{{announcement.body}}
{{announcement.title}}

From Zero to API in One Month!

DZone 's Guide to

From Zero to API in One Month!

Companies have many questions and they are quick to block an API initiative. That's why I'm advocating an MVP approach.

· Integration Zone ·
Free Resource

There are still several companies today that are not put on APIs and are still wondering how to do it. There are many questions and they are quick to block an API initiative. That's why I'm advocating an MVP approach.

lego blocks

This MVP approach may come as a surprise. We will easily ask ourselves questions about security, techniques, organization, API roadmap, the definition of good practices, and a framework that will seem essential. However, the questions are so numerous that we quickly find ourselves stuck. But really, an MVP approach could help you to start!

The Benefits of the MVP Are as Follows

  • Demonstrating the value:
    • By setting up the first API in MVP mode, it is possible to communicate more quickly to the operational what is now possible to do thanks to APIs.
  • Demonstrate feasibility:
    • Complex Legacy System, if siloed, there are many reasons to believe that it is long and complicated to implement APIs. Thanks to the existence of many Integration architecture patterns, we can break the locks! 
  • Get feedbacks:
    • Rather than IS audits and maturity audits, which are long & practically relevant to discuss, there is nothing like regular REXs to get a closer look at an API program. 
  • Prioritize initiatives :
    • Thanks to the first feedbacks, it is possible to know which initiative to prioritize. Should governance be studied? The integration strategy? Internal communication?
  • Reducing risks:
    • This allows for a fully integrated approach, with Go/NoGo decisions at every stage. You will, therefore, run a risk reduced to its most congruent portion.

This MVP approach can quickly result in a first API in a month, certainly imperfect, but usable. Do you think there will still be questions after a month? It's crystal clear: Yes! And you will even be able to prioritize them!

The Approach Can Be as Follows

  • Scoping (First week):
    • Choosing an API candidate: 
      • You will first choose a first API according to its ease to be instantiated and the value it will bring. Aim for the Quick Win!
    • List the expected: 
      • Several technical, functional, and other constraints are likely to be issued by the stakeholders. Collect them!
  • Visioning (Second week):
    • Architecture definition: 
      • You then need to define the target architecture of your API Manager. An On-Site? cloud solution? What matters is to do it quickly. It's not complicated to change the API Manager, you'll have time to change it later. Aim fast and cheap!
    • API definition: 
      • It then makes you define the API. See with operations and developers. Let's get to work!
    • Validation: 
      • Finally share your API and your architecture with everyone, to validate it.
  • Building (Third week):
    • Installation of the components: 
      • So it's time to install your API Manager!
    • Developments: 
      • And of course to instantiate your API, with its interface but also to develop the interface in question if needed.
    • Integration and tests with consumer systems: 
      • And yes, you have to test :)
  • Retrospective (Fourth week):
    • Technical & business REX: 
      • You can make technical and business feedbacks, by making a demonstration of your API.
    • Upcoming initiatives to prioritize: 
      • And it's the right time to list the next initiatives. As the feedbacks has just been done, there is no question of stopping everything. Keep up the good work!

You will be able to continue your API program with the work of prioritized initiatives, but most of all, keep the momentum going! And don't forget to communicate about your work!

Topics:
api management

Opinions expressed by DZone contributors are their own.

{{ parent.title || parent.header.title}}

{{ parent.tldr }}

{{ parent.urlSource.name }}