Refine
Document Type
- Conference Proceeding (8)
- Master's Thesis (1)
Has Fulltext
- yes (9)
Is part of the Bibliography
- no (9) (remove)
Keywords
- Mikroservice (9) (remove)
Institute
Even for the more traditional insurance industry, the Microservices Architecture (MSA) style plays an increasingly important role in provisioning insurance services. However, insurance businesses must operate legacy applications, enterprise software, and service-based applications in parallel for a more extended transition period. The ultimate goal of our ongoing research is to design a microservice reference architecture in cooperation with our industry partners from the insurance domain that provides an approach for the integration of applications from different architecture paradigms. In Germany, individual insurance services are classified as part of the critical infrastructure. Therefore, German insurance companies must comply with the Federal Office for Information Security requirements, which the Federal Supervisory Authority enforces. Additionally, insurance companies must comply with relevant laws, regulations, and standards as part of the business’s compliance requirements. Note: Since Germany is seen as relatively ’tough’ with respect to privacy and security demands, fullfilling those demands might well be suitable (if not even ’over-achieving’) for insurances in other countries as well. The question raises thus, of how insurance services can be secured in an application landscape shaped by the MSA style to comply with the architectural and security requirements depicted above. This article highlights the specific regulations, laws, and standards the insurance industry must comply with. We present initial architectural patterns to address authentication and authorization in an MSA tailored to the requirements of our insurance industry partners.
Cloud computing has become well established in private and public sector projects over the past few years, opening ever new opportunities for research and development, but also for education. One of these opportunities presents itself in the form of dynamically deployable, virtual lab environments, granting educational institutions increased flexibility with the allocation of their computing resources. These fully sandboxed labs provide students with their own, internal network and full access to all machines within, granting them the flexibility necessary to gather hands-on experience with building heterogeneous microservice architectures. The eduDScloud provides a private cloud infrastructure to which labs like the microservice lab outlined in this paper can be flexibly deployed at a moment’s notice.
In the context of modern mobility, topics such as smart-cities, Car2Car-Communication, extensive vehicle sensor-data, e-mobility and charging point management systems have to be considered. These topics of modern mobility often have in common that they are characterized by complex and extensive data situations. Vehicle position data, sensor data or vehicle communication data must be preprocessed, aggregated and analyzed. In many cases, the data is interdependent. For example, the vehicle position data of electric vehicles and surrounding charging points have a dependence on one another and characterize a competition situation between the vehicles. In the case of Car2Car-Communication, the positions of the vehicles must also be viewed in relation to each other. The data are dependent on each other and will influence the ability to establish a communication. This dependency can provoke very complex and large data situations, which can no longer be treated efficiently. With this work, a model is presented in order to be able to map such typical data situations with a strong dependency of the data among each other. Microservices can help reduce complexity.
Microservices build a deeply distributed system. Although this offers significant flexibility for development teams and helps to find solutions for scalability or security questions, it also intensifies the drawbacks of a distributed system. This article offers a decision framework, which helps to increase the resiliency of microservices. A metamodel is used to represent services, resiliency patterns, and quality attributes. Furthermore, the general idea for a suggestion procedure is outlined.
Portable-micro-Combined-Heat-and-Power-units are a gateway technology bridging conventional vehicles and Battery Electric Vehicles (BEV). Being a new technology, new software has to be created that can be easily adapted to changing requirements. We propose and evaluate three different architectures based on three architectural paradigms. Using a scenario-based evaluation, we conclude that a Service-Oriented Architecture (SOA) using microservices provides a higher quality solution than a layered or Event-Driven Complex-Event-Processing (ED-CEP) approach. Future work will include implementation and simulation-driven evaluation.
