appears
 6. Click on Copy

Description:
  - At some point, you may want to change the definition of an existing
    business view already in production mode, while keeping that one
    intact. Just copy the definition of the existing business view
    by performing steps 2 to 6.
  - We now have business view "Manufacturing cost in Boston_0002"
    in development mode. Notice that since we did not change the name
    of the original business view, the system automatically appends a
    version number to the name.
  - Now we are ready to go through the definition steps of a BV.


Operation: business view definition (high level)
 1. double click on BV named "Manufacturing cost in Boston_0002
 2. a window called "Business View: Manufacturing cost in Boston_0002"
    is displayed; it contains a notebook with 3 pages: column mapping,
    information and schedule
 3. click on the Information label on top of the notebook.
 4. click on the Schedule label on top of the notebook.

    the operational database. That information is represented in a
    separate table for each manufacturing location. If we are interested
    to look at the cost at each location individually, it makes sense
    to create a separate BV for each manufacturing plant. This is what
    we have done in our datamart implementation. (1), (2).
  - We could have been only interested in the production cost of each
    machine type, irrespectively of the manufacturing plant. In that case,
    we could have immediately defined a BV that was a gathering and
    summarizing data from multiple tables in the operational environment.
  - Immediately combining data from multiple tables at the source is a
    very powerful capability because it can tremendously minimize the
    amount of data being shipped over the network. Joining data from
    multiple tables or files is possible as long as all the data resides
    in the same resource database.
  - For each business view created, there are basically 3 types of
    information that the system needs:
    - column mapping: where to extract the data from, which BV columns to
      define, what mapping (aggregation, summarization, data cleansing)
      to perform between the original data and the BV being defined,
      which filter to apply on the extracted data. (3)
    - general information, the high level view of what the BV represents,
      what is its business name, and the associated table name in the
      datamart (4)
    - schedule: when to populate the business view.
  - Let us find more about each of these information types, and how
    these features translate into business value


Operation: business view definition (information page)
 1. click on the Information label on top of the notebook (in the
    window called "Business View: Manufacturing cost in Boston_0002)

Description:
  - (1) The information page contains the name of the business view, the
    name of the table that will store the data in the datamart, the
    security groups that are allowed to modify the BV definition.
  - Four other important features are specified here:
    - number of editions: each time the data is being inserted into the
      table, it can replace the previous content of the table, or it can
      be appended to previously stored data. This can be particularly
      useful if you need to perform some kind of trend analysis, and
      want to see the evolution of the data values over a 12 month period,
      a 52-week period, or any other period of time.
    - agent site: this is the site that is responsible to execute the
      pull of data for this business view. In order to achieve the best
      performance, consider having multiple agents, so they could work
      in parallel to populate the same datamart. Consider also putting
      the agent on the source or the target environment so that the data
      will move directly from source to target, and won't have to go
      through an intermediate "pumping" location. Look into the online
      documentation for the types of platforms currently supported as
      agent sites.
    - DB2 aliases: in our example, we are looking at version 2 of the BV;
      version 1 is probably in production mode, and we may want to switch
      users to version 2 of the BV at some point. This can be done
      without any disruption for the end users by having an alias name
      that refers to different business view names over time.
    - transient data: how many users will look at the business view and
      how often? How much data will need to be stored for that BV? In
      some cases, it may be worth considering not to store the data
      in the datamart, but to compute the SQL statement on the fly when
      a user requests that BV data. Transient data notifies the system
      that the data should not be stored in the datamart. Advantages
      of transient data are that the data requested is guaranteed to
      be up-to-date data, and this alternative requires less datamart
      storage space.


