New Professor investigates large-scale data management systems in collaboration with Maersk

- Today, we have so many types of data because of the many applications, software, and devices that generate different types of data. In these years, we’re digitalizing business processes from a lot of different industries, which creates more and more data sources. Therefore, we need flexible systems that can accommodate and incrementally add different types of data with various formats and models without exploding, says Yongluan.

From monolithic software to microservices

One of the implications of the profound digitalization is moving on from monolithic software to a new trend called microservices – a new software architecture where the whole application is divided into small services. This is, among others, used by large enterprises like Amazon who do not use a single large system but multiple services where different teams handle payments, user services, inventory management and so on.   

- The microservice trend creates a different data management problem because you have to develop a different way to match data across multiple microservices. Instead of having a single database that serves all the needs, we need a new data management system to help with this kind of architecture, Yongluan explains.

Event-driven software is the key

In connection to microservices, Event-Driven Architecture (EDA) is a key technology. Conventionally, updates do not automatically trigger new processes. The goal of event-driven software architecture is to enable a quick and agile response to the change of business and in that way seize the business moment.

- In order to create agile business processes, you want an event to trigger the next processes very quickly. For example, when something is sold, the system could immediately order more before the product is out of stock. Therefore, we’re working on developing data management systems to support complex digital systems where all events send updates to the relevant processes immediately or as soon as possible. In principle, every process should be event-driven, says Yongluan. 

He is the principal investigator of a new project called CEEDA (Consistent and Efficient Event-Driven Architecture), in collaboration with Maersk, who will be the end user, Software AG, who will provide of software, and Harry Consulting, who handles the connection to Maersk. The project envisions the next generation of actor-based programming framework with high data consistency and high system efficiency to simplify the development of complex EDA systems.

The foundation for future digital ecosystems

To Yongluan, this field of research is especially interesting because it provides important groundwork for other research areas within Computer Science:

- I believe that building software and data management infrastructure in terms of complex architecture – like event-driven microservices – is the foundation for future digital ecosystems. Without the infrastructure, my colleagues’ work on for example Artificial Intelligence and Machine Learning cannot be implemented or deployed in practice. Viable software and data management infrastructure enables materialization of developments from these areas, says Yongluan.

Besides the CEEDA project, Yongluan will also work on another new and big project at the SDPS-section about privacy and security. He has recently been working on building an efficient and scalable monitor based on a formal model that states who can access data and when, plus recording violation of the rules. The project is called “PAPRiCaS: Programming technology foundations for Accountability, Privacy-by-design & Robustness in Context-aware Systems”.

He will also be General Co-Chair of the EDBT/ICDT International Joint Conference which will be hosted by DIKU in March 2020, with 300 participants from all over the world.