Merge branch 'withdocs' of github.com:Akilan1999/p2p-rendering-computation into withdocs
This commit is contained in:
29
Docs/Client.md
Normal file
29
Docs/Client.md
Normal file
@@ -0,0 +1,29 @@
|
||||
# Client Module
|
||||
This module is incharge of communicating with the server and receiving the appropriate information back from the server.
|
||||
Note: To [read more about the functions](https://pkg.go.dev/git.sr.ht/~akilan1999/p2p-rendering-computation@v0.0.0-20210404191839-6a046babcb02/client)
|
||||
|
||||
## Functions of the Client Module
|
||||
- [Interact with the Server Api](#functions-of-the-client-module)
|
||||
- [Decision maker on how the ip table is created or updated](#decision-maker-on-how-the-ip-table-is-created-or-updated)
|
||||
|
||||
|
||||
## Interact with the Server Api
|
||||
This sections talks about the functionality implemented till now.
|
||||
- The client can start docker containers using the function StartContainer.
|
||||
This functions calls the route:
|
||||
```
|
||||
http://<server IP address>:<server port>/startcontainer
|
||||
```
|
||||
TODO: Outputs and how it's printed
|
||||
|
||||
## Decision maker on how the IP table is created or updated
|
||||
- Does a local speedtest to verify and see if the server IP's in the IP table
|
||||
are pingable.
|
||||
- Downloads the Servers IP table.
|
||||
- Tries to ping the servers IP Table addresses.
|
||||
- If it's pingable then it's added as a new entry in the IP table.
|
||||
- The following steps occurs in the clients IP table.
|
||||
- To ensure that the same servers are not being called to update the IP table. There is
|
||||
a temporary list of IP address which have already been called in relation to updating the
|
||||
IP table.
|
||||
- Based on the current implementation there will 3 hops done to update the IP table.
|
||||
61
Docs/Installation.md
Normal file
61
Docs/Installation.md
Normal file
@@ -0,0 +1,61 @@
|
||||
# Installation
|
||||
|
||||
Over here we will cover the basic steps to get the server and client side running.
|
||||
|
||||
### Install Go lang
|
||||
The entire the implementation of this project is done using Go lang.
|
||||
Thus, we need go lang to compile to code to a binary file.
|
||||
[Instructions to install Go lang](https://golang.org/doc/install)
|
||||
### Install Docker
|
||||
In this project the choice of virtualization is Docker due to it's wide usage
|
||||
in the developer community. In the server module we use the Docker Go API to create and
|
||||
interact with the containers.
|
||||
|
||||
[Instructions to install docker](https://docs.docker.com/get-docker/)
|
||||
[Instructions to install docker GPU](https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/install-guide.html#docker)
|
||||
````
|
||||
Do ensure that the docker command does not need sudo to run
|
||||
````
|
||||
|
||||
### Build Project
|
||||
To set up the internal dependencies and build the entire go code
|
||||
into a single binary
|
||||
```
|
||||
make build
|
||||
```
|
||||
|
||||
### Test if binary works
|
||||
```
|
||||
./p2p-rendering-computation --help
|
||||
```
|
||||
#### Output:
|
||||
```
|
||||
NAME:
|
||||
p2p-rendering-computation - p2p cli application to create and access VMs in other servers
|
||||
|
||||
USAGE:
|
||||
p2p-rendering-computation [global options] command [command options] [arguments...]
|
||||
|
||||
VERSION:
|
||||
1.0.0
|
||||
|
||||
COMMANDS:
|
||||
help, h Shows a list of commands or help for one command
|
||||
|
||||
GLOBAL OPTIONS:
|
||||
--Mode value Specifies mode of running (default: "client") [$P2P_MODE]
|
||||
--ListServers List servers which can render tasks (default: false) [$LIST_SERVERS]
|
||||
--CreateVM value Creates Docker container on the selected server [$CREATE_VM]
|
||||
--FilePath Testing for absolute path (default: false)
|
||||
--help, -h show help (default: false)
|
||||
--version, -v print the version (default: false)
|
||||
|
||||
```
|
||||
|
||||
### Add command to .bashrc or .zshrc
|
||||
```
|
||||
export PATH=~/<Project Path>/p2p-redering-computation:${PATH}
|
||||
```
|
||||
|
||||
|
||||
#### Note: The steps on how to use the commands will be added later.
|
||||
107
Docs/P2P-testing.md
Normal file
107
Docs/P2P-testing.md
Normal file
@@ -0,0 +1,107 @@
|
||||
# Testing P2P network
|
||||
|
||||
The objective would be to test the p2p network, and the effectiveness of updating
|
||||
the ip tables. The objective of would be to give the impression to the client and
|
||||
server of a Zero configuration setting. For testing there will be a
|
||||
test network set. In the testing scenario all will be client and
|
||||
server because the IP table does not store clients IP addresses. At current
|
||||
number of hopes would be 3 as default.
