Commit 540c16c2 authored by Simon Eismann's avatar Simon Eismann
Browse files

userguide now contains all changes to the mm aswell as the changes made by Simon Spinner

parent 6c473b38
No preview for this file type
......@@ -647,7 +647,7 @@ Figure~\ref{fig:m-example_deployment} shows an exemplary deployment of the runni
\label{fig:m-example_deployment}
\end{figure}
There is a datacenter that consists of two \model{ComputingInfrastructure}s, namely an \instance{ApplicationServer} and a \instance{DatabaseServer}. The servers' CPUs are modeled as active resources. The CPU of the \instance{ApplicationServer} consists of two processing units at a processing rate of 2.66~GHz. The \instance{DatabaseServer} CPU has eight processing units at a processing rate of 2.8~GHz.
There is a datacenter that consists of two \model{ComputeNodes}s, namely an \instance{ApplicationServer} and a \instance{DatabaseServer}. The servers' CPUs are modeled as active resources. The CPU of the \instance{ApplicationServer} consists of two processing units at a processing rate of 2.66~GHz. The \instance{DatabaseServer} CPU has eight processing units at a processing rate of 2.8~GHz.
The \instance{WebShop} instance is deployed on the \instance{ApplicationServer} node, a \instance{SQLDB} instance is deployed on to the \instance{DatabaseServer} node.
%##################################################
......
......@@ -20,15 +20,15 @@ and the performance influences of the different resource layers.
\Cref{fig:mm-resourcelandscape-diagram} depicts an overview of the structure of the resource landscape meta-model as a UML class diagram.
The root entity comprising all other model elements is the \model{DistributedDataCenter}, which consists of one or more \model{DataCenters}.
\model{DataCenters} contain \model{HardwareInfrastructures} which are either one of the three hardware infrastructure types \model{ComputingInfrastructure}, \model{NetworkInfrastructure}, and \model{StorageInfrastructure}, or \model{CompositeHardwareInfrastructure}.
A \model{CompositeHardwareInfrastructure} is a structuring element to group further \model{HardwareInfrastructures}.
\model{DataCenters} contain \model{ComputeNodes}, \model{StorageNodes}, and \model{CompositeHardwareInfrastructure}.
A \model{CompositeHardwareInfrastructure} is a structuring element to group further \model{Compute-} and \model{StorageNodes}.
For example, it can be used to combine servers to a cluster or to group them in a server rack.
Current architecture-level performance models usually abstract from these details and do not provide constructs to model the resource hierarchy and containment relationships explicitly.
However, for resource management at run-time, the description of the resource landscape and hierarchy of resources is crucial to improve reasoning about suitable adaptation operations, e.g., to decide if a VM can be migrated and where it should be migrated to.
Here, the novel aspect of our meta-model is that it allows to model the distribution of resources within and across data centers as well as their individual configuration.
When designing the resource landscape meta-model, we aimed at a generic approach to cover all types of infrastructure with the focus on \model{ComputingInfrastructure}.
More details on modeling storage and network infrastructures (i.e., the \model{StorageInfrastructure} and \model{NetworkInfrastructure} entities in the meta-model) can be found in the work of \cite{noorshams2013a,noorshams2013b,noorshams2013c} and \cite{RyKoZs2013-ThroughputPrediction,RyZsKo2013-DNI-meta-model}, respectively.
When designing the resource landscape meta-model, we aimed at a generic approach to cover all types of infrastructure with the focus on \model{ComputeNode}.
More details on modeling storage and network infrastructures (i.e., the \model{StorageNodes} entities in the meta-model) can be found in the work of \cite{noorshams2013a,noorshams2013b,noorshams2013c}.
\subsubsection{Containers and Containment Relationships}
\label{ssec:mm-containers}
......@@ -37,8 +37,8 @@ A common reappearing pattern in modern distributed IT service infrastructures is
This leads to a tree of nested system entities that may change during runtime because of virtual machine migration, hardware or software failures, etc.
% Because of this flexibility one single execution environment consisting of similar, reoccurring elements can have multiple different configuration states at run-time.
The central element of our resource landscape meta-model to model these nested layers of resources is the abstract entity \model{Container}, depicted in \Cref{fig:mm-resourcelandscape-diagram}.
We distinguish between two major concrete container entities: the \model{ComputingInfrastructure} and the \model{RuntimeEnvironment}.
The \model{ComputingInfrastructure} forms the root element in our hierarchy of containers and corresponds to a physical machine within a data center.
We distinguish between two major concrete container entities: the \model{ComputeNode} and the \model{RuntimeEnvironment}.
The \model{ComputeNode} forms the root element in our hierarchy of containers and corresponds to a physical machine within a data center.
This entity cannot be contained in another container, but it may have nested containers (\model{RuntimeEnvironments}).
