This blog is for Mainframe Modernization professionals to discuss and share perspectives, point of views and best practices around key trending topics. Discover the ART and science of mainframe modernization.

« October 2016 | Main | December 2016 »

November 16, 2016

Renew mainframes for digital readiness and new user experience [Part 4/6]

We briefly touched upon three modernization strategies from Infosys in an earlier blog. If you remember, it was stated that in order to address the challenges that come with modernization and provide a non-disruptive way to modernize mainframe systems, we have a new ART (Accelerate, Renew and Transform) approach. Detailing out the 'Accelerate strategy' in the previous blog, today's blog would talk about the RENEW strategy to Modernize Mainframes.

The RENEW strategy focusses on using the existing mainframe assets and exposing them as services or APIs - making them digital ready for enhanced user experience, reducing the development lifecycle time by employing the agile processes and automation through DevOps, and improving agility by modularization and rationalization of code modules and externalizing business rules from the code.

The below diagram depicts the different Renew Solution levers that could be applied for API enablement, business rule extraction and externalization, implementation of CI/CD on Mainframe portfolio and componentization and consolidation of business functions:


4. Renew_4of6.png


It's been observed that 43% of IT executives (a) believe that a failure to integrate with legacy systems is the biggest barrier to implement future initiatives and this barrier results in high IT costs*. There are various ways to integrate back with the mainframe applications and data and I have described the most popular methods below:

1. Emulator based integration: The Mainframe black screen (CICS, IMS etc.) can be integrated via Screen Scrapping technology tools like IBM HATS, Open Legacy, LegaSuite. This provides quick and easy GUI Integration and delivers a new user experience at a relatively fast time to market.

2. Process Integration: It enables monolithic legacy code to be exposed as service and easily build APIs. The Mainframe Services / APIs can be exposed to the distributed, digital systems. The different integration patterns utilize various compatible integrators that expose the mainframe components as services/ API's like CICS Transaction Gateway, CICS Web services (REST / SOAP), CICS Transaction Server Integration, MQ based integration, IMS Connect, zOS connect. The messaging to these APIs and services can be orchestrated through any ESB's or API Gateway. This option provides integration option to multi-platform applications and enables loosely coupled services.

3. Data Integration: It enables access to data from system of records repositories on zOS. The interfacing application is only dependent to fetch data from the Mainframe and business logic, and integration layer is built outside the Mainframe. This pattern is useful where the data can be fetched in one pass and the aggregation and business rules are applied in the integration layer outside the mainframe.

Just to give an example - for one of our health insurance client, as a part of a large digital initiative, we had to enable the "Find a Doctor" feature for various user provided search parameters. The book of record for maintaining the doctor information was existing on the Mainframe DB2 DB. On analysis, we realized that a few mainframe COBOL modules /API's were already available, that could be reused to realize this functionality. We were able to perform a process integration and use one of the prevalent integration pattern in the client ecosystem (CICS MQ based integration) and expose the mainframe components as services. This design option reduced the overall development time and helped in achieving the desired client goal, in a timely fashion.

Most of our clients also complain about how changing a small piece of code on their legacy mainframe applications takes a long time. When we analyzed applications for few of these clients, we discovered that most of the business rules were hardcoded and maintained over a period of time. The knowledge of those business rules was embedded in the code and SME knowledge had diminished over a period of time.

As part of our renew strategy, we recommend our clients to maintain the knowledge of business rules deployed and if possible, externalize them to be maintained by the business users. To reduce the time to introduce any change, we also recommend the adoption of CI/CD methodology for developing and deploying mainframe applications. The usage of CI/CD toolset helps in better productivity and quality with quick analysis showing application structure and relationships, free up MIPS for production use and eliminate delays by providing low cost environment, and accelerates delivery time with improved governance across the SDLC. We assist our clients in integration and configuration of the CI/CD tools, from IBM or Microfocus, in their eco systems and assist in formulating a rollout strategy.


Have you considered renewing your mainframe landscape to reuse the investments already made on mainframes?