|
||||
|
||||
### Test Network Scenario 1
|
||||
The test network consists of 5 nodes acting as a client and server.
|
||||
The objective would be to have the entire IP table Updated in each node
|
||||
with interacting with only 1 node once. Each node has knowledge of
|
||||
one node only.
|
||||
|
||||
|
||||
Fig 1.0 Visual Representation of testnet scenario 1
|
||||
|
||||
#### Result
|
||||
All nodes except node 1 where able to have information of IP addresses in the test net. This was due to the reason of 3 hops
|
||||
set as default. Node 1 had in it's IP table IP addresses of Node 2, Node 3, Node 4. Once the number of hops was set to 4 objective
|
||||
of the test was acheived.
|
||||
|
||||
|
||||
### Test Network Scenario 2
|
||||
The second test network has a scenario of a single peer which all the
|
||||
other nodes connect too. The scenario being when the other nodes
|
||||
connect to the single server they download information about nodes
|
||||
that have connected to the server node before.
|
||||
|
||||
### Testing Broadcast Module
|
||||
For testing the broadcast module 2 types of servers will be
|
||||
tested. One with a CPU only , another one with a CPU and GPU.
|
||||
The expected result being that the appropriate results are
|
||||
visible.
|
||||
|
||||
#### Results (CPU and GPU):
|
||||
```
|
||||
{
|
||||
"Hostname": "akilan-Lenovo-IdeaPad-Y510P",
|
||||
"Platform": "ubuntu",
|
||||
"CPU": "Intel(R) Core(TM) i7-4700MQ CPU @ 2.40GHz",
|
||||
"RAM": 7872,
|
||||
"Disk": 937367,
|
||||
"GPU": {
|
||||
"DriveVersion": "390.141",
|
||||
"Gpu": {
|
||||
"GpuName": "GeForce GT 755M",
|
||||
"BiosVersion": "80.07.A8.00.0F",
|
||||
"FanSpeed": "N/A",
|
||||
"Utilization": {
|
||||
"GpuUsage": "N/A",
|
||||
"MemoryUsage": "N/A"
|
||||
},
|
||||
"Temperature": {
|
||||
"GpuTemp": "66 C"
|
||||
},
|
||||
"Clock": {
|
||||
"GpuClock": "N/A",
|
||||
"GpuMemClock": "N/A"
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
```
|
||||
At the moment of the current implementation v1.0. Nvidia GPU
|
||||
are only compatible. As the Go code calls the command ``nvidia-smi``
|
||||
to get information about the GPU available.
|
||||
|
||||
#### Results (CPU only)
|
||||
```
|
||||
{
|
||||
"Hostname": "sv-t1.small.x86-01",
|
||||
"Platform": "ubuntu",
|
||||
"CPU": "Intel(R) Atom(TM) CPU C2750 @ 2.40GHz",
|
||||
"RAM": 7944,
|
||||
"Disk": 138793,
|
||||
"GPU": null
|
||||
}
|
||||
```
|
||||
As the ``nvidia-smi`` interface was not detected it only broadcasts
|
||||
the CPU specs available.
|
||||
|
||||
### SpeedTests
|
||||
The speed test has 3 parameters which are Ping , upload and download. The tests check if
|
||||
the results returned are approximately correct. The ping at the moment returns the correct
|
||||
result. The upload and download returned are inccorect at the moment, This is due incorrect
|
||||
implementation in for timer and will be patched in future versions.
|
||||
|
||||
### Unit tests
|
||||
All functions implemented on the P2P module returns type error.
|
||||
The units test call certain functions and check if the functions
|
||||
return an error or not. This proved sufficient as the point of
|
||||
the units tests was code coverage to check if certain functions
|
||||
return an error.
|
||||
|
||||
#### Functions tested
|
||||
This sections talks about the function called and represents
|
||||
code coverage.
|
||||
|
||||
1. ``TestServer_SpeedTest``: Function called LocalSpeedTestIpTable()
|
||||
2. ``TestReadIpTable``: Function called ReadIpTable()
|
||||
|
||||
The P2P module has a 100% code coverage in unit tests as both the unit
|
||||
tests call directly or call within the function all the functions used
|
||||
in the P2P module.
|
||||
|
||||
29
Docs/P2P.md
Normal file
29
Docs/P2P.md
Normal file
@@ -0,0 +1,29 @@
|
||||
# P2P (Peer to Peer module)
|
||||
In this repository the P2P module has been designed from sratch at the point of this implementation.
|
||||
[More about function implementation](https://pkg.go.dev/git.sr.ht/~akilan1999/p2p-rendering-computation@v0.0.0-20210404191839-6a046babcb02/p2p)
|
||||
|
||||
## Terminology
|
||||
1. IPTable: Refers to a json file which stores information about the current servers avaliable with the speedtest results ran from the Node that triggered it.
|
||||
```
|
||||
{
|
||||
"ip_address": [
|
||||
{
|
||||
"ipv4": "localhost",
|
||||
"latency": 14981051,
|
||||
"download": 8142.122540206258,
|
||||
"upload": 3578.766512629995,
|
||||
}
|
||||
]
|
||||
}
|
||||
```
|
||||
|
||||
|
||||
## Responsibility
|
||||
- To perform speed test to determine best node to connect
|
||||
- To ensure the IP table has nodes which are pingable
|
||||
- Using techniques such as [UPNP](https://en.wikipedia.org/wiki/Universal_Plug_and_Play). Still under development
|
||||
- Port Forwarding (To be introducted in a future release)
|
||||
|
||||
|
||||
## Note:
|
||||
If you are running in server mode it is recommended to use [DMZ](https://routerguide.net/when-and-how-to-setup-dmz-host-for-home-use/) to bypass the [NAT](https://en.wikipedia.org/wiki/Network_address_translation).
|
||||
35
Docs/Problems.md
Normal file
35
Docs/Problems.md
Normal file
@@ -0,0 +1,35 @@
|
||||
# Problems in implementation
|
||||
|
||||
### Number of Devices in the network:
|
||||
The current implementation has the major flaw which is that
|
||||
the server has to be in port 8088 to detected in the public
|
||||
network. As we know most personal networks have a single IP
|
||||
address. This means we cannot have duplicate ports. A fix can
|
||||
be to mention the open port on the IP table file.
|
||||
(Ex: Possible feild to be added)
|
||||
```
|
||||
{
|
||||
"ip_address": [
|
||||
{
|
||||
"ipv4": "localhost",
|
||||
"latency": 14981051,
|
||||
"download": 8142.122540206258,
|
||||
"upload": 3578.766512629995,
|
||||
"port": 8088
|
||||
}
|
||||
]
|
||||
}
|
||||
```
|
||||
|
||||
### Broadcast of container specs
|
||||
At the moment the container specs are not broadcasted rather
|
||||
just the machines specs and this module has yet to be tested
|
||||
rigorously.
|
||||
|
||||
### Intergration with GPU
|
||||
Certain machines have GPUs present in them which provide a
|
||||
huge advantage for those who want to do rendering and certain
|
||||
sort of computation. The Aim is to only allow it to be compatible
|
||||
with Nvidia. But an better idea would be to provide compatability
|
||||
with multiple GPU providers.
|
||||
|
||||
7
Docs/README.md
Normal file
7
Docs/README.md
Normal file
@@ -0,0 +1,7 @@
|
||||
# Table of contents
|
||||
1. [Architecture](#NotImplemented)
|
||||
1. [Client](Client.md)
|
||||
2. [P2P](P2P.md)
|
||||
3. [Server](#NotImplemented)
|
||||
2. [Installation](Installation.md)
|
||||
3. [Problems](Problems.md)
|
||||
12
Docs/Virtualization
Normal file
12
Docs/Virtualization
Normal file
@@ -0,0 +1,12 @@
|
||||
|
||||
Virtualization Design
|
||||
======================
|
||||
|
||||
The virtualization tool will be treated as a separate module. In our implementation
|
||||
we will use docker as it's easier to configure and set.
|
||||
|
||||
Methods to be created
|
||||
- Build OS Image from DockerFile
|
||||
- Run Image Built
|
||||
- Possibility to kill image by server admin or client side user.
|
||||
- Track stats of the docker container by server admin or client side user.
|
||||
Reference in New Issue
Block a user