openwsn-berkeley / mercator Goto Github PK

Example:

python mercatorRunExperiment.py local 03oos_mercator_iot-lab_M3.elf Traceback (most recent call last): File "mercatorRunExperiment.py", line 303, in <module> main() File "mercatorRunExperiment.py", line 291, in main expid = submit_experiment(args.testbed, args.firmware, args.duration) File "mercatorRunExperiment.py", line 256, in submit_experiment nodes = tb_json[testbed_name] KeyError: 'local'

setup

Two OpenMotes running the latest firmware

issue

After telling a device to send, it stops responding to further commands.

to reproduce

> rx /dev/ttyUSB0
> tx /dev/ttyUSB1

After a few seconds:

    `type             : IND_TXDONE`
> st /dev/ttyUSB0
`{'status': 4, 'type': 2, 'mac': (0, 18, 75, 0, 6, 13, 158, 223), 'numnotifications': 0}`
> st /dev/ttyUSB1
---------------------------/dev/ttyUSB1
None

Traceback (most recent call last):
File "mercatorRunExperiment.py", line 283, in
main()
File "mercatorRunExperiment.py", line 279, in main
site = site
File "mercatorRunExperiment.py", line 70, in init
self._doExperimentPerFrequency(freq)
File "mercatorRunExperiment.py", line 82, in _doExperimentPerFrequency
self._doExperimentPerTransmitter(freq,transmitterPort)
File "mercatorRunExperiment.py", line 111, in _doExperimentPerTransmitter
txfillbyte = self.TXFILLBYTE,
File "/senslab/users/brun/mercator/code/software/app/../lib/MoteHandler.py", line 178, in send_REQ_RX
txfillbyte,
struct.error: ubyte format requires 0 <= number <= 255

The new script will be called preprocess.py and will call the existing script.

Ex:
python preprocess.py strasbourg
This command should run pdr_dist.py, pdr_freq.py, rssi_dist.py ... for every dataset of the strasbourg site.

If a chart_config.json file is present in the dataset folder, the chart will be display in the Mercator website using this configuration.

After one our on a large experiment (>350 nodes), the following nodes are not responding:

-----------timeout--------------m3-220
-----------timeout--------------m3-1
-----------timeout--------------m3-30
-----------timeout--------------m3-189
-----------timeout--------------m3-286
-----------timeout--------------m3-285
-----------timeout--------------m3-283
-----------timeout--------------m3-282
-----------timeout--------------m3-281
-----------timeout--------------m3-280
-----------timeout--------------m3-289
-----------timeout--------------m3-288
-----------timeout--------------m3-306
-----------timeout--------------m3-307
-----------timeout--------------m3-304
-----------timeout--------------m3-305
-----------timeout--------------m3-302
-----------timeout--------------m3-303
-----------timeout--------------m3-301
-----------timeout--------------m3-308

Having everything in one branch starts to make the git repository slow to load and hard to navigate.

The following error occurs when running the experiment in local with two m3 nodes.
The second device does not respond anymore after the first step ("switch /dev/ttyUSB3 to idle").
It is not device dependent.
It does not append when using the mercatorCLI script and following the same steps.

/dev/ttyUSB1
/dev/ttyUSB3
freq=11 transmitterPort=/dev/ttyUSB3
    switch /dev/ttyUSB3 to idle
    switch /dev/ttyUSB1 to idle
-----------timeout--------------/dev/ttyUSB3
    switch /dev/ttyUSB3 to RX
    switch /dev/ttyUSB1 to RX
-----------timeout--------------/dev/ttyUSB3
    switch /dev/ttyUSB3 to TX
freq=11 transmitterPort=/dev/ttyUSB1
    switch /dev/ttyUSB3 to idle
    switch /dev/ttyUSB1 to idle
-----------timeout--------------/dev/ttyUSB3
    switch /dev/ttyUSB3 to RX
    switch /dev/ttyUSB1 to RX

Currently we can only run 255 transactions.

When the app is no able to connect with a mote, it crashes the entire experiment.

Should we force the app to continue ? (we can make this feature an option)

The following error is raised:

could not connect to m3-291, reason: [Errno 111] Connection refused
Traceback (most recent call last): 
  File "mercatorRunExperiment.py", line 283, in <module>
    main()
  File "mercatorRunExperiment.py", line 279, in main
    site = site
  File "mercatorRunExperiment.py", line 56, in __init__
    self.motes[s]    = MoteHandler.MoteHandler(s,self._cb)
  File "/senslab/users/brun/mercator/code/software/app/../lib/MoteHandler.py", line 45, in __init__
    raise SystemError(msg)
SystemError: could not connect to m3-291, reason: [Errno 111] Connection refused

We process the dataset to have "one_to_many" and "one_to_one" results.
Currently only "one_to_many" results are present.
We should create "one_to_many" and "one_to_one" to separate the future "one_to_one" results.

If some HDCL flags are corrupted, the node will not understand the command. However, this issue should not be fatal.

When running mercatorRunExperiment, the user should be able to specify the type of board.
Ex: -b m3

https://github.com/iot-lab/iot-lab/wiki/Getting-attributes-for-IoT-LAB-nodes

Instead of display the state of each node (ex: switch m3-325 to RX), display the current channel and the progression of the experiment.

https://github.com/chartjs/chartjs-plugin-zoom

The component should allow users to explore the processed dataset in a tree-like approach

On graphs such as waterfall plot, it would be interesting to see the average line on top of the existing values.

When looking at the logs, we can observe lines like this one:

timestamp,mac,frequency,length,rssi,crc,expected,srcmac,transctr,pkctr,txnumpk,txpower,txifdur,txlength,txfillbyte
2016-10-17_18.45.41,05-43-32-ff-03-df-93-67,11,100,-91,0,0,05-43-32-ff-03-db-c3-68,1,4129,10,0,100,100,0x0a

4129 is not valid as only 10 pkts are sent.

For preprocessed datasets with 3 metrics, display multiple graphs.

Example for PDR over each frequency along time:
The website should display one graph per frequency (I did not managed to have multiple frequencies on the same graph).

The preprocessed dataset should be under a folder with the following form: metric1_metric2_metric3
Ex: pdr_time_freq

With the following directory tree:

pdr_time_freq/
    11/
        time1.json
        time2.json
        ...
    12/
    13/
    ...

Currently, experiment type selection is done by clicking on buttons. To improve the user experience, the experiment type selection should be done when selection the observed nodes:

• when no nodes are selected, the overall graphs are presented (m2m)
• when selecting the first node, the observed node is the transmitting node (o2m)
• when selecting the following nodes, the observed nodes are receiving nodes (o2o)

This hides the one_to_one, one_to_many and many_to_many technical concepts.