Applications are evolving in ways that demand geographically distributed resources to co-operate in order to give users better Quality of Service (QoS). There is a plethora of ways to implement such systems, but a Heterogeneous Distributed System such as Jungle Computing System is the one that fits the above scenario the most. Additionally, utilizing the Microservices paradigm results in a dynamic system that can overcome today’s needs for reduced costs, heterogeneity, scalability and fault tolerance. Moreover, Load Balancing is a NP-Hard problem and finding the optimal solution becomes harder when the scale is getting larger. The development and simulation of a Heterogeneous System utilizing Microservices is presented in this article. This work aims at studying how existing load-balancing algorithms can reduce operational costs of such a system and introduces a new algorithm (SaMW) that results in decreasing expenses while keeping relatively low the Mean Response Time.

为了满足用户更好的服务质量（QoS），应用程序在不断发展，它们需要地理上分布的资源进行协作。有很多方法可以实现这样的系统，但是像丛林计算系统这样的异构分布式系统最适合上述情况。此外，利用微服务范式可以创建一个动态系统，该系统可以克服当今对降低成本，异构性，可伸缩性和容错性的需求。此外，负载平衡是一个NP-Hard问题，并且当规模越来越大时，找到最佳解决方案变得更加困难。本文介绍了利用微服务的异构系统的开发和仿真。