Operation: business view definition (column mapping: what data source
    and which tables in that data source to access?)
 1. click on the Column Mapping label on top of the notebook (in the
    window called "Business View: Manufacturing cost in Boston_0002)
 2. click on "Sources"
 3. click on one of the Available Tables in the Source Selection window
 4. click on Add in the Source selection window
 5. click on OK
 6. click 'Yes' at the warning message "Do you wish to continue?"
 7. click on "Join"
 8. click on "Table 1 Columns" pull-down tab
 9. choose one column name (e.g. serial)
10. click on "Table 2 Columns" pull-down tab
11. choose one column name (e.g. serial)
12. click on Add
13. click on 'Yes' at the warning message
14. click on "AutoJoin"
15. click on "Cancel"

Description:
  - (1) (2) In the column mapping pages, you provide all the information
    related to what data needs to be extracted and how it needs to
    be transformed.
  - You start by specifying which source data to access: in the source
    selection page, the available data sources are displayed (since
    this BV in development mode was not created from scratch, but
    was copied from another definition, the Select Source item cannot
    be changed).
  - From that selected source, you get a list of all the table names that
    Visual Warehouse knows about (we will see how in the resource
    definition section).
  - Choose one or multiple tables you want to extract the data from and
    click the Add push-button (3), (4)
  - In case of mistake, you can remove tables from the selected list
    by highlighting the table to be removed from the BV tables list,
    and clicking on the Remove push-button.
  - If multiple tables have been selected, you can now define how
    to join them together (5), (6), (7)
  - For each table, choose the appropriate column that needs to
    participate in the join (8), (9), (10), (11), (12), (13)
  - If many columns are involved in the join, it may be faster to
    depress the AutoJoin push-button: all the columns with the same name
    are then selected automatically to participate in the join (14).
  - You have identified so far the tables/files involved in the creation
    of the BV. You are now ready to specify what the data needs to look
    like in the datamart table (15).


Operation: business view definition (column mapping: what columns are
    needed, and how are the values in these columns computed from the
    source data)
 1. click on the Add push-button
 2. a window called "Add columns" is displayed
 3. click on the pull-down push-button for Select Table
 4. highlight a table name
 5. highlight the name of one of the available columns (e.g. GA)
 6. click on the Add push-button
 7. click on OK when the warning message about SQL statement appears.
 8. click on Close
 9. in the table under column name, highlight column named TOTAL_COST
10. click on the pull-down push-button at the end of that TOTAL_COST row
11. click on Cancel push-button
12. in the table under column name, highlight column named YEARMONTH
13. click on the pull-down push-button at the end of that YEARMONTH row
14. click on the pull-down push-button associated with Column Token
15. click on Cancel push-button
16. click on the pull-down push-button associated with Data Type for
    the TOTAL_COST row
17. click on DECIMAL in the menu list
18. go in the Length column and overwrite the value
19. go in the Column name and change TOTAL_COST into FULL_COST

Description:
  - (1) to (7) The definition of columns to be included in the datamart
    table is straightforward: just pick and choose the relevant columns
    from the original source. The name and type of the column is by
    default the same as the one in the source table.
  - (8) to (10) Some columns in the datamart table may be computed
    columns: in this example, the TOTAL_COST consists of the sum of
    2 columns in the source. Complex computation can be expressed here
    by using arithmetic operators, functions, constants, etc...
    From a business perspective, you get into the datamart just what you
    need: the refinement on the source data is expressed once here, and
    a point-and-click mechanism lets you specify what you need very
    easily.
  - (11) to (14) The value of a column can also be computed by using
    system tokens: these tokens provide many representations of date
    and time and the value is generated by the system when the BV table
    is populated. These tokens come particularly handy when you need to
    add a time component in your datamart table. This column provides
    then meaningful information to the business user of when the data
    item was inserted in the datamart. This column is also very useful
    if multiple editions of the data are kept within the datamart table.
  - (15) to (19) The column name and data type can be changed if
    necessary: for example, the original data might be an integer,
    and it could be changed into a decimal value. Also, changing the name
    of the column might help in providing clearer business meaning to
    the name. All these changes are very easy to perform.


