Why do Microservice integrations fail? And Path Forward

It had all started with the Internet. The Internet had brought a cultural change; the way people connect with each other or the way they share information. In the digital world, things had also moved from Intranet connectivity to the Internet world. The new paradigm shift was observed during the late ’90s when businesses started to think to interact with customers and other systems in the ecosystem, in a more modern way. This idea of interconnection between the various systems has erupted, and many integration tools and technology were developed. Enterprise service bus has come into existence to cater to such need of integration. However, monolithic systems became a bottleneck. Performance and scalability became a challenge.

Enter – Microservices Architecture! To bring speed and agility among distributed systems, Microservices became the lifeline of the IT landscape in the enterprises. What it does is separate the monolithic applications into independent services. All these services are independent and loosely coupled, connected through APIs and offer a seamless experience to the end-user. It also provides flexibility to the developer to upgrade or undertake maintenance of one service without affecting others. However, it also brings complications. Managing multiple APIs and network of connection is not easy, and hence it defeats the whole purpose of performance and agility when data connections grow with the period.

Microservices brings complications – With an increase in the number of services, complexity in microservices architecture increases, resulting in spaghetti network which no one wants to play with. In the current situation on COVID-19, when banks and financial institutions want to provide certain new services up and running, they are tangled in these complexities, which reduces their speed and agility.

Microservices Integration comes with a cost – Hmmmm, the organisation has to move to more agile and DevOps culture now. Microservices initiative will come with its own cost of the cultural shift. Also, services integrate with each other using API calls. Increase in the number of service calls, higher the cost of infrastructure to cater to bandwidth and servers for low latency.

Microservices can pose a security threat – Due to the increase in the number of applications or services and interconnections between them, security threats will become a challenge for the organisation. Any breach in the interconnection between the services can lead to access to the system, which poses a challenge to the organisation.

Microservices does not ensure Performance and Efficiency – Most of the time, interconnections are configured on request/reply protocol. It does not guarantee timely and effective communication. Message delivery becomes the bottleneck at application or service level, but the discovery of such service/ application is again very challenging. API management portals could be one of the solutions. However, they do not ensure the completion of the business task in the organisation within the stipulated time. The result is that the performance and efficiency of the system go down.

Complex MicroServices are New Monolith…!!

There! We said it. Now, the organisations have to revalidate their IT strategy when complex microservices become too big to handle. Automated business processes and business operations usually get affected. Enterprises in such situation should clear of the problem statement whether they want to:

• Orchestrate or Automate the workflow through services?

• Message – Persistence or Non-persistence way?

• Can they leverage their existing BPM tool or invest in new tool?

The challenge in front of the enterprise is to orchestrate and track the workflows with adequate performance and scalability without spending much on the new technological system.

The problem statement can easily be addressed with workflow management or Business process management tool. In our previous blog, we have covered this use case, where process orchestration and dataflow can quickly be processed with BPM engine. The pull and push capability ensures data drawing from one system into the business process, and data reaches to other systems. The communication can happen asynchronously as well as synchronously as well. With inherent case manager and process modelling, technical issue of identifying a process when it is stuck is easily addressed. It also addresses more pragmatic issues like authorisation and authentication at the user level, which is more complex in traditional microservices approach.