more work on report

This commit is contained in:
Claudio Maggioni 2021-05-16 17:29:55 +02:00
parent 7253804f26
commit 0d935c2112
5 changed files with 135 additions and 8 deletions

1
.~lock.status.ods# Normal file
View file

@ -0,0 +1 @@
,maggicl,Apple2gs.local,16.05.2021 14:55,file:///Users/maggicl/Library/Application%20Support/LibreOffice/4;

View file

@ -39,29 +39,141 @@ header-includes:
```
---
\tableofcontents
\newpage
# Introduction (including Motivation)
# State of the Art
- Introduce Ros\'a 2015 DSN paper on analysis
- Describe Google Borg clusters
- Describe Traces contents
- Differences between 2011 and 2019 traces
## Introduction
**TBD**
## Rosà et al. 2015 DSN paper
**TBD**
## Google Borg
Borg is Google's own cluster management software. Among the various cluster
management services it provides, the main ones are: job queuing, scheduling,
allocation, and deallocation due to higher priority computations.
The data this thesis is based on is from 8 Borg "cells" (i.e. clusters) spanning
8 different datacenters, all focused on "compute" (i.e. computational oriented)
workloads. The data collection timespan matches the entire month of May 2019.
In Google's lingo a "job" is a large unit of computational workload made up of
several "tasks", i.e. a number of executions of single executables running on a
single machine. A job may run tasks sequentially or in parallel, and the
condition for a job's succesful termination is nontrivial.
Both tasks and jobs lifecyles are represented by several events, which are
encoded and stored in the trace as rows of various tables. Among the information
events provide, the field "type" provides information on the execution status of
the job or task. This field can have the following values:
- **QUEUE**: The job or task was marked not eligible for scheduling by Borg's
scheduler, and thus Borg will move the job/task in a long wait queue;
- **SUBMIT**: The job or task was submitted to Borg for execution;
- **ENABLE**: The job or task became eligible for scheduling;
- **SCHEDULE**: The job or task's execution started;
- **EVICT**: The job or task was terminated in order to free computational
resources for an higher priority job;
- **FAIL**: The job or task terminated its execution unsuccesfully due to a
failure;
- **FINISH**: The job or task terminated succesfully;
- **KILL**: The job or task terminated its execution because of a manual request
to stop it;
- **LOST**: It is assumed a job or task is has been terminated, but due to
missing data there is insufficent information to identify when or how;
- **UPDATE_PENDING**: The metadata (scheduling class, resource requirements,
...) of the job/task was updated while the job was waiting to be scheduled;
- **UPDATE_RUNNING**: The metadata (scheduling class, resource requirements,
...) of the job/task was updated while the job was in execution;
Figure \ref{fig:eventTypes} shows the expected transitions between event types.
![Typical transitions between task/job event types according to Google
\label{fig:eventTypes}](./figures/event_types.png)
## Traces contents
The traces provided by Google contain mainly a collection of job and task events
spanning a month of execution of the 8 different clusters. In addition to this
data, some additional data on the machines' configuration in terms of resources
(i.e. amount of CPU and RAM) and additional machine-related metadata.
Due to Google's policy, most identification related data (like job/task IDs,
raw resource amounts and other text values) were obfuscated prior to the release
of the traces. One obfuscation that is noteworthy in the scope of this thesis is
related to CPU and RAM amounts, which are expressed respetively in NCUs
(_Normalized Compute Units_) and NMUs (_Normalized Memory Units_).
NCUs and NMUs are defined based on the raw machine resource distributions of the
machines within the 8 clusters. A machine having 1 NCU CPU power and 1 NMU
memory size has the maximum amount of raw CPU power and raw RAM size found in
the clusters. While RAM size is measured in bytes for normalization purposes,
CPU power was measured in GCU (_Google Compute Units_), a proprietary CPU power
measurement unit used by Google that combines several parameters like number of
processors and cores, clock frequency, and architecture (i.e. ISA).
## Overview of traces' format
The traces have a collective size of approximately 8TiB and are stored in a
Gzip-compressed JSONL (JSON lines) format, which means that each table is
represented by a single logical "file" (stored in several file segments) where
each carriage return separated line represents a single record for that table.
There are namely 5 different table "files":
- `machine_configs`, which is a table containing each physical machine's
configuration and its evolution over time;
- `instance_events`, which is a table of task events;
- `collection_events`, which is a table of job events;
- `machine_attributes`, which is a table containing (obfuscated) metadata about
each physical machine and its evolution over time;
- `instance_usage`, which contains resource (CPU/RAM) measures of jobs and tasks
running on the single machines.
The scope of this thesis focuses on the tables `machine_configs`,
`instance_events` and `collection_events`.
## Remark on traces size
While the 2011 Google Borg traces were relatively small, with a total size in
the order of the tens of gigabytes, the 2019 traces are quite challenging to
analyze due to their sheer size. As stated before, the traces have a total size
of 8 TiB when stored in the format provided by Google. Even when broken down to
table "files", unitary sizes still reach the single tebibyte mark (namely for
`machine_configs`, the largest table in the trace).
Due to this constraints, a careful data engineering based approach was used when
reproducing the 2015 DSN paper analysis. Bleeding edge data science technologies
like Apache Spark were used to achieve efficient and parallelized computations.
This approach is discussed with further detail in the following section.
# Project requirements and analysis
(describe our objective with this analysis in detail)
**TBD** (describe our objective with this analysis in detail)
# Analysis methodology
## Technical overview of traces' file format and schema
**TBD**
## Overview on challenging aspects of analysis (data size, schema, avaliable computation resources)
## Introduction on apache spark
**TBD**
## Introduction on Apache Spark
**TBD**
## General workflow description of apache spark workflow
**TBD** (extract from the notes sent to Filippo shown below)
The Google 2019 Borg cluster traces analysis were conducted by using Apache
Spark and its Python 3 API (pyspark). Spark was used to execute a series of
queries to perform various sums and aggregations over the entire dataset
@ -110,7 +222,11 @@ compute and save intermediate results beforehand.
## General Query script design
## Ad-Hoc presentation of some analysis scripts (w diagrams)
**TBD**
## Ad-Hoc presentation of some analysis scripts
**TBD** (with diagrams)
# Analysis and observations
@ -271,14 +387,24 @@ Refer to figure \ref{fig:figureV}.
## Potential causes of unsuccesful executions
**TBD**
# Implementation issues -- Analysis limitations
## Discussion on unknown fields
**TBD**
## Limitation on computation resources required for the analysis
**TBD**
## Other limitations ...
**TBD**
# Conclusions and future work or possible developments
**TBD**
<!-- vim: set ts=2 sw=2 et tw=80: -->

Binary file not shown.

Binary file not shown.

After

Width:  |  Height:  |  Size: 34 KiB

Binary file not shown.