*References: https://www.mulesoft.com/webinars/esb/unlocking-legacy-systems-enterprise-mobility

November 15, 2016

Unleash the Power of Mainframe using API Economy

The trend across today's industries is to move toward open web APIs. Enterprises are transforming their business models to embrace API Economy in a connected digital world with a large community of developers.

What has mainframe got to do with the API Economy?

Open API blog_jigsaw_small.png
Existing mainframe systems have been built up over the years and encapsulated enormous valuable business logics and data. These rich and mission-critical sources of business services and information are critical for digital transformation. As enterprises address business transformation, they will have to make their mainframe data and applications available as APIs in the total digital solutions. 

In my recent engagements with enterprise clients, clearly, both the healthcare and financial services sectors are becoming major drivers of digital innovation.  I have been frequently asked to advise how to revitalize, and incorporate enterprises' mainframe assets in digital environments. Here are some enterprise clients' business cases related to digital transformation:

financial services company, which handles majority of credit and debit processing, including prepaid cards, is looking to re-engineer their mainframe based complex and high speed transaction and data processing. Currently, I am leading an architectural proposal to modularize the mainframe portfolio and expose core capabilities as micro services and REST APIs to expedite future development efforts in meeting business demands on a timely basis.  Our proposal also addresses their skill issues with emerging technologies for decades to come. The proposal lends itself for a commercial project with defined outcomes.

This year, Infosys was chosen as the preferred partner for an American health insurance company's digital transformation initiatives. While our engagement teams were working diligently to deliver the transformation, this client also actively approached Infosys for mainframe integration. Their digital initiatives will not work without incorporating their family of systems, which are mainframe based mission critical business capabilities. The primary challenge they are facing is to expose their mainframe assets as REST services, with transactional tracking and problem determination capabilities across multi-components and platforms.

The other case is a Brazilian large banking and financial services company. This bank's core banking system resides on mainframe, and processes more than 2.5 billion transactions every day. With increased transaction volumes for digital banking, including a recent merger, and upcoming public/hybrid clouds, the bank is facing challenges for continuing integration with mainframe systems with same QoS and reduced MIPS. The mainframe revitalization proposals have been well accepted by the Bank's technical community for execution.

How to expose mainframe assets as micro services, and nature APIs?

Creating a successful API, however, is not easy. Creating APIs for mainframe data and applications is even more difficult, because it requires deep integration capabilities, especially in exchanging data between legacy and newly-written applications in the API digital world.

Solving such difficult and challenging problems of data exchange require certain technology like CAM (Common Application Metamodel). CAM is a standards based technology which facilitates data transformation from one language and platform domain to another, and also simplifies new applications creation. CAM is highly reusable and independent of any particular tool or middleware, and has been adopted and implemented as the underpinning in heterogeneous systems for enterprise application integration. Language based importers, which parse programming language source files and import relevant language metamodels,  have already been implemented in a variety of AD tools, and can also be re-used in any new tools to be created.

A comprehensive development tooling can therefore readily be developed by leveraging CAM and associated language importers to expose mainframe applications, e.g. CICS and IMS transactional applications, as micro services. Such tooling will read and parse legacy applications' interface declarations, e.g. COBOL copybooks, PL/I include files, and other source files, to generate REST / JSON services.

Further considerations also need to be evaluated regarding data transformation techniques. A development tooling can be designed for interpretive marshalling by generating metadata to be used during execution time. The very same tooling can also be designated to generate stub routines for compiled marshalling instead. In general speaking, compiled marshalling is faster than interpretive marshalling, though there are trade-offs between these two marshalling techniques. However, a hybrid approach between these two techniques should be considered to resolve the trade-offs.

Once REST / JSON services are created from legacy applications, these services can then be easily mapped as standard, language-agnostic interfaces to REST APIs, according to the OpenAPI Specification. These generated invocable APIs allow both human and computer to discover and understand the capabilities of the mainframe services. New apps can also be easily developed with these APIs to interact with mainframe assets with minimal implementation logics. 

With mainframe APIs, enterprises can now readily extend their digital reach for consumption in mobile applications, cloud applications and Internet of Things (IoT).

Subscribe to this blog's feed

Follow us on

Blogger Profiles

Infosys on Twitter


Categories