Abstract
Modern society is increasingly dependent on reliable performance of distributed systems. In this paper, we provide a precise definition of performance using the concept of quality attenuation; discuss its properties, measurement and decomposition; identify sources of such attenuation; outline methods of managing performance hazards automatically using the capabilities of the Recursive InterNetworking Architecture (RINA); demonstrate procedures for aggregating both application demands and network performance to achieve scalability; discuss dealing with bursty and time-critical traffic; propose metrics to assess the effectiveness of a performance management system; and outline an architecture for performance management.
Highlights
Online services, implemented by distributed computing systems, have displaced traditional ones in many areas: shopping; entertainment; social interaction; banking; and public services, to name but a few
This is enabled by developments such as software-defined networking (SDN) [3], which increase the dynamic configurability of networks
Background information on Recursive InterNetworking Architecture (RINA) and detail on how it interacts with performance management; More detail on the decomposition of ∆Q; A description of the process of aggregating QTAs; A new section on dealing with bursty traffic; Expanded conclusions and discussion of directions for future work; An appendix showing how ∆Q can be calculated from application behaviour
Summary
Online services, implemented by distributed computing systems, have displaced traditional ones in many areas: shopping; entertainment; social interaction; banking; and public services, to name but a few. There are growing societal concerns: firstly that such online services are not sufficiently reliable [1]; and secondly that implementing this trend using a traditional client–server model places too much power in the hands of a small number of service providers This is fuelling interest in service architectures that are less dependent on centralised control and management, for example distributed ledgers such as blockchains [2]. RINA provides nested scopes called DIFs within which management of performance is feasible, and control loops can be short enough to address congestion in a timely fashion [6] A new section on dealing with bursty traffic; Expanded conclusions and discussion of directions for future work; An appendix showing how ∆Q can be calculated from application behaviour
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have