Our approach to creating enterprise systems has caused us to "hit the wall" when it comes to addressing the needs of large complex conglomerates. The scalability issue is directly related to the fact that our industry takes an "integration last" approach to systems. Since the inception of systems, our approach is to give a team a list of requirements, create a system based on the team's preferences, implement it and then try to integrate it. This approach fails to account for the challenges that arise when we try to integrate more and more systems.
To illustrate the challenge, imagine implementing a fully integrated system for a large government that encompassed healthcare and education business entities. The goal is to ensure that data is only ever entered once and securely shared with every other business unit that has an interest in it. Furthermore, we need to automatically aggregate data from all business units into a data warehouse for AI.
Unfortunately, our inability to easily integrate systems and aggregate data for reporting and AI would immediately preclude us from addressing this challenge with many systems from many vendors. Even if we charged a single ERP Vendor with providing a solution, they would not be able to do it. ERP vendors created their systems based on a siloed mentality, making it difficult to implement instances by business unit and automatically share data between all instances and aggregate data for AI.
The best we can do is stuff as much functionality in a single system and use row level security, creating challenges for tailoring the system to meet specific business unit needs. Such systems can only get so big before they fail. Instead, we need an open-source platform and methodology that is designed from the ground up to be customized to meet the unique needs of a business unit and has inherent capabilities for exchanging data and aggregating it for reporting and AI.
By doing so, we could limit the complexity of a given system to the needs of a business unit, making the solution scalable to address any size problem. The key features of this approach include:
- Users within a business unit would have all functionality they required in a single system.
- Common business functions would become parameterized modules that could be shared amongst business unit systems and then tailored to meet unique needs.
- Data would only ever be entered once and securely shared with every other business unit that requires it.
- Data from all business unit systems could be automatically aggregated for reporting and AI.
An open-source platform and methodology has been created and tested to accomplish the new mandate. The hope is that this will be the first step in addressing our scalability issues with enterprise systems.
Thank you for your attention.
コメント