Sanjay Sarma is co-founder and research director at Auto-ID Labs, Massachusetts Institute of Technology (MIT). He is also a keynote speaker at NEXUS 17. 

Sanjay Sarma

	Regulations mandating serialization are forcing companies to consider how information technology (IT) can better support business operations and help people work smarter. Concerns over data security, management, and storage are driving many companies to consider how outsourced IT infrastructures can support them in navigating this new digital world.

	The upcoming serialization requirements in the European Union and the United States have presented pharmaceutical companies with the challenge of balancing data integrity with performance when designing the appropriate information architecture. But what makes a good architecture and what can the industry learn from the architectures that have changed the world for the better, but are often taken for granted?

Understanding architectures: a historic view

	In today’s modern society, a number of architectures add order and structure to the world. The highway/byway/rural road system is an example of a dominant architecture for road design in the United States. This system includes bypasses around cities, entry and exit ramps, and cloverleaf exchanges. In developing countries, a much simpler architecture is used consisting of ad hoc roads between settlements.

	Many companies have seen major architectural changes within their sectors in the past 10 years, which can impact both day-to-day activities as well as wider business operations. The pharmaceutical industry is currently faced with the need to introduce architectural change, as it prepares for the European Falsified Medicines Directive (FMD).

	From experience, it has become clear that industries succeed when there is a simple, dominant architecture. Although such architectures may at first seem counter-intuitive, they can unlock extraordinary value and capability.

	One example of an architectural change that has transformed our lives is the decision to use alternating current (AC) over direct current (DC). DC is the most obvious choice, sending an electrical current direct to the device that needs to be charged. However, the counter-intuitive architectural choice to use AC has made the modern grid possible, which is now the dominant architecture most commonly used today.

	Similarly, the basic idea of cellular technology is counter-intuitive. For example, two people in the same room can call each other, but their phones do not talk to each other directly. The call goes to a tower and then through the cloud, back to your home, back to the same tower, and back to the other phone - it takes a long round trip. However, this is the scalable alternative to the handheld transceiver (or walkie-talkie) developed during the Second World War.

	The World Wide Web is another fundamental, counter-intuitive architectural choice. In the 1980s, the Internet existed. However, Tim Berners-Lee set out to create a sub-set of it, a simple architecture consisting of a page, a hyperlink, and a protocol. While some people believed this was too simplistic, he had the vision to say that this counter-intuitive approach would unleash more possibilities and it is now the dominant architecture used today.

	The pharma sector is currently undergoing the process of introducing new systems and processes for serialization. A scalable solution that balances integrity with performance is likely to prevail and become a dominant architecture.

Software architectures-balancing integrity with performance

	Companies offering networks for the storing and sharing of vast amounts of serialization data are challenged with creating a shareable world that is also scalable. For pharmaceutical companies, serialization will require a paradigm shift in IT architecture to deal with the combination of the vast amount of data stored within network databases and the operational processes associated with the upcoming requirements.

	Traditional enterprise software architectures have prioritized integrity, security, and reliability. They were based on the “integrity first” principle. As they were designed for third-party use, they had to be error- and idiot-proof. The operator was not the creator of the software, so the creator had to take a comprehensive approach to make sure the user could not cause any issues.

	The fundamental change that has happened in software is the advent of the cloud. For example, a typical user will never use Google or Facebook’s software itself or write code within it; he or she will only interface with it. As such, Facebook’s primary concern is performance. Because if there are delays, people will stop using it. Similarly, Amazon cares about round trip time, measuring it in milliseconds to ensure the best possible user experience.

	However, when business-critical information is shared over the cloud, the architecture must exist in both the cloud and enterprise worlds; in the case of serialization, it must offer both integrity and performance. Vast amounts of new data will be created as companies begin to implement their serialization solutions, this will need to be stored, managed and accessed in a secure manner. For example, Facebook may lose a message from a profile, but it is not acceptable for a pharmaceutical company to lose a serial number. If the cloud software is slow, then sharing serial numbers could take seconds, and potentially cause a shipment to be missed as a result.

	When pharmaceutical companies consider how they will store, manage and exchange serialization data, the chosen IT architecture must offer integrity, performance, scalability, the ability to customize, security, and reliability.

	A native cloud architecture is decoupled from physical resource, enabling applications to be developed and deployed across different cloud platforms, helping them to run at scale. This can help companies balance the competing priorities between integrity and performance. It is likely that traditional on-premise solutions will struggle to keep pace with the rapid adoption of compliance requirements and the vast amount of data involved. In contrast, a centralized cloud network can provide instant access to transactional data at the click of a button, simplifying complex integrations with internal and external third-party systems and connecting companies throughout the whole pharmaceutical supply chain in a single network.

	This can bring about wider cost efficiencies as less resource is required when compared to managing traditional databases and full scalability can be achieved through leveraging the benefits of the cloud. Because of this and due to its ability to combine integrity with performance, we can expect the native cloud to become the dominant architecture within the pharmaceutical industry.

	In an industry as highly regulated and safety critical as the pharmaceutical sector, the integrity of the real world must meet the volume and performance of the virtual world. Cloud networks for serialization must offer security, together with speed and scalability especially with the introduction of new drug traceability requirements across the globe. 

 is co-founder and research director at Auto-ID Labs, Massachusetts Institute of Technology (MIT). He is also a keynote speaker at 

Articles

The author shares his opinion on the challenges presented by the Internet of Things and what companies need to consider when choosing suitable architectures to manage serialization data.