In this presentation you will learn about one of the hardest parts of IT transformations: a successful application modernisation approach. After only 50 years of IT experience we are in the rather slow learning curve of how to modernise our IT. The "just start over" approach has proven to be a nightmare in many cases. We arrived in an ongoing struggle between modernisation and legacy, mainframe versus cloud, stability versus agility. Learn where to start and how to create a step-by-step approach towards successful application modernisation. Spoiler alert: do not write off your legacy applications too soon...
NRB MAINFRAME DAY 03 - Toine Straathof - Legacy compute: modernisation is not just a hype…
1. Legacy compute:
Modernisation is not just a hype …
NRB Mainframe days
Toine Straathof
Digital Transformation manager
TStraathof@TRSAlpha.com
2. 2
Legacy has grown in the last 25+ years
Main issues …
1. Monolithic applications
2. Support core business
processes
3. Interrelation/overlap
4. Knowledge & scalability
of knowledge
Changes
Base
Platform
X
Platform
Y
Platform
Z
Complexity
?
3. 3
Legacy is bad!?
• Typical discussion: Old is bad, new is better?
• Performance, costs?
• Maintainability?
Example case: Life insurance renewal
Challenges of legacy:
• Strategic mistake of the original developers …
• Entanglement of legacy applications
• Knowledge of the application
• They seem to be expensive but are they really?
Example case: Replace Current account of a large French bank
4. 4
Mainframe versus Linux versus Cloud versus ?
• Each and every company with a mainframe has them: polarised, unstructured,
biased discussions
• Lots of shouting, not too much listening …
• Apples – Apples ???
− Costs: TCO!
− Performance?
− Stability?
− Maintenance complexity? (release upgrades)
− 7x24?
− Active/ Active?
• Rebrand mainframe towards optimal cloud computing?
So we will never leave the mainframe?
• For some parts of companies with certain specific functions: no
Is it anyway possible?
• Yes, of course. But be carefull …
6. 6
High level migration approach of legacy applications
• Multi year transformational programs
• Architecture based migration
• Complete rebuild of all systems vs parts?
• Buy or Build?
− Cobol ;-)
• Rebuild will replace legacy, but:
− Costs exceptional
− Timely, and needs focus
− Limits innovation
− Business value?
• Implementation strategy?
− 9 out of 10 situations: Core process!
− Big bang?
− Step-by-step?
− Hidden features?
7. 7
So, don’t write off your legacy too soon
Keep parts of the legacy alive when:
• It’s just doing its job
• Cut off front ends / mid offices
• Re-invest in knowledge & maintainability
• Availability of knowledge & resources is not a decisive issue
Do not (some bold statements …):
• Wrap
• Keep at all costs
• Position the legacy application & team as a dead end
• Migrate the legacy application towards a new technical platform
à Non biased analyses of the costs, maintainability & platform is necessary, but
really hard …
8. 8
Application modernisation? How?
• Not to scare you off but:
− Not too many (successful) examples world wide …
− Complex digital transformation so: 8 critical success factors
− Costly (from investment point of view)
• Creativity in your migration strategy necessary:
− Replace legacy by something better (…)
− Still be able to innovate! (many cases killed by this issue)
− Stability during a step by step replacement: involvement maintenance team
− Your maintenance costs will probably double during part of the migration
à This sounds and feels like a nightmare, and that is exactly what it is.
10. 10
Application modernisation: step-by-step approach
1. Create a target architecture, based on your business architecture (SOLL)
− Choice between buy versus build
− Keep some parts from the legacy alive?
− Choose your platform & integration strategy
2. Create a current architecture (IST)
− When not available
− Based on the current applications and integrations
3. Create migration strategy
− Step-by-step from IST to SOLL
− Max 3 months GoLive’s
− Create a dashboard (list of 7)
− Be aware that every in-between step needs to be production ready …
− Every step is a step in the right direction
− NO temporary solutions
4. Mistakes are extremely expensive: ask help if you lack experience in these kind of
transformations and follow a structured approach
11. 11
Lessons learned: a real life case
Large bank in The Netherlands:
• CRM renewal
• Replace 128 front end applications
• Replace legacy customer system, home grown, entangled with 2 older systems
• Culture … The old system is the best!
• 33.000 users
• Package implementation: challenge!
• All channels
• 5 year program
• Yearly budget 120-130 mio
Challenges:
• How to keep your budget alive over the years?
• Every 3 months go live
• Mainframe as the main database engine: tuning …