Scribd
Upload
314 views

BI Architect Questions

The document discusses the roles of several types of data architects including: - Data warehouse architect - Focuses on data modeling for the data warehouse as well as ETL, infrastructure, servers, and physical database design. - Data architect - Focuses on data modeling for transactional databases. - BI solution architect - Responsible for both the BI front end (reports, cubes) and back end (data warehouse, ETL). - Difference between roles depends on their focus area but there is often overlap of responsibilities.

Uploaded by

Anonymous L6bTPH
Copyright
© © All Rights Reserved
Available Formats
Download as TXT, PDF, TXT or read online on Scribd
314 views

BI Architect Questions

The document discusses the roles of several types of data architects including: - Data warehouse architect - Focuses on data modeling for the data warehouse as well as ETL, infrastructure, servers, and physical database design. - Data architect - Focuses on data modeling for transactional databases. - BI solution architect - Responsible for both the BI front end (reports, cubes) and back end (data warehouse, ETL). - Difference between roles depends on their focus area but there is often overlap of responsibilities.

Uploaded by

Anonymous L6bTPH
Copyright
© © All Rights Reserved
Available Formats
Download as TXT, PDF, TXT or read online on Scribd
You are on page 1/ 2

01. How to analyse the Execution Tree of a Query Plan.

02. What are the ideal performance conditions to be taken care with logical read
s & physical reads.
03. What are filtered indexes. What are advantages & disadvantages associated wi
th it.
04. What is the prominence of cross apply while using Scalar values and Table va
lued functions.
05. What are the disadvantages associated with Common Table Expressions.
06. What do you know about the column stored indexes.
-Describe advantages of the CIF architecture versus the bus architecture with co
nformed dimensions. Which would fit best in our environment given [some paramete
rs they give you] and why
-Describe snowflaking
-Describe factless fact tables.
-Draw a star schema of our business
-Describe common optimization techniques applied at the data model level
-How do you handle data rejects in a warehouse architecture?
-Describe common techniques for loading from the staging area to the warehouse w
hen you only have a small window.
-How do you load type 1 dimensions
-How do you load type 2 dimensions, and how would you load it given our [insert
business particularity]
-How would you model unbalanced hierarchies
-How would you model cyclic relations
-What major elements would you include in an audit model?
-How would you implement traceability?
-----------------------------What is the difference between a data warehouse architect and a data architect?
How about BI solution architect?
Both data architect and data warehouse architect do data modeling, as in ERWin s
tuff. Or Embarcadero ER Studio. But a data warehouse architect is more specializ
ed on designing the data model for a data warehouse, whereas a data architect is
more specialized on designing the data model for databases used by transactiona
l systems.
A data warehouse architect does a lot more than just data modeling. They also do
the ETL and the infrastructure. These are areas that a data architect doesn t do
normally.
For a data architect to be able to call themselves a data warehouse architect, t
hey don t only need to learn how to create a data model for a data warehouse (as i
n dimensional modeling). But they need to also understand the ETL architecture.
And they need to understand the servers. For example, they need to be able to sp
ecify the specification for the production data warehouse servers, i.e. CPU, mem
ory, disks. And other server stuff like clustering, mirroring and DR. And they n
eed to understand physical database stuff too, like table partitioning, file gro
ups and materialized views.
In my book I specify there are 2 sides of data warehouse architecture. The first
one is the logical architecture and the second one is physical architecture. A
warehouse architect needs to understand both.
A data warehouse architect in my opinion demands 4 separate skills: data archite
cture, ETL, database platform and physical infrastructure. By database platform I
mean SQL Server knowledge, Oracle knowledge, Teradata knowledge, Netezza knowled
ge, etc. For example, SQL Server 2008 Data Warehousing Features is a must know for a
DW architect in SQL Server. Whereas UPI & USI is a must know for DW architect in
Teradata. If we design a DW on Oracle 11g R2, we need to know Oracle specific DW
features, such as Initialization Parameter Settings, Partitionwise Joins, Compr
ession, Parallel Execution, etc.
A BI architect is more on the application side, as in SharePoint architecture, H
yperion architecture, Reporting Services architecture, and Analysis Services arc

hitecture. For example: a BI system where we have reporting services running on


SharePoint, plus excel services and PPS services on SharePoint, and SSAS cubes t
oo reading from a warehouse or mart. And on top of that some custom .NET coding
for authentication or customized security. Plus they allow some self-service BI
using Qlikview or PowerPivot.
Back to the data warehouse architect, the ETL aspect of the job is sometimes qui
te demanding. There is something called ETL architecture in warehousing, which i
s basically
a) the architecture of the overnight batch, i.e. the structure of the tasks and
workflows, the execution order of the workflows, the backup, the reconciliation,
the checking and alert, and the data quality. The overnight batch is not only a
bout data loading / ETL. It also has: data serving elements, i.e. processing rep
orts (stored as PDFs, ready to be served to achieve split second response time),
refreshing OLAP cubes.
b) the architecture of the continuous feed throughout the day for real time ware
housing
c) the physical infrastructure of the ETL, i.e. the servers, the databases, the
data connections
d) the methods to extract and load the data i.e. sliding window, swap partition,
flipping twin tables, identifying incremental extraction, changed data capture,
change tracking mechanism, (filtered) replication between warehouse and mart (o
r mirroring), how to extract data from cubes.
If a data warehouse architect only understands dimensional modeling, the company
will have problems in the ETL and infrastructure. The servers might not be suit
for purpose , for example disk configuration is not optimized for warehousing.
How about BI solution architect ? A data warehouse is the back end. Business Intell
igence is the front end. Data warehousing is about the data model, the ETL and t
he databases. BI is about the reports, the OLAP cubes, the analytical applicatio
ns, the data mining, the KPIs, the dashboards, the score cards, and the performa
nce management. It is a common perception that a BI solution architect is a front
end person. An application person. This perception is incorrect.
A solution architect is responsible for the whole solution. Not only the front end
, but also the back end. It is impossible for him to be responsible for the whol
e solution without looking after the back end. In the case of a BI solution arch
itect, he or she needs to look after both the BI front end (reports, cubes, perf
ormance management, etc.), and the back end (data warehouse, ETL). It is impossi
ble for a BI solution architect to be responsible for the BI without looking aft
er the data warehouse. In some companies, instead of calling the role BI solution
architect , it is called data warehouse solution architect . It doesn t mean that the
role is only responsible for the back end data warehouse, but he is also respons
ible for the BI front end. In some companies, they have not only one but several
BI solution architects, each responsible for a certain area.
There are several different types of architects in IT, for example: system archi
tect, data architect, solution architect, information architect and enterprise a
rchitect. This segregation of duties only happens in very large group of compani
es (enterprises). For example, a banking group with 3000 IT staff. If the IT sta
ff is only 100 usually the architect does multiple functions. A system architect
is responsible for infrastructure, including networks and servers. I ve mentioned
about data architect and solution architect. An information architect is respon
sible for the flow of information throughout the enterprise, including databases
and data quality. An enterprise architect is responsible for all the applicatio
ns in the enterprise, making sure they run in sync and adhere to the standards.
------------------------

You might also like