The API experiment

The purpose of any experiment is to prove something, educate someone or prove that something is not as everyone thought it was. Let’s do a small experiment to better understand the API economy and the need for an API platform, API libraries, and microservices.

Most enterprise-grade systems for any purpose support many technologies, systems, and ways of working with data. Most of them try to have one or many high-quality APIs, support BI, IoT, ML, AI, and so on. All of these companies are working day and night to provide their customers with the best tools to improve their business, products, customer relations, and many more aspects that make life a bit easier. 

The experiment is easy.

Expected result: Check if the cost for individual or poorly managed integrations between mission-critical systems are too expensive.

Steps to execute:

Step 1: Open your favorite search engine and search for the term: “Top 5 enterprise [system e.g., ERP, CRM, Billing, PLM, etc.] and their features”

Step 2: Look closely, for every system there is a big chance that you will find previously mentioned abbreviations and many more. And there is a logical explanation for why they are there and most of them fulfill their functions very, very well.

To produce a beautiful buckwheat pillow you would need Product Lifecycle Management software (PLM) which contains your “Bill of Materials” (BOM). Not to waste time and resources on material sources, BOM should be connected to the Supply Chain Management (SCM) system to ensure raw material and production flow. Both systems need to be connected to a fleet of IoT devices that are involved in the production, storage, delivery organization, etc. There needs to be an integration with your Enterprise Resource Planning system (ERP), Billing system, and Client Relationship Management system (CRM). The last one will require multiple integrations with different marketing tools as well as Configure Price Quote system (CPQ). Don’t forget that you may need to have only valid delivery addresses to integrate with your courier service which means integration with their API (one or many) and the local government facility responsible for a valid address registry. All of this will require integrations with BI systems to monitor business intelligence, integration monitoring tools to find and fix data issues, and various quality assurance tools. And the biggest headache is only to come: you have a sea of legacy systems that mostly does the same as the above systems, but for specific clients and they are too expensive to remove (so you need integrations with many of the above systems). And all of that is running in a hybrid environment (part on cloud, some on-premises). And this is only touching the surface, there are hundreds of services that you need to keep in mind (Service desk software, virtual call center, workspaces, storage spaces, etc.).

Step 3: Calculate how many resources would you need to develop integrations between these systems and how much more to monitor, maintain and support these mission-critical integrations.

Expected actions: based on the outcome of the calculations gather and compare information on how much would it cost to maintain an API platform in comparison to simple point-to-point integrations, poorly managed integrations and what additional benefits would you get as a result of implementing a highly sophisticated API platform.

To make your life easier you can sneak a peek at our findings:

1. Each system publishes its APIs, a centralized API team only maps them together and delivers as ready API to the consumer
2. Each system can create its own “building blocks” – a small API that delivers required functionality. Now any mid-level configurator in a business team can configure new products in minutes
3. Centralized maintenance and support of integrations
4. API Library – any team can see what data they can get from any other system, get access and consume
5. APIs can be delivered as microservices – group necessary functions (APIs) in business domains, procedure points, or functions. If 1 microservice falls, the remaining services continue their work
6. If you change one microservice, you don’t have to re-test all E2E solution
7. Automatic horizontal and vertical scaling. Simple and easy throttling.
8. Circuit breaker – if functionality is not responding, data can be sent via a different route to establish continuity (if service 1 fails, data are sent to service 2 which is developed as a safety measure (if your ship is sinking and you get on a lifeboat and it starts to sink, you get to try lifebuoy) 
9. Guaranteed delivery

As you can see there are many benefits to a well-maintained API platform both in financial terms (small, but effective maintenance/support team; API functionalities are delivered inside respective teams, no need for an extensive API development team; centralized maintenance can help you avoid extremely costly downtimes; cost reduction for systems that licensing relies on the number of clients or integrations, etc.).

To find out more about these and many more benefits get in touch with us and our experts will be happy to communicate and share their expertise!