Você está na página 1de 6

Problematic application releases such as the Knight Capital trading software glit

ch in August 2012
and the
bumpy rollout of Healthcare.gov in late 2013 serve as cautionary tales in the an
nals of enterprise IT history. A
software bug that costs a brokerage firm $440 million in half an hour catches pe
ople s attention. But a solution
to the ber-technical problem of generating test data to empower proper, thorough,
speedy, secure application
testing as part of the development lifecycle? Not so much.
It s a rare exec who sits bolt upright in a cold sweat at 3 am over the sudden real
ization that great testing
tools are precisely what has been lacking all this time, says Joshua Greenbaum, p
rincipal analyst at
Enterprise Application Consulting. That s because the rest are sleeping the sleep o
f the ignorant.
Project failure is really death by a thousand cuts, and one of the issues that c
uts deeply is the problem
with software testing.
There are myriad problems with software testing, Greenbaum notes, but a big one
is that test data is given low
priority and executed using last century s tools and last week s college graduates.
Not all companies are using old-school approaches to application development and
testing, of course. Most
enterprises have embraced Agile methodologies based on iterative, incremental de
velopment. Agile replaces
the waterfall approach of old, and, when executed as intended, delivers mind-numbi
ngly fast development
cycles (called sprints) particularly important in the race for adding features and
keeping pace in the mobile
and cloud spaces.
Unfortunately, however, the number of sprints in any Agile project magnifies ine
fficiencies and increases
costs related to test data management. So if you re not applying Agile in testing,
or if you leave test data
management as an afterthought, or if you re still using manual methods for generat
ing test data, the lowered
cost and increased speed you thought you d get from Agile development move beyond
your reach. Why
adopt Agile if you re going to use outdated methods of generating test data? Manua
l test data creation
methods can t keep up with today s agile, dynamic environment, nor can they protect
you from the risk of
accidentally testing with sensitive data.
Garbage in, garbage out
That phrase is probably too simplistic to fully encompass what happens when you
attempt to develop an
application without quality test data. But it s not far off. As Greenbaum notes, i
t s a rare exec who fully
understands how skimping in this area can backfire in the long run. Consider the
specifics of life without an
automated test data management solution:
Wasted time, high costs. Gartner reports that development and testing, when comp
ared with other activities,
consume the largest slice of time 24 percent, or almost a quarter in an application
development
lifecycle.
1
2

3
2
In a recent survey conducted by Informatica
, tasks associated with managing test data defining
test data to be used in a test case, copying or creating test data, securing or
masking sensitive data, and
so on are taking up more than half a developers and QA team s time during the testing
phase. Clearly,
for enterprises engaged in hundreds of application development lifecycles at any
given time, the manual
process of creating and provisioning test data is inefficient and costly.
http://en.wikipedia.org/wiki/Knight_Capital_Group
Gartner IT Key Metrics Data, December 2011
TechValidate. TVID: 988-ED5-CCD
3
Cost Per Test Cycle = $2.7M
100 Days, 300 Testers, $90/hour
20% application defects due to bad test data
Application quality
Cost of a defect found later in the release cycle = XX
With Test Data Management
De?ne Test Data
Requirements
Improve Tester Ef?ciency by 30%
Cost Per Test Cycle = $1.6M
60 Days, 300 Testers, $90/hour
90% reduction in test data-related defects
Cost avoidance
Provision
Test Data
Provision Test Data Test Execution
Test Execution
Production
Application
Production
Application
Improve Tester Ef?ciency by > 30%
Reduce Testing cycles by > 25%
Improve application quality by 90%
Why You Need Test Data Management
3
4
ARE YOU A TDM
CANDIDATE?
If you answer yes to two
or more of the following
questions, you could benefit
from TDM and data masking
solutions
Are your testers and
developers unable to or
wasting time to get the
data they want, when
they want it?
Are bugs being uncovered
in production because
testers could not test
every scenario in your
applications?
Are you experiencing a

high percentage of testing


defects that are caused by
bad test data?
Is privacy a key
consideration of your
application testing?
What is test data management?
Test data management (TDM) is the process of creating and working with a data se
t that s
representative of that used by enterprise applications. A TDM solution can save
hours in
test data creation, build efficiency into your test process, ensure you avoid ri
sk of exposed
sensitive data, and ultimately reduce costs associated with testing. A great use
case of TDM
is Cognizant, a leading global provider of IT, consulting, and business process
services.
Using a test data management solution, Cognizant reduced overall test data manag
ement
costs for one of its clients by 40%
(see Cognizant s experience on page 5).
An enterprise TDM solution should help you meet these core application-testing o
bjectives:
Provision test data that models the existing production environment. Basic funct
ional
4
and regression testing should include a representative set of production data th
at is
considered baseline. Without this capability, test results are not reliable.
Provision test data that models a hypothetical production system for purposes of
new
development. If it s a new application, the design team needs to provide the testi
ng team
with the definition of a typical test data set to be used as the baseline. This
capability will
save testers hundreds of hours from manually creating hypothetical test data.
Test Data must not expose the risk of violating privacy laws. The challenge here
is
that obfuscated data must still be usable or you risk introducing errors. Tester
s must be
concerned about privacy because legislation has worldwide coverage from each US
State (such as CA SB 1386), to industry laws (HIPAA, GLB), to regional laws (Eur
opean
Data Protection Directive). Your organization must de-identify sensitive data th
at is shared
with third parties or outside borders and use reasonable practices (such as mask
ing) for
internal use.
Test data should be variable enough to allow what-if testing. Extreme-conditions (
aka
negative ) testing ensures that if something doesn t go as planned, the application
responds correctly. This is the only way to test how the application handles err
ors. Not
only should the data be variable, but the definitions need to catalogued and reu
sable
for control purposes.
Support many projects simultaneously with limited database administrator resourc
es.
Manually provisioning test data consumes significant database admin cycles and
introduces inconsistent results. Automating test data generation eliminates bott

