Decision Systems In Rail

The entire transportation industry is exposed to massive disruption due to the significant improvements in safety, efficiency and reliability that can be achieved by applying modern information technology to age-old business practices and processes. It is estimated that between 30% and 50% of all functions currently performed by humans in the transportation industry will be replaced by automated and semi-automated systems within the next ten to fifteen years.


The huge infrastructure assets of major railroads provide significant protection against disruption from outside the industry, but internal disruption (through rapid adoption and deployment of modern and emerging technologies) is vitally necessary if railroads are to meet the safety, efficiency and reliability standards that their customers and other stakeholders will demand.


Just as building the rail infrastructure was fundamental to the creation of the industry, building the right information and decision systems infrastructure is now fundamental to the creation of an optimally performing future rail network.


Due to the enormous power of artificial intelligence and machine learning systems, this future data infrastructure must facilitate the aggregation, normalization and cleansing of all data from all sources across the network, so that this data can then be effectively utilized by the relevant decision systems. In the modern context, these data sets will inevitably include huge troves of unstructured (image, video and text) data as well as the more traditional structured data that current systems are designed to manage.


The decision systems that will harvest all this data will operate as frequently at the edge of the network as at the core. (Think of how powerful the edge processing systems are in a Tesla automobile!) Importantly, any given decision may require any of many data elements to ensure it is optimal for overall rail network performance. Thus, data will need to flow unimpeded in many directions, but particularly both from edge to cloud and from cloud to edge.


These next generation data systems will not only be huge, but will be hugely complex and very fast changing. They must be conceived and designed to be massively scalable and highly adaptable while remaining tightly secure, extremely robust and completely reliable.


The expertise to design, build and operate this data infrastructure of the railroad of the future generally does not reside within the current railroad industry, but resides only within the enterprises who operate at the leading edge of the technology industries. One only needs to look at the enormous growth of Amazon Web Services to see how much outsourcing of data systems infrastructure has already occurred across multiple industry verticals. (AWS revenue has grown at a 55% compound rate since 2013, despite dramatic price reductions over the same period.)


To survive and thrive over the next decade, railroads must immediately convert from seeing their core information systems providers as vendors to seeing them as essential partners with whom engagement is deep, continual and collaborative. Optimally, rail management should engage multiple (carefully chosen) technology partners in a functional and highly agile consortium, working together to design, build and operate the entire data management and decision system across the entire enterprise.


The time to choose this approach is now!


Originally posted by Richard Arnold  |  Executive Director