Operation: business view definition (column mapping: filtering the source
           data)
 1. click on the Where push-button
 2. click on the Cancel push-button
 3. click on the SQL push-button
 4. click on the Cancel push-button

Description:
  - (1) In the datamart table, you may not be interested to get all the
    rows from the source data. You can specify any filter necessary,
    using the power of the SQL where clause. A point-and-click mechanism
    can help you easily define the filter, or you can just write the
    where-clause in the Where box.
  - (2) (3) At this stage, you have defined what needs to be extracted
    as well as the transformations. You can have a look a the SQL
    statement that has been generated, and still modify it if needed (4).


Operation: business view definition (column mapping: invoking a UDP
           (User-Defined Program))
 1. click on the UDP push-button
 2. click on the Cancel push-button
 3. click on 'Yes' at the warning message.

Description:
  - It may not always be possible to express all the transformations
    required (especially if data cleansing must be performed) in one
    single SQL statement. There may also need to be some business logic
    that needs to be expressed to compose the business view.
  - In this case, you can invoke a user defined program as part of the
    definition of the business view. (1)
  - The program that includes all the logic to transform the source
    data can be any EXE, DLL or BAT program.
  - In this sample demo, we use a UDP for the definition of the BV called
    "Revenue - Europe"; it does a simple data cleansing job which
    consists of replacing the column with currency value with the
    associated country. (2) (3)
  - So far, you have performed the steps involved with the definition
    of WHAT the data in the datamart will be. One more step must be
    performed: this step consists of defining WHEN the table in the
    datamart must be populated


Operation: business view definition (schedule)
 1. click on the Schedule label on top of the notebook (in the
    window called "Business View: Manufacturing cost in Boston_0002)
 2. click on Schedule Type
 3. click on Weekly item in the Schedule Type menu pull-down
 4. change the counters (hour, minute, ...) if you want
 5. click on the Add push-button
 6. click on starts
 7. click on one of the BVs in the Starts window (there will not be
    any BV in that window if you do not have other BV in development mode)
 8. click on the Add push-button
 9. click on Cancel push-button
10. click on 'Yes' at the warning message.

Description:
  - You can request the data to be put in the datamart table on a
    scheduled basis: hourly, daily, weekly, yearly, on a continuous
    interval, or at specific dates. (1) to (5)
    As a business user, you are confident that the right data gets now
    in your datamart when you need it on a scheduled basis, so you can
    get you report each Monday for example, and do trend analysis on
    the most current data.
  - The scheduling capability does much more than just scheduling BV
    on a regular basis. The population of a business view can be chained
    for example to the successful execution of another BV. (6) to (8)
  - After the BV named "Manufacturing cost in Boston" has been populated,
    you may decide to always populate the BV named "Manufacturing cost
    in San Jose". (9) (10)
  - With this approach, you can make sure that several business views
    are always activated in a sequence. Even if the schedule for the
    first business view is changed, the other chained BV still run
    always after the first BV has successfully completed.
  - In this sample demo, you can verify that all the BV related to
    revenue at each location are activated using the sequencing
    capability: in the schedule window, you find which BVs are started
    automatically by the current BV, and which BVs start the current one.


Operation: business view definition (schedule triggered by program)
 1. double click on the "Corporate Income" BV in the Business View list window
 2. click on Schedule page
 3. click on the Program push-button
 4. click on the Cancel push-button in the Cascade Program window.
 5. click on the Cancel push-button in the Busines View window.
 6. click on 'Yes' at the warning message

Description:
  - To model more complex business processes, you may want after a BV
    has been populated to verify through a program that the data just
    being inserted satisfies a certain condition; in this case only,
    you want another BV to be populated. (1) (2)
  - After the "Net Income" has been generated, the CEO of our fictitious
    corporation, may decide to give some awards, but only if the total
    net income generated by all the products is over a certain
    predetermined value. (3)
  - The program named TRGAWARD calculates the total revenue generated,
    verifies that it falls over the predetermined value, and in that case
    the BV named "Corporate Awards is activated. (4)
  - This schedule type is called "conditional". (5) (6)


