'An API romance'

Chapter 1 - The scene

Thai was 3 years old when he picked up a textbook and scribbled in it crudely with a pencil. This was ironic because although his future was to be defined by technology, his first chance interaction with statistics was as low-tech as it could have been. The reference was his father’s who worked as an engineer for a manufacturing company in the city, and although Thai didn’t comprehend at his young age, the act of crudely mimicking the strange hieroglyphics must have instilled a deep-seated confidence in the subject which then endured.

Thai was bright, but never top of his class. The competition at school was fierce and although his parents had saved hard and pulled official strings in order to get him into the best school in their province, the expensive private tutoring that many of his fellow students were privileged to was out of his reach. His imagination however was beyond that of his peers, and aged 12 he was writing crude mobile phone games and beginning his journey into the world of technology.

At university, Thai studied hard and was an unusual natural in statistics. His curiosity and innate confidence in the subject led him to play around with his own ideas; he enjoyed creating algorithms in R-based analytics frameworks which drew on data he sucked from the internet using web-spider programs, the frameworks for which he had downloaded for free. By now, even home computing was ‘in the cloud’ and so Thai was not limited by the processing or space of his second hand desktop computer which would have been the only option a few years before. Besides, 'cloud' was now ubiquitous; the population density in his country had eventually resulted in a cost-effective internet backbone which would have turned anyone in Europe, the US, or especially in Australia crimson with envy just ten years earlier.

Thai took the data and attempted to reverse-engineer all sort of unusual correlations using his knowledge by building statistical learning trees. He found a way to extract social media comments en masse, and turn them into a prediction of any commentator’s social background, emotional state, education level, and other factors. Although Thai’s English was reasonable, his algorithms were language neutral; he used the Google translate service to identify languages automatically, and the patterns in the writing when compared to known commentators such as politicians, business people, and anyone whose background could be ascertained through their comments or web-searches, revealed all through the power of his algorithms.

Incidentally, Thai certainly wasn’t the first to come up with this discovery, nor were his algorithms necessarily the best. Such analysis had been in the realms of governments and social media companies with far greater resources since at least ten years before Thai was working on his. Like anything, turning his algorithm into a commercial success would require dedication, great execution, marketing, contacts, and most difficult of all for him and his family, a large amount of capital.

This is where the story could easily have ended devoid of romance, as it would have in the past for many potential inventions and discoveries without access to the right contacts and commercial backing.

Chapter 2 - The technology

In the past decade, technology had continued to change most of the corporate world. It was normal now even for large and security-paranoid organisations to utilise software-as-a-service systems hosted outside of their networks. In fact, increased security breaches and highly publicised incidents over the past decade had almost relegated the concept of private and public networks to the museum shelf. Encryption was now standard in virtually every communication on any network, and public and private networks had to some degree fused. The further development of virtual computing and advanced management and control systems, had allowed composite applications to be created out of different systems and services hosted anywhere in the world. Thai’s university for instance used one such service for student life cycle management, and a different service for lecture scheduling and rostering which were integrated together such that students and lecturers could get the information they needed without ever knowing that there were two systems involved.

With this boon of specialist applications available online, corporations became gradually disillusioned with their monolithic systems whose value, in practice, was gradually being eroded by the return of the best of breed fashion to implement these new cloud application services. After all, their functionality and domain-specific innovation were better than the traditional big enterprise system providers could offer in their large modular systems. As a side effect, most corporate IT landscapes had been slowly returning to the bad old ages of spaghetti-integration, and with the politics and confusion of a continued dark decade of bi-modal IT combined with prolonged economic difficulties and cost cutting, many were desperately hoping for a new light.

Amidst this however, a revolution was slowly taking place thanks to the creation of a new Platform for application services. At its core it offered a comprehensive but standardised data model for almost any type of business data or transaction, which were related together using a sophisticated hierarchy management and processing system based on graph database technology. Products, locations, organisation structure, the relationships between nodes, and the subsequent modelling of an entire organisation was contained within this. An advanced workflow processor allowed the life cycle and processing of the documents and transactions to be tailored and linked back to the hierarchy, so as to match the individual business requirements of every scenario in the organisation. Inventory for example was just a relationship between nodes in physical location and product hierarchies, and a price a relationship between nodes in product and business partner hierarchies.

This was a nice new model but not massively different to many other ERP and other systems previously developed, however the Platform as a whole had some big differences. Within the workflow processing was the capability to link the documents and data objects to custom application processing functionality encapsulated into discreet services via a powerful API management system. These custom application services all sat on top of an underlying technical infrastructure platform which allowed the corresponding service application to be instantiated in a secure container as needed, with dynamic and elastic infrastructure. All data and any user configuration for each service was stored in the underlying platform thus ensuring security, and a group of other services provided by the underlying platform ensured standardised monitoring, error handling, ticket and problem management, service level monitoring, and so on across all the different services that ultimately joined together and as a whole, started to replace the workflows and processing steps provided by traditional corporate IT systems.

