Logo des Repositoriums

Auflistung nach Schlagwort "Deployment"

1 - 4 von 4
  • Zeitschriftenartikel
    A Case for a New IT Ecosystem: On-The-Fly Computing
    (Business & Information Systems Engineering: Vol. 62, No. 6, 2020) Karl, Holger; Kundisch, Dennis; Meyer auf der Heide, Friedhelm; Wehrheim, Heike
    The complexity of development and deployment in today’s IT world is enormous. Despite the existence of so many pre-fabricated components, frameworks, cloud providers, etc., building IT systems still remains a major challenge and most likely overtaxes even a single ambitious developer. This results in spreading such development and deployment tasks over different team members with their own specialization. Nevertheless, not even highly competent IT personnel can easily succeed in developing and deploying a nontrivial application that comprises a multitude of different components running on different platforms (from frontend to backend). Current industry trends such as DevOps strive to keep development and deployment tasks tightly integrated. This, however, only partially addresses the underlying complexity of either of these two tasks. But would it not be desirable to simplify these tasks in the first place, enabling one person – maybe even a non-expert – to deal with all of them? Today’s approaches to the development and deployment of complex IT applications are not up to this challenge. “On-The-Fly Computing” offers an approach to tackle this challenge by providing complex IT services through largely automated configuration and execution. The configuration of such services is based on simple, flexibly combinable services that are provided by different software providers and traded in a market. This constitutes a highly relevant challenge for research in many branches of computer science, information systems, business administration, and economics. In this research note, it is analyzed which pieces of this new “On-The-Fly Computing” ecosystem already exist and where additional, often significant research efforts are necessary.
  • Konferenzbeitrag
    AI in the Wild: Challenges of Remote Deployments
    (INFORMATIK 2023 - Designing Futures: Zukünfte gestalten, 2023) Dede, Jens; Wewetzer, David; Förster, Anna
    The effect of humanity on the earth becomes more and more apparent. Besides the publicly discussed climate change and overpopulation, also the number of conflicts with wildlife increases. The technological progress of the past years helped to understand these challenges better. Monitoring solutions, known to the public as the Internet of Things (IoT), increase the amount of collected data, whereas artificial intelligence (AI) supports analyzing and gathering a deeper understanding. Most projects in the area of wildlife try to achieve a more sustainable usage of natural resources and a better coexistence with our environment. The mAInZaun project focuses on the conflict between wolves and livestock. It aims to introduce these new technologies into grazing management and foster non-lethally coexistence between livestock and predators. Artificial intelligence (AI) analyzes images and videos of the areas surrounding the pasture. The algorithms detect possible attackers or predators, such as wolves, stray dogs, bobcats, etc. In the second step, these animals are scared away using adaptive technologies. These can be sound, ultrasound, scent, light, etc. These systems are usually operated in remote environments, raising challenges like hardware design, power requirements, and maintenance. This paper will discuss these challenges and how we address them in the mAInZaun project.
  • Konferenzbeitrag
    Local energy demand of functionality distributions
    (INFORMATIK 2024, 2024) Schmalriede, Florian; Winter, Andreas
    Heterogeneous and distributed systems allow many different functionality distributions across devices involved. Various distributions lead to different local energy demands, so that runtime requirements are met or not. As part of deployment planning, it is necessary to find alternatives that meet runtime requirements. Therefore, alternatives must be analyzed and compared regarding energy demand. In this paper, the influences of functionality distributions on local energy demand are investigated using a smart health example. Automated energy measurements across different functional distributions and input data are collected and compared. It is shown that functional distributions and input data influence local energy demand.
  • Konferenzbeitrag
    Model-Driven Allocation Engineering – Abridged Version
    (Software Engineering 2017, 2017) Pohlmann, Uwe; Hüwe, Marcus