Operation: promote the status of the business view
 1. click on the BV named "Manufacturing cost in Boston_0002"
 2. click on the status menu in the menu bar
 3. click on "Promote to Test" (after this, you can promote to production)
 5. click on the status menu in the menu bar
 6. click on "Demote to Development"

Description:
  - You have gone through all the steps required to define the business
    view. You can now promote the business view to test mode and then
    to production mode (1) (2) (3)
  - In test mode, you can verify as an administrator that the table
    is populated correctly. In test mode, the schedule specified in
    the business view definition is not active. It becomes active when
    the BV has been upgraded to production mode.
  - Now demote the same business view back to development mode (4) (5) (6)


Business value of business views
  With business views, you are able to build summarized and
  aggregated data that can be the result of complex joins and unions
  between many tables. An example of such a BV is the business view
  named "Sum of revenue by location". You can browse through its
  definition using all the steps reviewed so far.
  Sometimes your operational data may be inconsistent, or different values
  might represent the same data. Data needs thus to be consolidated and
  cleansed. The advantage of data warehousing is to build consistent data
  with the same meaning for all users.
  One of the values of the datamart is the ability to build progressively
  higher level business views from existing ones, and enabling specific
  groups of users to have direct access to the data with the right level
  of summarization and aggregation.
  Another advantage of Visual Warehouse is the fact that the definition
  of each business view is stored in the control database, so that it
  is very easy to know the process that was used to collect and transform
  the data delivered to the end user. In fact all this definitional
  information can be exported to DataGuide, where a business user can
  browse through the existing information, its meaning, its accuracy,
  the transformations that were applied to build the data, etc...

          ....................................................

3.3 Examining the relationships between business views.

Section summary and objectives
  If there are only a few business views that have been defined, it is
  easy to guess the relationships between the BVs; but how do you quickly
  find these relationships when there are hundreds of BVs? Which BVs will
  be affected if you decide to modify the definition of a certain BV?
  In Visual Warehouse, there is a very easy way to display these
  relationships.

Operation:
 1. click on the View menu on the menu bar in the Business View window
 2. click on the Tree menu item under the View menu
 3. click on any + sign in front of a business view name
 4. click on the pull-down pushbutton on the Tree Structure line
 5. click on the line that reads "show the BVs that receive data from
    that BV"
 6. click on any + sign in front of a business view name
 7. close the Business Views for Finance window

Description:
  - Instead of displaying a detailed list of business views as you have
    done till now, you can display the busines views in a tree structure.
  - (1) (2) (3) You can display the business views that contribute data
    to a business view: this is a top down representation of the BVs,
    starting from the highest level of summarization, aggregation or
    refinement of the data. When there is no + sign in front of the
    BV name, it means that this BV was created from data out of an
    operational data source.
  - (4) (5) (6) The tree can also be displayed in other directions:
    for eaxample, you may be interested at viewing all the BVs that
    receive data from a particular BV: this would be useful if you
    consider changing the definition of a business view and want to
    see all the BVs that will be affected and that need to be demoted
    before you modify the BV definition.
  - We have also demonstrated how the execution of several business views
    can be cascaded: a tree can be represented to display the cascade
    relationships. In case of maffunction at execution, you can identify
    with this tree structure what are the BV that should have run but
    could not because a cascading business view did not complete
    successfully. (7)

          ....................................................

3.4 Monitoring the datamart activities.

Section summary and objective
  The data warehouse administrator can monitor from a single panel the
  whole data warehousing activity: such as the current processes that are
  running to populate the data warehouse, or when that activity is scheduled,
  or whether some previous activity could not complete successfully.