The usage of microservices promises a lot of benefits concerning scalability and maintainability, rewriting large monoliths is however not always possible. Especially in scientific projects, pure microservice architectures are therefore not feasible in every project. We propose the utilization of microservice principles for the construction of microsimulations for urban transport. We present a prototypical architecture for the connection of MATSim and AnyLogic, two widely used simulation tools in the context of urban transport simulation. The proposed system combines the two tools into a singular tool supporting civil engineers in decision making on innovative urban transport concepts.
In microservice architectures, data is often hold redundantly to create an overall resilient system. Although the synchronization of this data proposes a significant challenge, not much research has been done on this topic yet. This paper shows four general approaches for assuring consistency among services and demonstrates how to identify the best solution for a given architecture. For this, a microservice architecture, which implements the functionality of a mainframe-based legacy system from the insurance industry, serves as an example.
Microservices is an architectural style for complex application systems, promising some crucial benefits, e.g. better maintainability, flexible scalability, and fault tolerance. For this reason microservices has attracted attention in the software development departments of different industry sectors, such as ecommerce and streaming services. On the other hand, businesses have to face great challenges, which hamper the adoption of the architectural style. For instance, data are often persisted redundantly to provide fault tolerance. But the synchronization of those data for the sake of consistency is a major challenge. Our paper presents a case study from the insurance industry which focusses consistency issues when migrating a monolithic core application towards microservices. Based on the Domain Driven Design (DDD) methodology, we derive bounded contexts and a set of microservices assigned to these contexts. We discuss four different approaches to ensure consistency and propose a best practice to identify the most appropriate approach for a given scenario. Design and implementation details and compliance issues are presented as well.
Zusammen mit der Microservice-Bewegung werden immer häufiger synchrone Request-Response-Schnittstellen nach dem REST-Paradigma entwickelt, um Service-Landschaften zu integrieren. Die Einfachheit des Paradigmas verleitet viele Organisationen, nahezu die komplette Interprozesskommunikation ihres Ökosystems über diese Art von Schnittstelle abzuwickeln – nicht ohne Konsequenzen.
Diese Arbeit entwickelt Ansätze, wie die Integrationsprobleme, die bei übermäßiger Verwendung von REST entstehen, mithilfe von Event-Driven Architecture gelöst werden können, ohne den Status quo dieser Organisationen außer Acht zu lassen. Dafür werden der gegenwärtige Zustand der Integrationsmuster und eingesetzten Infrastruktur von Event-Driven Architecture kritisiert und Kriterien erarbeitet, die pragmatische und zugängliche Integrationsansätze erfüllen müssen. Um die Einführungskosten gering zu halten, wird eine Middleware entwickelt, die in bestehende REST-Schnittstellen eingesetzt werden kann und auf Basis der API-Aufrufe Events generiert. Darauf aufbauend werden vier Integrationsmuster entwickelt, die eine schrittweise Transformation zu Event-Driven Microservices ermöglichen. Um die Zugänglichkeit der Eventing-Infrastruktur zu erhöhen, wird außerdem wird die Standardisierung der Event-Struktur durch die CloudEvents-Spezifikation vorgeschlagen. Um die Zugänglichkeit weiter zu erhöhen, erfolgt die Kommunikation der Services nicht direkt mit dem Event-Broker, sondern über Proxies, die die Events per HTTP annehmen oder ausspielen. Um die Transparenz über den Datenfluss im System zu wahren, werden alle Produzenten und Konsumenten werden mitsamt ihrer Events durch den Beschreibungsstandard AsyncAPI dokumentiert.
Nach einer Validierung dieser Ansätze mithilfe eines Prototyps kommt diese Arbeit zu der Erkenntnis, dass der Einsatz der entwickelten Middleware für alle Organisationen sinnvoll ist, die bereits viele REST-Schnittstellen im Einsatz haben. Die Standardisierung der Event-Struktur und des Event-Protokolls mittels CloudEvents und HTTP-Proxies sowie die Dokumentation durch AsyncAPI empfiehlt sich auch unabhängig des Status quo für alle Organisationen, die Event-Driven Microservices entwickeln möchten.