We began our analysis with the fixed alpha values, but now we are heading towards another dimension in TULIP which is calculating the distance based on the RTT's. We start our work by pinging each landmark to another landmark as we already have landmark latitudes and longitudes we can easily find the distance between the two. These results are in the form of tsv file containing the records for landmark source to landmark destination their rtt and distance an example of this file is given below
Landmark Source |
Landmark Destination |
RTT |
Distance |
|---|---|---|---|
icfamon.rl.ac.uk |
csplanet02.cs-ncl.net |
19.242 |
83.02 |
planetlab1.cs.vu.nl |
csplanet02.cs-ncl.net |
22.986 |
357.77 |
host4-plb.loria.fr |
csplanet02.cs-ncl.net |
33.77 |
548.47 |
After getting these results, I then edited the available schema of TULIP database and added a new table to cater for this change. The schema for new table is given below:
+-----------------------------------+------+
| Field | Type | Null | Key | Default | Extra |
+-----------------------------------+------+
| ipv4Add_f | varchar(50) | YES | | NULL | |
| ipv4Add_t | varchar(50) | YES | | NULL | |
| rtt | varchar(50) | YES | | NULL | |
| distance | varchar(50) | YES | | NULL | |
+-----------------------------------+------+
This table is then populated with recently collected data with the help of perl script "rttdistinput.pl"deployed at:
/afs/slac.stanford.edu/package/pinger/tulip