Monitored Information in Clouds - Review and Analysis
Elasticity and pay-as-you-go are only two features of many cloud computing features. However, such and other key characteristics of cloud computing can only be fully enabled if there were no waste of cloud resources. This means, only the exact used amount of cloud resources should be allocated and paid for. To ensure this case, both cloud providers and cloud customers need a monitoring activity for their cloud resources. This monitoring activity should: a. enable measuring states of used, currently needed or available resources, b. enable monitoring (changing of states and recongurations) of these resources and c. ideally to predict future usage and allocations. Cloud Monitoring is not only important for providers, but also for customers. Providers need monitoring to ensure the needed amount of infrastructure, measure their client usage, and billing purposes. On the other hand, customers need monitoring to ensure paying only for the needed amount of resources that meet their needs and not more. Important aspects and indicators to take decisions in monitoring activities are delivered by pre-dened attributes called Monitored Information. These attributes and measurements provide information about states of cloud entities. Depending on their denitions, most Monitored Information have a type and related value. An example of monitored information is the virtual cpu utilization (vcpu-utilization) that could have dierent values in a time window such as 6%, 37%, or 93%. The Monitored Information are indicators of specic observed aspects of entities in clouds in a specic time. These entities could belong to dierent cloud layers such as Physical Layer (PL), Virtual Layer (VL), Logical Application Architecture Layer (LAAL), or Application Business Logic Layer (ABLL). This work is a structured literature review. The purpose of this work is to analyze dierent architectures of cloud monitoring systems and the monitored information used by these systems in dierent cloud layers. This study consolidate the collected monitored information in a single section. The contribution of this work are the first insights about monitored information used in cloud environments.