Microservices and API-First Design Stalking Mainframe Practices
Rome wasn’t built in a day.
While the rationale varies within software engineering specialties, the concept of brick-by-brick construction underlies much of component-based approaches to building and maintaining software.
Much less obvious – and the subject of many debates spanning decades – is just what those components ought to be, who should write them, and how.
Go-to’s were bad because they interfered with component engineering. Components ought to be reusable. One component should not interfere with the operation of another. Components should not be monolithic, in order to promote maintainability. If you’re associating this idea with the 80’s music like A ha’s “Take on Me,” you’d be wrong. Try a 1968 paper by one M. D. McIlroy. Way back then, he wrote:
“Existing sources of components – manufacturers, software houses, users’ groups and algorithm collections – lack the breadth of interest or coherence of purpose to assemble more than one or two members of such families, yet software production in the large would be enormously helped by the availability of spectra of high quality routines, quite as mechanical design is abetted by the existence of families of structural shapes, screws or resistors.”
That was then. Today the trend continues with two concepts that are transforming software in the current era: API-first design and Microservices.
Brick by Brick
Last year IBM published a Redbooks document, “Microservices from Theory to Practice: Creating Applications in IBM Bluemix Using the Microservices Approach.” The impact of the API-first software engineering movement on mainframes is being pushed from two directions:
(1) An explosion of service requirements for large users, and
(2) An evolving role for API in middleware and Service Oriented Architecture (SoA).
API-First and Microservices Cause Rethinking of Mainframe SDLC.
Credit: IBM Redbook “Microservices from Theory to Practice: Creating Applications in IBM Bluemix Using the Microservices Approach.”
Diverse and Demanding
In scenarios playing out in settings as diverse as Citibank and the 170+-year-old Economist magazine, IT shops are being pressed to support ever more diverse communities within their ecosystems.
In an Information Age column, Subrata Mukherjee, VP of Product at the Economist, told the media outlet that his IT team had been busy with initiatives including “. . . the Zuora subscription platform, Salesforce CRM, Gluu’s open source authentication and API access management, and a new data master management platform.” Motivating these ambitious concurrent undertakings was a new business model that needed to be data-driven. Legacy software implementations at the Economist sometimes left the company having to purchase its own data back in order to power new information-hungry business models.
“Data is king . . . We needed all our data, so we decided to put all our master data in Salesforce and then send it out to different systems through APIs,” Mukherjee explained.
In a very different setting, Citibank found itself taking on a similar transformation. Citibank serves 62 million clients in 35 countries, and, according to another report on APIs by Ben Rossi in Information Age, the bank “uses APIs to build many of its consumer-facing digital products.” Both Citibank and MasterCard took the approach of building cloud-based resources that mirrored production APIs. Both discovered that this approach accelerated developer adoption within a matter of weeks.
API-first, Not for Dummies
Elsewhere, IBM would like you to download their Limited Edition copy of APIs for Dummies by Claus Jensen (Wiley, 2015). It’s a document, which, they suggest, is “your guide to applying the power of APIs to business challenges ranging from changing business models to embracing a world of devices and sensors.”
In a chapter aptly titled “Managing APIs – and Not,” Jensen points to the need for API security, appropriate authentication, optimization, and manageability. API governance, he argues, is an essential element to effective use of APIs across the enterprise.
Auditing API use – resource consumption, access, configuration – can be implemented through efficient API gateways. As explained in “What is that Mainframe Really Doing?,” products like Syncsort Ironstream can be ideal tools to capture both direct and indirect traces of API. Such records can power debugging, tuning, analytics and forensics.
Microservices in z/OS
In a separate Redbook piece authored by Jennifer Foley et al. titled “How Walmart Became Cloud Services Provider,” the role of microservices in z/OS CICS was similarly highlighted:
“Microservices architecture is becoming the standard approach for developing, deploying, and managing applications and application components. Small teams work independently and use their technology of choice to create services that are accessed by easily consumed application programming interfaces (APIs). CICS provides the multi-language run time for such teams on a robust, secure, highly scalable z Systems platform. The resulting microservices are stand-alone pieces of code, each with a finite function that can be accessed by anyone on the team and independently deployed. The APIs can call other services or be used alone. It is all about the service that is faster, cheaper, and more flexible than traditional IT provided” (p. 2).
Managed Microservice Gateways in z/OS
Projects choosing to implement API-first, microservice approaches on z/OS can achieve visibility through robust logging. As Foley et al. point out, “CICS and z/OS evolved with a high degree of visibility into even the most granular level of system resource usage.” Using RMF and SMF records, plus third party tools such as Ironstream, can improve operational intelligence stability as well as secure the API-first ecosystem.
Chances are, these practices will also spin off tools that are helpful for developers – such as the ones IBM targets in “CICS and DevOps: What You Need to Know.”
A consequence of this trend is that the mainframe environment will have both more and smaller moving parts. As Beyoncé explains in her “Flaws and All” song: “All the pieces aren’t even in the box.”
To better understand SMF records and the value of having them readily available for analytics and visualization, download the free eBook, “What is that Mainframe Really Doing?”