mirror of
https://gitlab.com/netravnen/NetworkLabNotes.git
synced 2024-12-23 20:57:53 +00:00
52 lines
2.2 KiB
TeX
52 lines
2.2 KiB
TeX
\chapter[QoS]{Quality-of-Service}
|
|
|
|
\gls{qos} is used to guarantie a minimum of service level to select applications. Often this encompasses \gls{voip} applications and \gls{av} applications being allocated the highest priority. It\tsq{s} not uncommon to allocate \gls{nmt} to the high priority queue, too.
|
|
|
|
Different mechanisms of handling access to network ressorces is used.
|
|
\begin{itemize}
|
|
\item \itemhead{\gls{nac}}
|
|
\begin{enumerate}
|
|
\item Which applications has access to what level of network ressources.
|
|
\end{enumerate}
|
|
\item \itemhead{Traffic Control}
|
|
\begin{enumerate}
|
|
\item Scheduling of traffic,
|
|
\item classifiying traffic,
|
|
\item marking packets based upon priority,
|
|
\item marking packets based upon shaping traffic.
|
|
\end{enumerate}
|
|
\end{itemize}
|
|
|
|
\section{Concepts}
|
|
|
|
\gls{qos}\tsq{s} goal is a differentiated prioritazion of packets parsing thorugh the network based upon the following concepts:
|
|
\begin{enumerate}
|
|
\item Bandwidth,
|
|
\item latency,
|
|
\item jitter\footnote{Latency Variation}
|
|
\item realiability\footnote{Pct. of packets discarded by any a router}.
|
|
\end{enumerate}
|
|
|
|
\section[Congestion Mgmt]{Congestion Management}
|
|
|
|
There are different ways to do congestion management. Which is in it\tsq{s} essence sorting of packets when a link reaches full capacity usage in the outgoing direction.
|
|
|
|
\begin{enumerate}
|
|
\item \gls{fifo}: Classic store-and-forward. Oftentimes the default algorithm.
|
|
\item \gls{pq}: Made to give stict priority to important traffic at each point \gls{pq} in the network where \gls{pq} is used.
|
|
\item \gls{cq}:
|
|
\item \gls{wfq}: Traffic is diveded into flow based upon charactaristics \begin{mylist} \item \gls{dst} address, \item \gls{src} address, \item protocol, \item port number, \item socket. \end{mylist}. Flows is then allocated a part of the bandwidth relative to the number of ongoing conversations/flows.
|
|
\begin{enumerate}
|
|
\item Flow-based \gls{wfq},
|
|
\item \gls{cbwfq},
|
|
\item \gls{dwfq}.
|
|
\end{enumerate}
|
|
\item distributed class-based \gls{wfq}
|
|
\item \gls{ip} \gls{rtp} priority
|
|
\item \gls{llq}
|
|
\begin{enumerate}
|
|
\item \gls{llq}
|
|
\item Distributed \gls{llq}:
|
|
\end{enumerate}
|
|
\end{enumerate}
|