lenecks and
frees up administrators to focus on other priorities.
Create smaller but representative data sets. This speeds the testing process and
optimizes
the testing team s time. Infrastructure costs can escalate when unencumbered serve
r and
database cloning is part of the testing culture.
Allow testers to provision test data in a self-service model. It is common to ha
ve multiple
testers share a single test environment. Whether test data is sourced from large
r test data
sources or synthetically generated, testers require the ability to reset test da
ta at
an individual or functional level to avoid interfering with or overwriting each
other s
data sets.
Cognizant Improves Tester Productivity and Reduces Customer Application Testing
Time by 20 Percent
with Informatica: http://www.informatica.com/Images/02472_cognizant-improves-tes
ter-productivity_
cs_en-US.pdf
4
Cognizant s experience
No matter your industry, most enterprise IT groups face the same business challe
nges:
Increase efficiency, reduce costs, and mitigate risk. When it came to generating
application
test data, Cognizant needed to solve all three challenges. As one of the world s l
argest
quality assurance and engineering outsourcers with more than 4,000 testers acros
s the
globe, Cognizant needed to deliver testing services more efficiently while ensur
ing testers
were not exposed to sensitive data.
Here s the background: Cognizant created testing Centers of Excellence for priorit
y
customers these centers control the entire application testing function within a c
ustomer
organization and range from 300 to 500 Cognizant resources deployed at the custo
mer
site. They were using an internally developed test data management solution to d
eliver the
Cognizant Testing Services (CTS). As their business expanded, however, their tes
t data
management solution restricted their ability to scale in line with customers data
growth.
The custom tool introduced challenges to cope with the wide variety of data, inc
luding
packaged applications such as SAP, Siebel, or other cloud sources. Cognizant nee
ded to
find an alternative test data management solution to maintain competitiveness.
It ended up choosing and deploying Informatica s Test Data Management solution, an
d,
along with its domain expertise and testing experience, Cognizant was able to de
liver great
results. For instance, it helped a U.S. retailer lower its test data provisionin
g time by more
than 95 percent and achieve a 33 percent reduction in overall test cycle time. T
he solution
also reduced test data provisioning time by 80 percent and test data size 60 per

cent for an
employee benefits management application project.
We are shrinking
customers development
cycles by working with
smaller sets of test data, and
lowering IT costs through
the use of smaller data sets
that require less storage and
fewer system resources.
Pradeepkumar Govindasamy,
Director, Cognizant Testing
Services
Why You Need Test Data Management
5
6
Why Informatica?
Informatica
Test Data Management enables your IT organization to create functionally intact
, secure test
data subsets of database applications from a single platform. With Informatica T
est Data Management,
your organization can:

Reduce the number of testing tools by leveraging a common TDM platform across di
sparate, heterogeneous
applications to provision integrated test data sets.
Speed deployments using automated metadata discovery, test data profiling, and p
ackaged application
accelerators for SAP, Oracle E-Business Suite, and Salesforce.com to name a few.
Reuse the same test data definitions across on-premises and cloud-based applicat
ions.
Integrate sensitive data discovery, data masking, and data privacy validation to
ensure compliance with
local and international regulations.
Leverage Informatica skillsets for existing Informatica customers.
Have confidence that you are using market leading technology Gartner positioned In
formatica as a leader
in its 2013 Gartner Magic Quadrant for Data Masking Technology
.
To learn more about how Informatica Test Data Management can help minimize appli
cation development
5
5
and testing costs, read more here: http://www.informatica.com/us/solutions/appli
cation-informationlifecycle-management/test-data-management/.
2013 Gartner Magic Quadrant for Data Masking Technology: http://www.informatica
.com/magic-quadrant-datamasking/
About Informatica
Informatica Corporation
(Nasdaq:INFA) is the world s
number one independent provider
of data integration software.
Organizations around the world
rely on Informatica to realize
their information potential and
drive top business imperatives.
Informatica Vibe, the industry s first
and only embeddable virtual data
machine (VDM), powers the unique

Map Once. Deploy Anywhere.


capabilities of the Informatica
Platform. Worldwide, over 5,000
enterprises depend on Informatica
to fully leverage their information
assets from devices to mobile
to social to big data residing
on-premise, in the Cloud and
across social networks. For more
information, call +1 650-385-5000
(1-800-653-3871 in the U.S.), or

Você também pode gostar