While many cluster and grid computing frameworks are available, the task of building secure distributed systems or implementing distributed algorithms continue to be a challenging task due to the inherent distributed nature of such systems with multiple failure modes and security issues. In this paper, we present the design and development of remote method delegation (RMD), which is a secure lightweight grid computing platform with load balancing and code migration. RMD is focused on improving the usability issues that plague related industry solutions. The platform is implemented on the JVM (Java Virtual Machine) and supports the Java and Kotlin programming languages, however, the platform should theoretically work with other JVM languages. RMD was designed to simplify the implementation of distributed algorithms by providing a Kotlin DSL (domain specific language) that allows the programmer to define jobs within dedicated code blocks. Users from around the world can donate their own computing resources by hosting their own job server. RMD is secured by placing all untrusted code within a sandbox environment that prevents potentially malicious actions from taking place. To demonstrate the feasibility of the proposed model, a proof of concept implementation has been constructed with real examples demonstrating the usefulness of the proposed solution.

尽管可以使用许多群集和网格计算框架，但是由于此类系统具有多种故障模式和安全性问题，其系统固有的分布式特性，因此构建安全的分布式系统或实现分布式算法的任务仍然是一项艰巨的任务。在本文中，我们介绍了远程方法委托（RMD）的设计和开发，该方法是一种具有负载平衡和代码迁移的安全轻量级网格计算平台。RMD专注于改善困扰相关行业解决方案的可用性问题。该平台是在JVM（Java虚拟机）上实现的，并支持Java和Kotlin编程语言，但是，该平台理论上应与其他JVM语言一起使用。RMD旨在通过提供Kotlin DSL（域特定语言）来简化分布式算法的实现，该语言允许程序员在专用代码块内定义作业。来自世界各地的用户可以通过托管自己的作业服务器来捐赠自己的计算资源。通过将所有不受信任的代码放在沙盒环境中来确保RMD的安全，以防止可能发生的恶意操作。为了证明所提出模型的可行性，已经用实际例子构造了概念证明的实施，证明了所提出解决方案的有效性。通过将所有不受信任的代码放在沙盒环境中来确保RMD的安全，以防止可能发生的恶意操作。为了证明所提出模型的可行性，已经用实际例子构造了概念证明的实施，证明了所提出解决方案的有效性。通过将所有不受信任的代码放在沙盒环境中来确保RMD的安全，以防止可能发生的恶意操作。为了证明所提出模型的可行性，已经用实际例子构造了概念证明的实施，证明了所提出解决方案的有效性。