Towards holistic multi-tenant monitoring for virtual data centers

Abstract

Cloud computing becomes increasingly prevalent for outsourcing IT functions. The basic feature of offering virtual data center slices to customers has been in use for some time now. So far, customers only get the raw resources, with only little insight and control of their resources. But to let customers build reliable services on top of the rented infrastructure, they need adequate monitoring and control capabilities. In the future, we expect operators to offer such functions to their customers. In this paper, we introduce our approach towards offering a holistic monitoring system to data center customers. It offers generic monitoring information propagation and storage covering various types of resources (network, servers, and applications), all kinds of monitoring information, and all tenants. As virtualized data centers are usually large and multitenant, our solution is built with these properties in mind.

Monitoring as-a-service to drive more efficient future system design

“…A monitoring system is comprehensive if it supports different types of resources (physical and virtualised) and is multiple tenants [11]. The latter requires:…”

“…To improve Intrusiveness and Comprehensiveness, [11] proposes an architecture based on agents that monitor directly the flow of information through the same workflow system. They are connected with adapters, which abstract from data of a specific technology.…”

“…Most of the proposed architectures use a subsystem to propagate event announcements [7,10,11,42] or agents for collecting, filtering, and aggregate data [10,11,23,26,31]. Sensu [26]'s use of the publish/subscribe pattern of communication allows for automated registration and de-registration of ephemeral systems, allowing you to dynamically scale the infrastructure up and down.…”

Cloud monitoring: Definitions, issues and future directions

“…A monitoring system is comprehensive if it supports different types of resources, several kinds of monitoring data, and multiple tenants [29]; it is extensible if such support can easily be extended; it is intrusive if its adoption requires significant modification to the Cloud [30]. Most non-Cloudspecific monitoring systems were already designed to provide extensibility and low intrusiveness.…”

Hierarchical Multi-log Cloud-Based Search Engine

“…Further, Lee et al [5] propose that multi-level IDS (Intrusion detection system) is required for protecting cloud computing environments from various security threats and attack patterns. Today major cloud service providers include capacity and resources for their services and QoS guarantees in their SLAs, specified in service level specification (SLS) including SLO (service level objective) [11], [12]. Logging can provide insights into Key Performance Indicators (KPI) about cloud provider SLA and QoS based on analysis.…”