The \model{RuntimeEnvironment} is the second type of container.
It can contain further \model{RuntimeEnvironments}.
......@@ -88,7 +88,7 @@ Also, we can assume that model instances can be built automatically or with tool
\label{ssec:mm-resource_configuration_specification}
Each \model{Container} has its own specific resource configurations that describe the container's influence on the system performance.
In our meta-model, we distinguish between three different types of configuration specifications: \model{ActiveResourceSpecification}, \model{PassiveResourceSpecification}, and \model{CustomConfigurationSpecification}, depicted in \Cref{fig:mm-resource_configuration_diagram}.
In our meta-model, we distinguish between three different types of configuration specifications: \model{ProcessingResourceSpecification}, \model{LinkingResourceSpecification}, \model{PassiveResourceSpecification}, and \model{CustomConfigurationSpecification}, depicted in \Cref{fig:mm-resource_configuration_diagram}.
\begin{figure}[ht]
\centering
......@@ -208,7 +208,7 @@ showing the hierarchy of the different resources as well as their configuration
\label{fig:resource_landscape_example}
\end{figure}
The root element is the local \model{DataCenter} in our computer science department at KIT, which contains a \model{CompositeHardwareInfrastructure} (a cluster environment called \instance{AcamarCluster}), and a separate \model{ComputingInfrastructure}, the \instance{DatabaseServer}.
The root element is the local \model{DataCenter} in our computer science department at KIT, which contains a \model{CompositeHardwareInfrastructure} (a cluster environment called \instance{AcamarCluster}), and a separate \model{ComputeNode}, the \instance{DatabaseServer}.
% Other infrastructure of the data center that is not part of our experiments is omitted in the figure.
The cluster in this example consists of five \instance{ComputeNodes}, connected by a 1~Gbit Ethernet LAN.
Each compute node runs XenServer~5.5 as a hypervisor.
......@@ -219,11 +219,11 @@ To ease the resource configuration specification of the other containers, we use
The resource configuration specification templates for the different container types are stored in the \instance{ExperimentEnvironmentContainerSpecs} container repository.
The \instance{Compute\-NodeTemplate} specifies the hardware resource configuration of the cluster compute nodes.
A compute node has two \model{ActiveResourceSpecifications} modeling its two CPUs.
A compute node has two \model{ProcessingResourceSpecifications} modeling its two CPUs.
Each has four cores with 2.66~GHz and \model{PROCESSOR\_SHARING} as scheduling policy.
The \instance{XenServer5.5Template} is a template for a \model{RuntimeEnvironment} of class \model{HYPERVISOR}.
It refers to a \model{CustomConfigurationSpecification}, which refers to a \model{CustomResourceConfigurationModel} for the XenServer~5.5 hypervisor.
Further details of this custom model have been presented in \cite{HuQuHaKo2011-CLOSER-ModelVirtOverhead}.
Finally, the \instance{VMTemplate} specifies the configuration of the VMs hosted by the XenServer.
This \model{RuntimeEnvironment} is of class \model{OS\_VM} and has only one \model{ActiveResourceSpecification} for its VCPU.
This \model{RuntimeEnvironment} is of class \model{OS\_VM} and has only one \model{ProcessingResourceSpecification} for its VCPU.
It has two cores with 2.66~GHz and \model{PROCESSOR\_SHARING} as scheduling policy.
\ No newline at end of file
This source diff could not be displayed because it is too large. You can view the blob instead.
\relax
\providecommand\hyper@newdestlabel[2]{}
\@setckpt{chapters/titlepage}{
\setcounter{page}{2}
\setcounter{equation}{0}
\setcounter{enumi}{0}
\setcounter{enumii}{0}
\setcounter{enumiii}{0}
\setcounter{enumiv}{0}
\setcounter{footnote}{0}
\setcounter{mpfootnote}{0}
\setcounter{part}{0}
\setcounter{chapter}{0}
\setcounter{section}{0}
\setcounter{subsection}{0}
\setcounter{subsubsection}{0}
\setcounter{paragraph}{0}
\setcounter{subparagraph}{0}
\setcounter{figure}{0}
\setcounter{table}{0}
\setcounter{parentequation}{0}
\setcounter{Item}{0}
\setcounter{Hfootnote}{0}
\setcounter{bookmark@seq@number}{0}
\setcounter{ContinuedFloat}{0}
\setcounter{subfigure}{0}
\setcounter{subtable}{0}
\setcounter{example}{0}
\setcounter{lstnumber}{1}
\setcounter{su@anzahl}{0}
\setcounter{LT@tables}{0}
\setcounter{LT@chunks}{0}
\setcounter{ecnt}{0}
\setcounter{section@level}{0}
\setcounter{lstlisting}{0}
}
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment