pmsosa / eagleeye Goto Github PK

On PacketCapture fix the issue of timesteps not being set to start every 5 seconds (or X seconds).

#Pseudocode Fix Idea
if packet.time > timestep:
   new_timestep = packet.time - (packet.time - first_timestep)%5

# e.g. first_timestep = 0; new_timestep = 13;
# new_timestep = 13 - (0-13)%5 = 13 - (3) = 10.

Per Client Graphs:

• Sent & Recieved Packets per Time
• UDP & TCP Packets per Time
• Top Ports per Time

Big Cummulative Graph (all clients together):

• Total Sent & Recieved Packets per Time
• Total UDP & TCP Packets per Time
• Total Top Ports per Time

Look online for this type of thing. Have a function (or add to the packet capturing function)

While not as important, it would be interesting to gather our information and do a chart that shows the channel spectrum.

On each client table we have an entry titled "Usage". We have yet to actually fill this in.
There are multiple ways we could fill in for usage.

• A running average: the average of packets sent from the beginning of the monitoring session. (i.e Client A has sent an average of 5,000 packets/second)

• Instantaneous usage: compared to the total number of packets sent on the past timewindow (5 seconds) how many belonged to the given client. (i.e. Client A sent 60% of all packets on the last 5 seconds)

• A mix of the two previous options: Compare to the total number of packets sent from the beginning of the monitoring session, what percentage were from A. (i.e. Client A accounts for 75% of all packets sent on the network).

Or perhaps we change "usage" for some other metric, instead we can change it for something like Avg. Sent or Avg. Recieved, using the above ideas.

Adding a Dropped Packets Graph if possible. I guess a way to do it would be:

• For each packet that we send/recv save their ID.
• If we see a packet (within a close timeframe) sent/recv with the same ID. count this as a drop.
• If x amount of time passes, the packet most likely passed and thus we can erase that id from our list.

This however cant be part of clients. (otherwise we are exchanging way to much info with the front-end). Idea to avoid jsonifying this information

Calling localhost:1993/getdataset every single time is a huge waste. Make a call that returns the latest info. Perhaps the localhost could define the latest timestamp of the window of data it got, and the server sends any collected data after that.

We need to do this to be more precisce, since the line charts in chart.js are continuos and they try to estimate non existent points. In other words, charts.js will do the following:

• the green line is what chart.js estimates.
• the red is how it actually should be.

Look at doing client fingerprinting using Scapy (and nmap module).Have a function set up for it.

As of 34e2250 a help page has been added. All that is missing is to finish adding help icons (?) to the terms that might need it.

After we are done with the other Front-End tasks. We must beautify it. Perhaps with Bootstrap

What is your age?
Have you ever set up a wi-fi router?
Did you set your own password or router name on the router?
Do you know about changing wi-fi channels?
How do you solve problems with your router?

• Call ISP, Turn Off and On, Check in-router sites
  Do you consider yourself tech savvy?

Probably want to change those things/ do those calculations inside main.py

Define were the charts and the layout in general

Can you add the description like video screen-recording how to configure and how it works etc..
Finding difficult to configure ...Thus video will be helpful.Thanks in Advance.....

Eg. Timestep = 1 : {'80' : '32' , ..... , "123": "100"} ((port):(# of packets))
The front-end will do the work of filtering through what is important and what it wants to show

I 'm just testing yours and curious to find how this works.Can you explain how this works and it will be good if u clear me with what procedure to follow to run from ubuntu without any external adapter like tp-link and to use inbuilt adapter? Thanks in advance