The beauty of the Platform was that the application services themselves were not pre-defined as would have been functionality inside a traditional ERP system, but instead were provided via a global ‘service store’ marketplace. Because of the standard nature of the underlying data and the flexibility of APIs provided by the platform, anyone who wanted to could write a new piece of functionality as a service which would then be compatible with a large number of businesses globally, if they wished to use it. Ultimately it was the commonality of the data which was the key, and which facilitated the network effect and ease of adoption.

Facebook continued to provide a standard for posting holiday pictures, arranging social events, and exchanging messages. YouTube continued to do the same for entertainment and advertising. The Platform did a very similar thing, only for business transactions and functions. It was a lot more complicated than either of its technological predecessors but once it was established business users quickly couldn’t imagine a world without it. IT departments began to change, and the long anticipated vision of IT as a ‘service broker’ started to come to fruition. Systems and operations finally began their journey into one.

In time, order management, service management, pricing, promotions, forecasting, transportation, logistics, financial planning, and almost any other service imaginable sprung up, with the majority of the pioneers initially being individuals and small companies.

The Platform didn’t suit everyone. Some companies continued to purchase and upgrade their ERP systems with all the usual heavy customisation. There was a degree of risk in using the platform to start with, but as time went on and services became established, for those companies for whom the economic benefit of standardisation versus customisation paid off, they flocked to it as did thousands of developers and teams around the world.

Not ‘open source’ but ‘open service’; another stepping stone in the global information economy.

Chapter 3 - The romance

Thai discovered the platform and decided to write his algorithms as application services on it. He didn't need to worry about infrastructure, security, memory, disc space, or bandwidth. The platform provided him with all the tools he needed to manage the service, including developer tools, diagnostics, monitoring, reports, and issue management. Customers data and specific configuration was always segregated and safe inside the platform, and their transaction data always kept safe and encrypted inside the secure container technology. Thai couldn’t steal his customers’ data, even if he wanted to.

Best of all, the platform handled billing, and Thai defined his own standard service levels and charging model, and the platform then automatically charged customers for their service usage wherever they were in the world, with Thai receiving a good percentage of the income after the Platform took a very surprisingly small operating charge. Most customers paid for Thai's service as part of a package of many other application components and services and the Platform helped to provide a predictable overall commercial model. Overall the costs were significantly less than for a traditional modular ERP or separate best of breed applications integrated together.

Any issues that were created inside the service when something went wrong, or as the result of a customer inquiry were also captured in a standard way as part of the Platform operation, and as a result of this Thai was able to purchase a service from another company which helped him provide a multi-language help desk, communications and responses in order to help provide great customer service. Soon people all over the world were using his service as a small but important part of their overall business systems. One customer even tried to use his algorithm to work out when their employees were having a bad day based on their email language patterns, and then responded using another service to send them "happy messages". The latter service didn't make much money.

The romance didn't involve a yacht or a Swiss bank account; it was more the classical 'agape'. Thai and his friends never became tech billionaires - their service was just a small piece of functionality amidst all the other millions of systems and services used globally. The service was however a continued success and provided customers with useful insights and good service as Thai built an organisation around him to continue to develop and support it. The Platform ultimately enabled them to offer a software service globally in a standard, safe, and predictable way that customers could trust, and was used alongside components from much larger organisations to help power organisations all around the globe. Very simply, the Platform had enabled Thai to commercialise his ideas and creation globally with fewer barriers than ever before. Supply and demand took care of the rest.

With their new-found little slice of the global enterprise software market available to them through the Platform, Thai, along with thousands of others around the world, was able to give his family a quality of life that previous generations could never have believed to be possible. Globalisation in the end turned out to be digital, and facilitated not by drones or robotics, but ultimately by data standards.

The Platform itself was an enigma; it wasn’t called ‘Facebook’, ‘AWS’, ‘Netweaver’, 'SalesForce', nor anything you would recognise from ten years before. The creators were also shrouded in mystery. One conspiracy theory was that the platform itself might be an accident of artificial intelligence analysing the databases and code of corporate systems, but that the global giant who created it had chosen to keep its true nature secret so as not to affect the 'coolness' of their consumer and entertainment brands. Others speculated that it was a team of disillusioned employees who could no longer afford to rent their share-tents in their tech-hub capital city, and had created The Platform during an open-ended surfing and coding adventure. After all, such trips had become very popular since the late 2010s after Agile and Couchsurfing had successfully revitalised each other's businesses by converging into a single development philosophy.

The sceptics however believed a different story. With access to aggregate real-time information on every single business transaction in the world, as the creators of the Platform in theory might have had access to, and the opportunities that would provide, who would need to charge much?

Jonathan Gardiner, April 2017