Open Source adoption in a Mexicon Second tier Bank
-
Upload
wso2 -
Category
Technology
-
view
664 -
download
1
Transcript of Open Source adoption in a Mexicon Second tier Bank
Open source adoption in a Mexican second-tier Bank
by
Nelson RaimondIT Manager
Sociedad Hipotecaria Federal
Introducing myself
Self employed IT consultant for 8 years• Client server, web and DWH projects for Insurance,
Government, TV and Entertainment sectors
7 years. National Savings Bank• Implemented software platform for handling international
remittances and disbursing subsidies for poverty diminishing
2 years as IT manager. Federal Mortgage Society• Middleware renewal, BPM suite awakening from slumber
and open source demystifying
Sociedad Hipotecaria Federal(Federal Mortgage Society)
Mexican second-tier Bank
• Do not serve the public directly. Rely on the assistance of financial brokers to make SHF resources available to people. Brokers are responsible for granting and managing loans, from opening to conclusion
Promotes the growth of the mortgage markets
• Grants guarantees to build, acquire and improve of housing, preferentially for the lower-income bracket
Main products
Long-term loans for acquisition
Medium-term for rural housing acquisition
Micro credits for housing improvement
Key values in SHF. Efficiency
• Done right the first time with minimal resources and meeting the expectations of our clients and stockholders
Key values in SHF. Innovation
• Encourage creativity to develop new processes, services and products or improving them
Key values in SHF. Teamwork
• Achieve goals with a spirit of cooperation and shared responsibility, encouraging systemic thinking and the integration of value chains
Main objectives to align IT with business
• Achieve efficiency and decrease TCOMake more with less
• Increase levels of serviceWork faster. Keep your client happy
• Standardize processes using tools and methodologies based on open standards and best practices
Don’t reinvent the wheel
Obj. #1 Make more with less
Which were the urban legends in SHF related to proprietary software vendors in early 2010?
They won’t let you down. The best performance ever• Microsoft, Oracle, Sybase have best of breed
technology, R+D and worldwide support• Just get new boxes to avoid stepping on the corns
The more expensive, the better• Liabilities coverage are proportional to licensing and
consultancy costs. Don’t repair in expenses
They shield you from obsolescense. Won’t disappear
But…
• Now servers are underutilized. Plenty of them. Management became complex
• Obscure application integration. Lots of frameworks and worst practices
• Older versions of RDBMS & AS
• Have not asked for support
Obj. #2 Work faster. Keep your client happy
• To avoid underuse of proprietary technology, formal learning and getting certification is desirable
• “In theory, there is no difference between theory and practice. But, in practice, there is.”- Snepscheut
• Hiring VAR consultants will speed up things. A high hour rate assures you will get pros
• How fast are results visible?
• Is our people getting their hands dirty?
In practice
• Loss of control on how are things done. Everything but agile
• Junior consultants (newbies that charge like pros)
• No one questions the outsourcer skills or if they do right things
Obj. #3 Don’t reinvent the wheel
• Proprietary high-end technology should enforce us to adopt standards and best practices
• Instead of understanding the rationale behind those standards and best practices?
• Humans tend to apply the minimum effort law
• 20% depends on technology the rest, on people• “Why to make it difficult if we can turn it impossible?”
• “Don’t leave it for tomorrow if you could do it the day after tomorrow”
• “It always been done this way”
So far it seems we are doomed
Diagnosis
• Lots of paradigms. Fear of changing
• Low level of reusability. Copy and paste seems to be an easier way of coding
• Lots of services. Operations=services. All services must be WS. Everybody talks through the bus
• Complexity. Systems integration and mediation is rocket science
• Personnel avoids to get hands dirty
Prescription
• Break the proprietary software paradigm
• Think about business processes instead of systems
• Take advantage of being on the second-tier business: low volume of transactions, few customers, 5x12 instead of 7x24
• Reduce application integration complexity
• Get rid of anti-patterns and code less
• Let personnel be involved in the whole application life cycle
But where to start?
“Start with whatever is right in front of you” – Muhammad Yunus
First of all, your boss support• Needs to belief on change
• Needs to break paradigms
Begin from scratch• One or two simple use cases
• Quick hits
So what is the deal?
• Adopt a methodology
• Assemble a squad team
• Implement a lean middleware
• Follow David Wheeler approach
• Identify
• Review
• Compare
• Analyze
• Test drive and exploit the solution
Adopt a methodology
That is simple: KISS for instance
Assemble a squad team
Enthusiastic
Shares values & principles
High performance oriented
Self-directed
Middleware eval. Step 1. Identify
• Many options: open source and proprietary
• Proprietary software will take too long. Public bid is needed. Can’t be under evaluation for long time. Not an option now
• OSS is at hand. Will enable internal capabilities without having to pay in advanced
• Players found: Fuse, JBoss, Mule, OpenESB and WSO2
Middleware eval. Step 2. Review
• Googled deeper and gather information:
• Features
• Community and Customers
• Support and consultancy model
• Documentation
• Benchmarks
• Pros, cons and FAQ’s
• Embedded third party software
• Reviewed “Open source SOA”. Davis, Jeff (2009)
• Experienced customer service
Middleware eval. Step 3. Compare
Fuse• Asked for info but they did not reply
JBoss• Bounced me between US and LA support. After several
weeks they asked me for project info: size, number and type of servers involved, number of architects and programmers
Mule• Tried community edition
• More coding than configuring
• If you want more need to pay for EE
• Asked for pricing but they did not reply
Middleware eval. Step 3. Compare - cont’d
OpenESB• Received a demo from a SUN partner
• Too much coding
• Complex to model composite services
• Just ESB. JavaCAPS and Metro if want more
WSO2• Lean
• Quick install
• Easy to learn
• Excellent customer service
Middleware eval. Step 4. Analyze
• Decided to go with WSO2
• Took a glance to the samples
• Started playing with WSO2 ESB for simple proxy and services implementation
• Reviewed how main mediators work
• Also reviewed WSAS and Mashup• Not very lucky with BPS
Test drive
• “Rule #38: To think big, start small”–Alan Webber
• Define standards
• Maintain a base line
• Start with simple services
Exploit the solution
Spread the word• Adhere to standards
• Train the trainers
• Get feedback and refine
Continue improving• Add more functionality
• Try the unused features
• Compose
Do not go alone• Prepare to be part of the community
• Get professional support
Not everything is plain sailing. Another lessons learnt
• Good enough is good enough• Could be better or nicer but, do you need it?
• Feel the force. If something goes complex, then there must be something wrong from its design
• Upgrade to new versions as soon as possible• If it is too hard to upgrade, then something may be wrong
• Defining standards doesn’t mean that everybody is following them
• Naming conventions, fault handling, filing, coding style and how-to’s were the main issues
"We don't have mistakes here, we just have happy accidents”
Questions?