Cloud Monitoring

Cloud Monitoring

Peer Hasselmeyer (NEC Laboratories Europe, Germany), Gregory Katsaros (National Technical University of Athens, Greece), Bastian Koller (High Performance Computing Centre Stuttgart, Germany) and Philipp Wieder (Gesellschaft fuer wissenschaftliche Datenverarbeitung mbH Goettingen, Germany)
DOI: 10.4018/978-1-4666-1631-8.ch006


The management of the entire service landscape comprising a Cloud environment is a complex and challenging venture. There, one task of utmost importance, is the generation and processing of information about the state, health, and performance of the various services and IT components, something which is generally referred to as monitoring. Such information is the foundation for proper assessment and management of the whole Cloud. This chapter pursues two objectives: first, to provide an overview of monitoring in Cloud environments and, second, to propose a solution for interoperable and vendor-independent Cloud monitoring. Along the way, the authors motivate the necessity of monitoring at the different levels of Cloud infrastructures, introduce selected state-of-the-art, and extract requirements for Cloud monitoring. Based on these requirements, the following sections depict a Cloud monitoring solution and describe current developments towards interoperable, open, and extensible Cloud monitoring frameworks.
Chapter Preview


Network and systems management has been an important topic for data center and communication service providers for decades. Appropriate management does not only ensure continued service availability and superior quality levels; it can also be a competitive advantage ensuring a quick response to customer requests and reduced costs for operations.

Management is typically separated into the two functional areas of monitoring and control. Monitoring is concerned with taking measurements of various parameters of the infrastructure, while control is about adjusting the infrastructure's configuration. Monitoring provides the information needed to reach appropriate control decisions. Monitoring is therefore essential to properly control the infrastructure.

Monitoring is typically done by adding instrumentation to the hardware and software infrastructure. The instrumentation provides measurements that relate to key performance indicators (KPIs) of the operator's systems. Proper management requires information from the complete set of components involved in service delivery, usually including hardware such as servers, network, and storage, and software such as operating system, middleware, and applications. Measurements include diverse metrics such as CPU utilization, network bandwidth consumption, transaction throughput, and the number of active users. From the collected information, the current state of the infrastructure can be assessed and, ideally, future requirements and utilization can be predicted in order to proactively adjust the available resources and their configurations to meet future demand.

Cloud computing has recently received a lot of attention as a new paradigm for the use of computational power, electronic services, and storage resources. In such scenarios, an IaaS provider has a contract with a service provider while the service provider in turn has a contract with an end-user, who can be private or an enterprise (Figure 1). The services of the service provider are supported by the infrastructure from the IaaS provider. The quality of service offered by the service provider depends on the quality of service that the IaaS provider supports. The service provider therefore relies on the infrastructure provider, and both parties need adequate monitoring to ensure the required service levels.

Figure 1.

A typical service provisioning value chain


As initially stated, monitoring is the basis for resource control which in turn is required to ensure proper operation of the infrastructure in order to ensure adequate service provisioning. The need for monitoring is not shrinking in the Cloud. In fact, it becomes even more important as performance parameters become less predictable with the use of shared, virtualized resources. Ideally, the Cloud infrastructure provides the same information as local installations do, combined with the ease of use of Clouds.

Although the acquisition, use, and decommission of virtual machines are well addressed by commercial providers and open-source software packages, monitoring of the rented Cloud infrastructure is not as advanced as their deployment. Some basic capabilities are offered, but they are usually much more limited than what is commonly available on locally operated hardware. In particular, the capabilities are mainly geared towards the need of the Cloud operator, not the Cloud users.

The focus of IaaS providers is on keeping their systems running at an operating point that maximizes their profits. They are therefore only interested in low-level information about the health of their physical infrastructure. In addition, they need information about the resource usage of customers' virtual resources, mainly for billing purposes but also to see how the Virtual Machines behave, and to evaluate their performance and quality on their system. IaaS customers, on the other hand, need low-level information about their virtual infrastructure as well as all the software that is running on top of it. It is obvious that in most cases there is little overlap of the sets of information needed by providers and customers. Only some low-level measurements of virtual infrastructure parameters are shared. Only these parts are made available by Cloud providers to their customers. This is clearly not enough to get a complete overview of the state of applications and services deployed in the Cloud. Any additional information needed by customers must be dealt with by customers themselves.

Complete Chapter List

Search this Book: