27 Jan SmartWork cloud-based architecture: the backbone of a system improving ageing at work

January 2020: Mr. Charalampos Vassiliou, Project Coordinator from BYTE, provides a first insight concerning the SmartWork System Architecture and overall specifications, and how the system will provide a robust and seamlessly integrated framework.

SmartWork aims to support an age-friendly working and living environment through the utilization of an ICT-enabled framework for the integration and support of unobtrusive and ubiquitous ICT tools and services, taking into consideration novel, scalable and viable business models, based on feedback from large-scale and multicountry pilot applications.

The SmartWork ecosystem (Smart Services Suite) is built on the characteristics of existing reference architectures (such as the ones provided by universAAL) and targets at integrating existing solutions into a cloud-based service-oriented architecture. Advances with respect to existing cloud computing shortcomings are targeted, such as those regarding resource utilisation, capture, store, processing and communication of information to the end-user devices for a reliable and user-acceptable provision of AAL services, adhering to the specific demands for quality of service imposed by the elderly users. Advanced features are aimed at being incorporated to project components and new features developed where appropriate to ensure the efficient handling of captured data and services to be provided. Cloud computing resources such as networking, storage, servers are conveniently requisitioned on demand by AAL services. To this end, an extensive API is being developed for obtaining resources programmatically by components.

Taking into consideration the various SmartWork mobile and desktop applications, the overall web platform components will be deployed on top of a cloud computing infrastructure to provide the basis for delivering a scalable and cost-efficient solution. Each service will be packaged and deployed as an independent component using state-of-the-art container virtualization technology, namely Docker.

A cluster management tool, also known as “orchestrator” will be used on top of the cloud infrastructure to manage containerized application workloads across multiple machines. Google Kubernetes fits this purpose well by providing open source, out-of-the-box functionality for common needs of distributed applications running in production such as naming and discovery, resource monitoring, application health-check, upgrades, replication, load-balancing, auto- scaling and so on. Using Kubernetes the underlying infrastructure (servers / operating systems, network etc) is decoupled from the applications running on top, creating an architecture that is more portable and easier to maintain and scale across clusters of compute and storage nodes.

The SmartWork services will run in a load-balanced stateless tier and access persistent data by communicating with shared external data stores or databases. A polyglot persistence architecture will be targeted where multiple data storage technologies will be used based upon the way data is accessed by the individual applications. MongoDB will be used for providing a scalable and highly-available solution for fast storage and retrieval of data that do not have fixed schema and require support for indexing and querying. PostgreSQL on the other hand, being a relational database that supports database transactions will be used for storing data that have a fixed more or less schema and are more sensitive such as the older worker accounts and the patient profile. Furthermore, in order to avoid extensive disk I/O and bad overall performance that would be the result of always accessing data from databases or file storages, a distributed in-memory caching layer based on Hazelcast and Redis will also be used to store data that is accessed frequently.

The integrated environment described above is a constantly evolving structure as the project matures. This will be deployed in a highly reliable data center supported by BYTE and will be utilized during the semi-controlled and larger field trials to take place at the offices of Cáritas Coimbra and Aarhus Municipality in 2020 and 2021. Feedback from users on the overall functionality, performance and effectiveness of the platform, services and tools will be continuously collected and analysed to identify rooms for improvement and further development.