Simplifying the Software Infrastructure of Computer and Communication Systems

这项研究方案涉及计算机和通信系统的软件基础设施。这种软件基础设施的例子有操作系统内核、系统库、网络路由器中的转发引擎、命名和会合服务、传输协议等等。这些基础设施组件是所有IT系统的核心，因此对于强烈影响每个个人和整个社会的许多电子系统的性能、安全性和可靠性至关重要。 该研究项目的总体目标是研究软件基础设施的设计和实现的根本更简单的方法。在操作系统和虚拟化域，这主要涉及算法和数据结构，而在网络和分布式系统领域，通信协议是主要的主题。具体地说，这项研究的目的是为低级别系统的设计寻找和揭示简单的和原子的底层模式。有了这些模式，就可以使用基本构建块的简单模块化组合来分解、重新设计和重新构建系统。这将导致更简单的系统，更容易分析、验证和教学。最终，这也将导致更好地绘制整体设计空间的图表。 将研究的具体领域是一种新型操作系统内核，除其他设计思想外，该内核将与一种新型通信基板集成，该基板可以有效地从核心内通信扩展到广域网络。这一建议的独特之处在于综合考虑了整个研究问题的操作系统和网络方面。结果将通过科学出版物和开放软件传播。由此产生的开源软件生态系统将对加拿大工业非常有利。

This research proposal concerns the software infrastructure of computer and communication systems. Examples of such software infrastructure are operating system kernels, system libraries, forwarding engines in network routers, naming and rendezvous services, transport protocols, and many more. These infrastructure components are at the heart of all IT systems and as such, are critically important for the performance, security, and reliability of the many electronic systems that strongly affect every individual and society at large. The overarching goal of this research project is to investigate radically simpler approaches to the design and implementation of software infrastructure. In the operating system and virtualization domain, this mostly concerns algorithms and data structures, while in the area of networks and distributed systems, communication protocols are the primary subject. Specifically, the research is aimed at finding and uncovering simple and atomic underlying patterns for the design of low-level systems. With these patterns, it will be possible to decompose, redesign, and rebuild systems using a simple and modular composition of basic building blocks. This will lead to simpler systems that are easier to analyze, verify, and also teach. Ultimately, it will also result in a better charting of the overall design space. The specific areas that will be investigated are a novel operating system kernel that, among other design ideas, will be integrated with a novel communication substrate that can efficiently scale from intra-core communication to wide-area networking. The unique proposition of this proposal is the integrated consideration of both operating systems and networking aspects of the overall research problem. The results will be disseminated through scientific publications and open software. The resulting open-source software ecosystem will be very beneficial to Canadian industry.