diff options
Diffstat (limited to 'script/main.tex')
| -rw-r--r-- | script/main.tex | 34 |
1 files changed, 29 insertions, 5 deletions
diff --git a/script/main.tex b/script/main.tex index a524e96..8acb153 100644 --- a/script/main.tex +++ b/script/main.tex @@ -1,5 +1,6 @@ \documentclass[12pt, a4paper]{article} \usepackage{palatino, enumitem, parskip, xspace} +\usepackage[top=1.5cm, bottom=1.5cm, left=2cm, right=2cm]{geometry} \usepackage[dvipsnames]{xcolor} \newcommand{\eg}{\emph{e.g.,}\xspace} \newcommand{\todo}[1]{\textcolor{Blue}{\textbf{TODO(#1)}}} @@ -18,8 +19,6 @@ As you know and as you will likely see in the upcoming presentations today, datacenters are important. But, I would like to shortly mention this myself. - - A single GPU is already very complex. Within a Google Datacenter, there are hundreds of server racks, with tens of such GPUs. This begs the question: How are we going to manage this large of a datacenter, that has so many \emph{layers of complexity}? @@ -41,16 +40,41 @@ \item \textbf{Research Questions}\\ We wish to enable the development of predictive analysis components for DCDT's by designing a predictive DCDT. - We ask the following research questions. \todo{Read from slide boxes.} + We ask the following research questions. \todo{Read from slide boxes and explain for each (1) describe why it's important (2) say why it's challenging (3) say what makes it scientific.} \item \textbf{Literature Survey}\\ + This is the most exciting part of the thesis for me. + To answer \textbf{RQ1} we conduct a comprehensive literature survey. + We did not conduct the systematic literature survey by Kitchenham \etal, instead we relied heavily on snowballing and manual search for works in Google Scholar and DBLP. + + Google Scholar referred us to ACM Digital Library, IEEExplore, Science Direct and others. + We used structured queries such a ``datacenters'' \texttt{AND} ``digital twinning'' or plainly ``datacenter digital twins''. + + To filter out relevant work we read the abstract, introduction and conclusion and afterwards decided whether to include the article. + The results are in \textbf{Table 1.1}. + \todo{"Read the slide box."} \item \textbf{System Model}\\ + We also created a holistic system model of DCDTs. + We decided to make a system model instead of a taxonomy, because we discuss the design of a set of systems, and there are not that many to consider making a full \emph{Linnaeus} tree and a taxonomy. + + The system model is in \textbf{Fig. 1.3}, and what I found to be the most interesting while reading the literature was the lack of the connection between the two twins. + As such, what makes this design special is the \emph{Digital Thread}. \todo{Read the slide box.} \item \textbf{Reference Architecture and Prototype}\\ - A reference architecture. - The design is more specific than the reference architecture. + From the literature survey, we gathered the potential use-cases of our system, which we omit for brevity. + From the use-cases we developed as set of functional and non-functional requirements, based on which we created the reference architecture. + + The most innovative part of the data pipeline is the use of both in-band and out-of-band data pipelines, by including both a short-term cache and a long-term database. + + The most interesting thing that I devised myself, is the predictive analytics component. + \todo{Go through the elements in the plot.} + + Given this reference architecture, we created a prototype, called \emph{Sunfish}. + We evaluate this prototype in the following slides. + \item \textbf{Novel Evaluation Method}\\ + Now we go to the most difficult part. In order to evaluate a prototype, we propose a novel approach. Many researchers do not have a real facility to experiment with. We propose to use a second simulator to act as the real datacenter. |