Operation: monitoring the work in progress
 1. click on the Operations menu on the menu bar in the IBM Visual
    Warehouse window.
 2. click on Work in Progress menu item under the Operations menu
 3. click on the New push-button under the "Business Edition Actions" box
 4. scroll down the Run Business View" window and find Business View name
    "Revenue - Atlanta"
 5. click on "Revenue - Atlanta" business view name
 6. click on OK
 7. click on the Scheduled push button (to have the list ordered on this
    column on descending order) (click once or twice)

Description:
  - As a Visual Warehouse administrator, you have a powerful tool to
    monitor the business views activity (1) (2)
  - In the Operations Work in Progress window, you find all the BVs that
    have been scheduled, the BVs that completed successfully, the
    BVs that failed, as well as the ones that are currently running.
    In fact, you can monitor the activities of multiple datamarts from
    this single Work in Progress window.
  - This list of activities associated to all the BVs can be viewed in
    ascending or descending order just by clicking on the column name
    item  (Business View Name, Status, Edition ID, Scheduled, Completed)
  - You can also filter just the business view activities you are
    interested in (click on the View menu and choose the Filter item)
  - Let us run a business view now: ((3) to (7))
  - In this scenario, you will see several business views that are
    being executed: this occurs since a cascaded schedule was defined
    on the business views (Revenue Atlanta starts Revenue Seattle,
    which starts Revenue Mexico, and so on). This is a very powerful
    way to have a sequence of activities performed in a completed
    automated fashion.
  - You can now run another Business View: try BV named Net Income;
    maybe you see that Corporate Awards BV is laso being executed.
    This is perfectly normal because of the trigger that has been
    activated. If corporate awards is activated, it means that the
    condition checked in the trigger program was verified as true.

Operation: monitoring the statistics and the logs
 1. click on any business view name in the Operations window
    (for example, highlight the BV named Corporate Income)
 2. click on Statistics push-button in the "Log Actions for selected
    edition" box
 3. close the window named "Business View Statistics Details"
 4. close the window named "Statistical Information for VWSAMP2"
 5. close the window named "Operations Work in Progress"

Description:
  - The Visual Warehouse system keeps detailed statistics about the
    execution of the business views. The statistical information
    includes for each BV the number of rows retrieved and the elapsed
    time for the fastest, the slowest, the largest and the smallest
    pull. (1) (2)
  - As an administrator, you may want to have a report about the
    execution anomalies, or summarized statistical information.
    You can use the Visual Warehouse technology to build business views
    that will collect just that information. There is a separate sample
    called VWSAMP that shows business views defined to collect this
    kind of statistical information. (3) (4) (5)

Business value:
  Visual Warehouse offers a single place from which to monitor all the
  activity in the data warehouse, and to easily locate the source of
  any problem, such as a data source not available, a BV that cannot be
  executed because the underlying BV is not current, ...

          ....................................................

3.5. Identifying the operational resources

Section summary and objective
  Before the administrator creates the business views, he first needs to
  register the resources (databases, files) that will be used to gather
  the information; it is also necessary to set up the definition for the
  target data warehouse.

Operation:
 1. in the IBM Visual Warehouse window, click on the File menu in the
    menu bar
 2. click on New Source menu item under the File menu
 3. click on the Cancel push-button in the Resource Type window
 4. click on Manufacturing in the Resource Name column
 5. click on the File menu in the menu bar
 6. click on Properties menu item under the File menu
 7. click on Agent Sites to open the Agent Sites page
 8. click on Database to open the Database page
 9. click on Tables to open the Tables page
10. click on the Import push-button
11. click on the OK push-button in the Catalog Import window
12. click on JOHN.COST12 table in the "Table list for COST" window
13. click on the "List Columns" push-button
14. click on the Cancel push-button in the "Column list" window
15. click on the Close push-button in the "Table list for COST" window
16. click on Security to open the Security page
17. click on the Close push-button to go back to the IBM Visual Warehouse
    window.

