Slicing as a distributed systems primitive

Abstract

Large-scale distributed systems appear as the major in- frastructures for supporting planet-scale services. These systems call for appropriate management mechanisms and protocols. Slicing is an example of an autonomous, fully decentral- ized protocol suitable for large-scale environments. It aims at organizing the system into groups of nodes, called slices, according to an application-specific criteria where the size of each slice is relative to the size of the full system. This al- lows assigning a certain fraction of nodes to different task, according to their capabilities. Although useful, current slicing techniques lack some features of considerable practical importance. This pa- per proposes a slicing protocol, that builds on existing so- lutions, and addresses some of their frailties. We present novel solutions to deal with non-uniform slices and to per- form online and dynamic slices schema reconfiguration. Moreover, we describe how to provision a slice-local Peer Sampling Service for upper protocol layers and how to en- hance slicing protocols with the capability of slicing over more than one attribute. Slicing is presented as a complete, dependable and inte- grated distributed systems primitive for large-scale systems.