Enterprise IT past and futur?
Enterprise IT never ending cycle
-
Enterprise IT is now a 60 years old industry and it still takes more time to implement most of large IT projects than to build a skyscraper and the maintenance of these systems take much more than cleaning teams.
​
-
We moved from centralised to decentralised approach for the different technical components but the complexity link to having different business components speaking one to another has not been resolved.
​
​
-
The common approach when 2 components have to be connected is the rule of the jungle. The “stronger” side imposes its interface to the weaker and provides at best a data dictionary that is rarely very helpful.
​
-
As we are moving to a service and API architecture the number of interfaces increases and the cost of developing and maintaining them is increasing and most of the cost of IT team in large enterprises is spent in connecting systems more than creating new features.
After long experience in large enterprise project Bailam founders have taken a unique approach of addressing the data integration without focussing on the data but on the information it represents developing an innovative methodology and platform to help all actors to share and collaborate.
The reality behind the line
-
All It architecture diagrams have a nice set of boxes connected by lines, most people focus on the boxes assuming the lines are the easy part. This line usually becomes the bottleneck of a project because each part focuses on their needs and defines their one language in a form of data model and assumes the others will speak the same language.
​
-
Explaining to a humain the exchange of information will be easy because they represent the same concept but the translation into different data models make it hard.
Bailam approaches allow each part to keep the focus on their side but ask each side to share in a human readable form the scenario they are covering. Then using machine learning Bailam can understand the language of each side to create a mapping process.