Description:
  - During the definition of the business views, the source tables
    and the columns that composed each table were readily available.
    How did the Visual Warehouse system know about this information? (1)
  - Before creating the business views, the Visual Warehouse
    administrator defines to Visual Warehouse the available systems,
    the source and target databases. This is done via the File function
  - You see in the Resource Type window that many data source types
    can be accessed (2).
  - Instead of defining a new resource, let us open a resource that has
    already be defined. ((3) to (6))
  - The definition of an information resource is done by completing
    information on several pages of a notebook.
  - In the General page, you specify some overall descriptive information
    about the resource, such as the name, a description, a contact person,
    and a location (7)
  - In the Agent Sites page, yoy specify which of the available agents
    can be used to access this resource: you can allocate different
    agent sites to work with different resources. (8)
  - The Database page identifies the name and type of the database
    to access. If the resource type had ben defined as a non-DB2
    family, you would find other database types than the ones listed
    here. After you have gone through this section, you can start
    again with the resource named SALESEMP, which is of type flat file,
    and you will notice differences in the notebook pages. (9)
  - The tables page displays the names of the operational tables that
    are known to Visual Warehouse (and recorded in the control database)
    It is here that you can import the table and column definitions as
    they exist in the operational environment, so that when you create
    business views, Visual Warehouse knows the name and type of the
    columns in the operational environment, and you do not need to re-key
    any of that information. ((10) to (13)).
  - If some columns in the operational data source will never be used
    in the datamart, you do not need to import the definition of these
    columns. (14) (15) (16)
  - Since many different resources could be accessed, you can assign
    specific groups to have access to specific resources: the security
    page lists the groups that have access to this resource.
  - (17) Now, I suggest that you go through the same section for a non
    DB2 resource (You can choose SALESEMP resource, which is a flat file)


Business value:
  Visual Warehouse keeps a metadata catalog with all the catalog info
  that has been extracted from various sources. It enables the
  administrator to easily gather metadata information about the sources,
  and to fill in additional descriptive information as desired. It can
  serve as an important reference tool to understand where the data
  warehouse gets the data from.

          ....................................................

3.6. Setting up the security

Section summary and objective
  Another category of functions that can be run with Visual Warehouse
  is the security setup. It provides a way to assign different roles
  and responsibilites to users (such as Business View coordinator,
  data warehouse operations controller, administrator, ...)

Operation:
 1. in the IBM Visual Warehouse window, click on the Security menu
    in the menu bar.
 2. click on the Privilege menu item under the Security menu
 3. click on OK
 4. click on the Security menu in the menu bar
 5. click on the Groups menu item under the Security menu
 6. double click on Budget in the Name column

Description:
  - Visual Warehouse also provides a set of functions related to the
    security (1) (2)
  - Visual Warehouse users may have different privileges depending on
    their responsibility: you can see the privileges associated with
    this user by selecting Privileges in the menu. In your installation,
    you may assign a person with the role of Business View definition,
    while the administrator would probably have Administration, Operations
    and Resource Definition privileges. (3) (4) (5) (6)
  - Users belong to one or more groups; each group has access to specific
    sources and targets. The group notebook lets you specify and browse
    through that information.

          ....................................................

3.7. Setting up agents

Section summary and objective
  In Visual Warehouse, several agents at different physical locations
  can be defined. Each agent can be responsible to handle specific
  business views. By setting up multiple agents, you can achieve even
  better performance because multiple agents can act in parallel to
  populate different business views simultaneously in the same datamart.


Operation:
 1. in the IBM Visual Warehouse window, click on the Settings menu
    in the menu bar.
 2. click on the Agents Definition menu item under the Settings menu
 3. double click on Manufacturing site in the Name column of the
    Agent Sites window
 4. click on Resources to open the Resources page
 5. click on Business Views to open the Business Views page

Description:
  - For performance and scalability, multiple agent sites can be
    defined; each agent has access to specific resources and handles
    a set of business views. The Visual Warehouse manager activates
    the appropriate agent when a business view must be executed.
    (1) to (5)
  - You can browse the different pages of the agent notebook to see
    how resources can be assigned to an agent, and to have a list of     s
    business views handled by this agent.

Operation:
 1. click on the Close push button in the Agent Site window
 2. close the Agent Sites window
 3. close the Visual Warehouse window

Description:
  - You have now completed the whole demo, and seen the major features
    of Visual Warehouse product.


.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-

We thank you for your interest in Visual Warehouse, and hope you will
realize the many business benefits this solution has to offer today!

