The list of major disruptions induced by digital activities is long. And these disruptions put the company under pressure like never before and make the information control a strategic issue.

Expansion of mobile technologies and proliferation of devices; explosion of social networks and behaviors that challenge models where each contact is a moment of truth; development of the Internet-of-Everything; advances in data analysis and artificial intelligence; demands on personalized and contextualized omni-channel experience; adoption of cloud computing; sources of information available more numerous and varied; volumes exploding face to processing needs more and more just in time ... The list of major disruptions resulting of digital is long. And these disruptions put the company under pressure like never before and make the information literacy a strategic issue.

Faced with these, companies that survive and thrive, regardless of their industry, will be those able to adapt their information systems to the radical changes of the new hyper-connected world. An information system that is becoming more than ever a System-of-Systems and multimodal hybrid and a major headache for enterprise architects. And if CIOs are aware today, the business departments are still struggling to understand the foundations needed by the digital IF and mandatory changes to operate in its architecture to accomplish. Yet an essential understanding for the SI business owners to support the strategic investments carried by the IT Department and especially to avoid their tactical IT vertical requirements as responses is still too often the case, sclerosing more agility and velocity information and made the company 3.0 system.

A complex digital equation whose resolution requires the implementation of a data-centric architecture, only able to ensure the effectiveness and efficiency of this SoS. An SoS usually describes an overview of the different IT systems of the company 3.0.


The Systems-of-Records usually designated by conventional systems, enterprise systems or the Back-end applications (ERP, CRM, HR System, Financial System). They constitute the bulk of the historic heart of the information system and have received most IT investments over the last thirty years. Their purpose remains the heart support business processes with significant de facto life. They are focused on transactional performance and rule over a highly structured data island which they guarantee high quality. They are the single point of truth of critical business information. Their architecture is monolithic and they were not originally designed to connect to the outside world. They dump their data most often in batch mode in the Systems-of-Insights, commonly named before the advent of Big Data, decision-making systems that produce reports and historical analyzes (backward / mirror analysis). If they are more than essential to the SI, the main purpose is not to continue to complicate but rather to lighten them. So if modernization work must be undertaken at their level, it should be to make them more agile and scalable. The imperative is to reduce the life cycle of their development including limiting their collateral impact and avoiding having to completely test them each time. For this, we must seek to transform their monolithic structure towards greater modularity and lightness, re-urbanizing them where this makes sense around micro-services, that is to say services with a single function , thought and not reuse efficiency (unlike SOA) and having a strong autonomy for deployment of automated and independently.

The Systems-of-engagement that form the edge systems, often referred to as front-end, that is to say systems In touch with customers & partners that connect all stakeholders of the extended ecosystem of the Networked Enterprise (website, mobile app, API Partner ...). Unlike Systems-of-record with whom they are integrated to achieve the transactional documents and obtain status, their life cycle is very short in nature to the connected client era. They can not be satisfied, as they were, to be just a natural extension and only offer a simple transaction capture interface. They should focus on the user and not the business process to create a differentiating experience, exciting, unforgettable. They must take into account all that is known to the user (its history, its preferences, its potential ...) and what he saw at the time of interaction (its device, its location, mood, behavior, time, weather ...). They must enable paperless interactions and omni-channel, real-time, mobile and social between all internal and external actors. They must be able to support complex, collaborative processes and a share of unexpected nonlinearity (Case Management). They have a zero downtime requirement and must be able to function even if the Systems-of-Record is unavailable.

The Systems-of-Insight which are historically the reporting and analysis systems in charge to inform about what happened today in the heart of the digital experience. They must evolve to meet the character more immediate of decision-making and their consequences, and anticipate, if not predict what will happen (forward / predictive analysis). They must be able to monitor and analyze different behaviors in the Systems-of-commitment at the time of occurrence (data in motion) and stored information and current transactions in the Systems-of-Records (data in motion). They must be able to analyze data of any kind to detect new opportunities, trends, risks and operational behavior and help achieve the appropriate actions to improve the performance and security of business. They must be integrated closer to the action and most at the end of the line. They must be permanently active, automated and self-learners to continuously improve reliability, refine their acuity. As Systems-of-engagement, they must be agile to meet the exponential needs analysis and high performance to support huge volumes of data generated by ever more automated and online customer interactions and objects that explode process. They should allow to conduct exploratory analyzes to answer to unknown questions in advance and get the best out of the data (Data Science). Finally, they must be able to filter events, analyze and correlate to power orchestrations effectively process.

The Systems-of-Systems Interaction or Systems-of-Interoperability and Systems-of- Integration are systems which role is to ensure the integration, coordination and collaboration between Systems-of-record, the Systems-of-Engagement and Systems-of-Insights. They should allow synchronous and asynchronous exchange and guarantee loose and standardized coupling between systems. For this, they must support the SOA patterns, Microwave Services and EDA (Event Driven Architecture).

The Systems-of-Operation which constitute the backbone of the runtime infrastructure must also evolve to deliver the promise of the dynamic provisioning environments (Software Defined Infrastructure) and continuous integration (DevOps) facing technologies growing and changing and facing the hybridization of infrastructure resources between internal (Data Center) and external (Cloud). They must provide security brick adapted to cyberspace and Internet-Of-Everything.

The business directions must understand that re-architecting the information system around this combination of SoS is vital and must support the CIO in his crusade. In terms of priority, Systems-of-Insight are those who can get the most business value because only short-term can transform the mass of information in decision and action. They are the keystone to propose Systems-of-Engagement smoother, easier to use, more personalized and contextualized and without deep processing of Systems-of-record. These can also benefit as well to improve decision making they wear.

If deployment can be gradual, it must simultaneously be accompanied by the establishment of a data-centric architecture supported by a Shared Data Backbone or System-of-Data that is to say an industrial base to capture, collection, federation, treatment and storage of all types of data and integrated into continuous integration process (DevOps). This base should be multilingual to bring all All-Data, that is to say all historical operational data and all digital data, the famous Big Data required for System-of-Insights. It must make them available and optimized to Systems-of-Insight, Systems-of-Engagement and Systems-of-Records by performing all pre-processing data operations (as implementation, filtering, aggregation ...) . It must of course be backed with Systems-of-Interaction whose capacities must be extended to support the variety, velocity and volume of data.

But to achieve the IT Department will have to evolve in parallel skills (eg Data Architect, Data Scientist), in its design practices (eg Design Thinking, Data-Driven SOA, μSOA) and development and data governance to ensure success of this digital re-engineering and the success of its digital SoS.

Source: Journal du Net, 28th April 2016 - by Jérôme Besson