Output one min/max value per interval.
Output a curve.

20190723 10:34:06 - 0.001888 0.000722 0.003263 0.600000 0.000000 1.000000 0.000000 0.000000 0.000000 1.000000 0.000000
20190723 10:34:06 - 0.001888 0.000721 0.003263 0.600000 0.000000 1.000000 0.000000 0.000000 0.000000 1.000000 0.000000
20190723 10:34:11 - 0.002062 0.000711 0.003490 0.800000 0.000000 1.000000 0.000000 0.000000 0.000000 1.000000 0.000000
20190723 10:34:11 - 0.002062 0.000711 0.003490 0.800000 0.000000 1.000000 0.000000 0.000000 0.000000 1.000000 0.000000
20190723 10:34:16 - 0.002133 0.000782 0.003179 0.800000 0.000000 1.000000 0.000000 0.000000 0.000000 1.000000 0.000000
20190723 10:34:16 - 0.002133 0.000783 0.003180 0.800000 0.000000 1.000000 0.000000 0.000000 0.000000 1.000000 0.000000
20190723 10:34:21 - 0.002613 0.001238 0.003393 0.800000 0.000000 1.000000 0.000000 0.000000 0.000000 1.000000 0.000000
20190723 10:34:21 - 0.002613 0.001238 0.003393 0.800000 0.000000 1.000000 0.000000 0.000000 0.000000 1.000000 0.000000

Example output:
Get request received is x% from all received requests
Function Inputs:(dest IP, dest port, src IP, Src port, monitoring time window)

server.c - connection_handler

for example:
https://abc.xyz.com/path/somefilename.xy?param1=blie&param2=bla
https://abc.xyz.com/path/somefilename.xy
https://abc.xyz.com/path/somefilename.xy/
https://abc.xyz.com/path/logout
https://abc.xyz.com/path/search/?param1=blie&param2=bla
https://abc.xyz.com/path/search?param1=blie&param2=bla

Display the CDF graph of the most popular X requested objects(e.g:somefilename.xy ) or the method (search or logout).

Example output:
Responses with 200 is X% of the the total returned responses
Responses with 304 is Y% of the the total returned responses
Function Inputs: (dest IP, dest port, src IP, Src port monitoring time window)

Output one min/max value per interval.
Output a curve.

server.c

• direction (1 graph with 6 curves), min/max/avg
• move the client filter for all the metrics
• number of bytes (tcp payload size)

1 -Global To
2- Standalone function for each metric.

• input/output
• with r w/o params

3 -min/max refactoring

4 -ST

• for all client : destination

TP
a specific From

num of clients

cumulative cdf for request path

parametrable number

• specify a client and server
• chose the direction(req/rsp)
• number of bytes/s
• min/max/avg

main.c
int GP_CAP_FIN = 0

I needed to substitute this global var for a shared var among thread. Want to discuss if this change can be problematic.

The sniffer should send a close message to the front-end once the analysis is done

Restart it on failure

remove the validation based on request seq_next and response ack

http://tuxthink.blogspot.com/2013/02/using-read-write-lock-in-pthreads.html

Measureing the cpu and memory utilization difference for the sniffer between the following two scenarios:

pcap_next paramter is set to 1000 ms

pcap_next paramter is set to -1

The components (front end, back end and sniffer), the technology of each one, the communication protocol along with the test application (YCSB) components.

System architecture (Components/ communication protocols used/ function(s) of each componet)
Installation guide/configuration/environment of each component along with issues met and how are resolved
code structure and documentation.

data.c

We can maybe remove them, as only one thread write these values.
One lock per variable?

• rst
  -rstmin
  ...

in main.c - capture_main

//TODO: Check with this is commented! Memory leak.
// Is it to solve the seg Fault?
//packet_free(packet);

Change was made in

• process_packet_queue
• process_flow_queue

We notice a problem with the response stamp (greater than the request's one)

server.c - extract_data
For the saddr/daddr allocation.

The flow is moved from the hash_table to the flow_queue once the connection is closed. This is a problem when the connection is persistent.

• add a division for the interval
• add the axis unit