1. CUSTOMER STORY
Unnecessary code conversion avoided after SIG review
ING chooses intelligent alternatives
for legacy systems
ING OIB (Operations & IT Banking) degrades the stability of these sys- out of service as part of the exten-
has been involved in an extensive tems strikes directly at the reliability sive transformation programme
transformation programme since of the company itself. for which Van Dalen is responsible.
2008. Technology and process im- This programme is essential to
provement are included among Tried and tested applications permanently securing ING’s future.
the goals, as is creating the best The result is that many banks rely on It also contributes to ING’s overall
possible working environment for tried and tested applications that strategy of offering financial pro-
employees. All the IT systems are are sometimes decades old. This ducts and services in the way cus-
being subjected to a critical review applies to ING as well. At the time of tomers desire. This impinges on
as part of the transformation pro- writing, approximate- t h e r e p l a c e-
gramme. A number of those sys- ly 160 of this bank’s ment of the PL/
tems are legacy applications, some systems still use the 1 applications,
of which are decades old. SIG hel- PL /1 programming
“Phasing out as investigation
ped answer the question as to language and run on is required into
whether and to what extent these IBM mainframes. The
older systems the feasibility
legacy applications could be rene-
wed.
systems originate lar-
gely from the ‘NMB
reduced our total of running the-
se applications
legacy’, i.e. from befo- management o n h a rd wa re
As consumers, we have become re 1989, when NMB other than the
accustomed to always having ac- a n d t h e Po s t b a n k costs.” current main-
cess to the latest in digital techno- merged to form ING f ra m e s . Va n
logy. We consider a computer that and had to integrate Dalen believes
operates under Windows XP to be their individual data centres. “These that this is probably possible. “But
as hopelessly out-of-date as a mo- systems are used to process count- we also wanted to use the oppor-
bile telephone from 2005. At the less transactions every day”, says tunity to investigate whether a
same time, we accept that brand- Ans van Dalen, the head of the ING conversion to Cobol would be pos-
new software and devices almost OIB Transformation Programme. She sible. Cobol is after all considered
always suffer from teething pro- has freed up time for us during a by the financial world to be the
blems. In a business environment busy morning. “For example, the PL/ standard language for transaction
however, the situation is someti- 1 applications are part of our pro- systems. We chose to set up a
mes surprisingly different, particu- cess for some ATM payments. Securi- proof of concept. This involved
larly if the organisation handles ties and mortgage portfolios are converting the PL/1 application
sensitive or valuable information also handled by these systems. So that processes the PIN payment
and consequently sets extremely you can understand that we have a and cash withdrawals to a Cobol
high standards in terms of mana- strong desire to stick with stable system and then testing whether
geability, reliability and availability. and robust applications for activities the new application met require-
The financial world is the most like this and only replace them with ments.”
obvious example. Every single eu- new software if there is no other
ro, every transaction, every in- choice and also no loss of reliability.” Realistic risk assessment
vestment: they are ultimately a The conversion would take place
collection of zeros and ones in the ING’s IBM systems are on the list of automatically and, if successful,
ICT systems. And anything that systems that may possibly be taken we could use the same approach
2. to convert all the other PL/1 systems, small stages is much easier to ma- these processes in a second set of
at least that was the idea. However, nage. In addition, they felt it essen- securities systems that already
Van Dalen wanted to dig a little dee- tial to complete existed and had
per before starting the project. “Our the proof of some degree of pa-
experience with conversions has not concept first “We were able to rallel functionality.
always been positive”, she admits. “In before initia- The current data
many cases, the quality of the code
degraded and maintenance costs in-
ting the other
conversions.
save a lot of money could be migrated
and no more than a
creased. I knew SIG from an earlier They also indi- storage facility was
project within ING and decided to call cated that this
by migrating required for older
them and ask whether they could proof of con- data in order to
check the code quality of this test cept would
processes and data comply with the
conversion. This seemed justified as probably requi- statutory period for
the expense of converting all the sys- re more time.
to other retaining records.
tems to Cobol would amount to ap- This was be- After moving the
proximately 10 million euros. A lot of cause not all
applications.” data and processes,
money and I wanted a realistic as- the code came we were able to
sessment of the risk beforehand. out of the conversion correctly and phase out the older systems, which
Partly because of this, I asked SIG to more tests were required. At that in turn reduced our total manage-
state their findings as clearly as pos- time, we hadn't yet decided how ment costs. All of these solutions
sible in their report.” much time and capacity we needed seem very obvious, but this simply
to free up for the work.” demonstrates SIG’s added value in
The result was an eye-opener. SIG Van Dalen’s opinion. “Clearly, we
reported that the conversion project Different perspective don’t have the same level of per-
would become unmanageable be- Van Dalen decided on a change of manent in-house expertise as SIG.
cause of all of the testing. SIG’s con- course. “I appointed a new project Hardly surprising, as that is not our
sultants also advised against the ‘big leader and drew up a long list of core business. We are a bank and
bang approach’ that ING envisaged. questions for him to answer with our goal is handle customers’ fi-
“Our plan was to perform all the con- input from SIG. SIG played devil’s nancial data with due care. We
versions first and then test them at advocate in this process and encou- gladly call in external, objective
the end in one large batch”, Van Da- raged our people to think diffe- quality auditors to ensure com-
len explained. “However, SIG advised rently and outside the box. That pliance. SIG has the necessary ex-
us that converting and testing in was one of their conclusions: find perience, asks the right questions
alternatives. We also looked at and their consultants remorselessly
other businesses and investigated question and debate every aspect
how they solved their conversion and element of a project. The result
problems. It became clear that limi- is an objective picture that we as
ting the conversion as much as pos- insiders could never paint without
sible made good sense. The list of outside assistance.”
PL/1 applications included various
systems that we intended to phase
out anyway within a few years. So
why perform a major conversion
that would cost millions? SIG’s in-
put helped us come to the conclu-
sion that we could save a lot of
money by migrating processes and
data to other applications. That
was a completely different perspec-
tive.”
We were able to discontinue a
Picture: Ans van Dalen, Head number of securities applications,
Transformation Programme ING OIB. for example, and accommodate