An outside in approach considers change and layers of systems from the customer’s perspective (outside the organisation) down into the operations, i.e. customer experience layer, then interaction and process layer ( x channel) down to the application layers, data and finally infrastructure. This drives a vertical domain or micro-service service based view where services are assembled (orchestrated) to meet customer needs. It contrasts with a traditional IT model where systems are built ground up focussing on horizontal infrastructure, then data and apps with user experience bolted on top.
This drives a different model for the management and monitoring of services. The user experience can be delivered as an orchestration of services, some built and cloud hosted, some bought as SaaS and some on premise, with the ability to plug and play new services as needs and opportunities change. This means different monitoring tools compared to traditional IT estates, focussing on user experience, end point, security and transaction flows in the first instance. Each domain or micro-service, whether external or on premise, then needs defined service performance parameters and clear accountability, manifested in an SLA. The monitoring within the service ‘black box’ can then vary by service type and owner. The services need to be designed to ‘self protect’ and assume volumes and behaviours around can change to outside design boundaries. This ensures end to end integrity when multiple services are orchestrated together as no assumption is made in the domain/micro-service design of what it is combined with. This approach contrasts with a traditional IT design model which typically relies on boundary protection, but once past the boundary end to end models are developed assuming existence of all elements, with some limited load testing to validate performance under certain off design scenarios.
This outside in model is thus focussed on the ability to change or ‘plug and play’ services, with confidence that services are loosely coupled to enable independent releases. This compares with a traditional approach which relies on a heavily controlled management of change to deliver operational stability. In the new model one needs to be confident with constantly touching operational services as opposed to a heavily locked down traditional model.
In summary, if one is setting off on a digital journey, underpinned with a micro-services architecture, then from the beginining the management of those live services needs to be considerd. A SIAM (Service Integration and Application Management) model need to ve developed which can bridge both new and existing services, taking an outside in, customer view.