Vous êtes sur la page 1sur 3

Choose your ETL Tool : The best fit for your purpose :

The selection of right ETL Tool for the Strategic Datawarehousing Project for enterprise
level solution is very much critical. Once you decided the right ETL the 70% initial effort
is done. Once you finalize the ETL, then either you have to hire experts of this
Technology or you can arrange for proper training to make an efficient Technical group.
From this point of time you can start expecting a properly populated Datawarehoue on
which you can make desired reporting generation activities and analysis activities.
An enterprise-class ETL selection is a very challenging activity. If you are the person
among the committee of selecting the appropriate ETL Tool then first thing you need to
make a checklist for this purpose. Consider the following mention points while you will
evaluate the ETL Tool best fit for your Project Purpose.
• Is the price tag for the ETL Tool manageable with your IT budget? Of course, the
cost is important in a new system. Not only should the initial cost of purchasing or
creating a system be considered, but the long-term costs of maintaining the
system should also be explored. An evaluation of the overall expense should
consider a variety of variables including ongoing training, licensing fees for the
ETL Tool, maintenance and technical support etc.
• When you are comfortable with price tag, you should also look at what your
organization will get in return. Choosing an ETL that will lift the system what
will maximize the return on investment, or ROI, is a primary component of the
decision process for such kind of large investments.
• One of the important considerations for any large IT investments is to help for
removing your organization’s inefficiencies, which subsequently leads to a
reduction in operating costs. How much efficient is your ETL Tool for this
purpose?
• Is the human resource for the corresponding technology is readily available in
market, else is it possible to train people for making an effective technical group
who will really able to take the responsibility for quality and timely deliverables
for ETL development work?
• In order for your stuff to use the ETL Tool effectively and efficiently, will you
incur additional time and expense for training? Evaluate it.
• Would you be paying for a lot of features that you don’t really need? Does the
cost of the product, including extra features you do not need or would not use plus
training, provide you with the returns you need in order to justify the expense?
• Does the ETL Tool easy to use and readily comprehendible from Development
and Maintenance perspective?
• Is your ETL process is really integrable with your business process? What all
extra effort you need to do for integrate it with your pre existing business
platform?
• Does it meet all your requirement, Business and Technical both?
• For understanding the technical requirement analysis, please do consider the
below mentioned points where I illustrated the technical needs in detailed
manners.
• Check about it’s scalability. Whether is it supporting faster processing of large
volume of Data scattered through out the enterprise? You can perform peak load
testing to evaluate it.
• Will it able to support the growth rate of your enterprise wide Data volume for a
considerable long span of time?
• Is it scalable in terms of number of sources and targets?
• Is it able to extract data from heterogeneous source systems, scattered enterprise
wide, if it requires to meet your business need?
• Is it able to access all brands of relational and legacy databases, including those
on a mainframe through native gateways and is able it to take the full advantage
of query optimizers and bulk loaders?
• How effective the ETL Tool for Integration purpose? A real time requirement
drives ETL and EAI convergence. At present I am working for a US based bank
Datawarehousing Project where we need to extract data from 64 different source
systems scattered enterprise wide. We wanted a consolidation strategy. As all data
are coming from various platforms, so check out with ETL Tool gives you the
facility for a wide spectrum of connectivity and ability to standardize the data
those are coming from different sources, which will be feed to your ODS or
Datawarehouse through ETL process as a single true source of your Data. We find
ETL Tool named, Ab Initio is a very good choice for this purpose.
• Whether is it supporting various platforms? Some organization may like to deploy
their ETL Tool across various platforms.
• How much efficient is it to bring dirty data in shape? It is observed many
organizations want to skip the Data cleansing activity before ETL, but for
business aspect it is a very important and unavoidable step. So keep a check
whether the ETL is smart enough to take care of it (do not expect Fuzzy Logic
Level solution with it, if you desperately would like to go for data cleansing then
do not rely on ETL Tool, take the use of some Data cleansing Tool, like
FirstLogic, Trillium etc. ).
• For enterprise level development work it is expected your ETL Projects and
Objects are shared between the authorized groups of developers across
geographically dispersed location. So Check out whether the ETL Tool supports
Check in and Check out facility for version controlling and provide the sufficient
care for Locking on sharing.
• How much coding effort is reduced with the use of the ETL Tool? Is it providing
reusability, readily comprehendible processing? Is it providing easily
maintainable solutions? How well Graphical Development Environment is it
providing?
• Is it support process restart in case of failure from Checkpoint? Is it support
Phasing concepts? Is it support calling native codes for Data processing?
• Is it able to implement complex transformation logic and able to load a number of
Targets parallely.
• We need to check whether the components of the ETL Tool supports parallelism
and at what extent? I personally find ETL Tool Ab Initio as really remarkable
choice for this point.
• Is the ETL Tool taking proper care of Meta Data Management?
• Is the ETL Tool providing ready support for environment migration?
• Is it providing a repository as an integrated part of it or any other means for better
management of ETL Project and Objects.
• Does it supports scheduling or your IT stuffs need to put extra efforts with some
other technology to achieve it?

Vendor selection is a lengthy and arduous process. An enterprise-class ETL can be


evaluated on basis of all the attributes mentioned here above. Please be cautious about the
selection process, do not make a comparative study among the ETL Tools available in the
market on the basis of aforesaid points. Consider what points and aspects are most
important for your Business need and in those areas which is the best choice—consider
your choice then check whether is it satisfying all other business constraints assigned for
your project. If so then check out for real life essence with this ETL Tool either by load
testing the tool and by interviewing the vendor’s customer references. Analyze Vendor’s
customer case studies, observe what type of solutions they achieve at what kind
environment and at what extent they really able to meet their requirements.
Based on all of these factors, you must make a decision for selecting the Vendor for your
ETL product. The general rule of thumb is this: if your business problem is common, go
for the mostly available ETL product in market, but make sure that you do not want or
do not like to consider all other issue and aspects provided by the other ETL vendors
which are also in market with lesser market share. May be in context of your business
needs and issue the ETL with lesser market share will be the perfect fit.
To support my statement I would like to draw your attention to some crude statistical
data. The total market share for the most popular ETL Tool Informatica decreased from
25% (June, 2001) to 23%(June, 2004), where as the market share of ETL Tool Ascential
increased from 13%( June, 2001) to 15% (June, 2004) and for ETL Tool Ab Initio the
market share is also increased to 7%( June, 2004) in accelerating mode and it is expected
as a very potential ETL Tool for future Datawarehousing market specifically in
Insurance, Finance, Banking domain; where the a large data volume is needed to process
in faster time frame in a very secured and systematic manners, as well as it provides
connectors to all the conventional platforms like DB2, TeraData, Oracle, SAS etc.
Cheers,
Atish Banerjee
atish_ban@yahoo.co.in
Practicing IT solutions in Datawarehousing area in various verticals, like – Insurance,
Banking etc. Presently working with Covansys India Pvt. Ltd. as Bearing Point (KPMG
Consulting) Offshore Team.

Vous aimerez peut-être aussi