Purpose
Analyze new results for TULIP2 (using PlanetLab Landmarks) with TULIP1 (Using only PingER landmarks). We have used Speed net sites as our sample data. The location of these sites is already known from Speed Net database. The results using TULIP1 (PingER only sites) can be downloaded here
North America
There are about 52 Speednet sites in North America region. We conducted some first using TULIP1 (PingER only landmarks). We have 10 PingER landmarks in North America region. Then we conducted some more tests using TULIP2 (using both Pinger and Planetlab landmarks). We used about 80 Planetlab landmarks in North America region. When we compared the two results. there were some obvious improvements in the results. The results can be found here Excel sheet can be downloaded here . Here is the frequency of landmarks used in North America. The most frequent landmark used is Chicago_UnitedStates_PL.
Chicago_UnitedStates_PL |
Europe
There are about 52 Speednet sites in Europe. We also compared TULIP1 and TULIP2 for Europe and there were obvious improvements in the results. We have 8 PingER landmarks in Europe. We used 48 Planetlab landmarks in Europe. The results for Europe can be found here . The results can be downloaded here
Interesting Observations
While doing the analysis we came across an interesting result for a speed net site. We conducted the initial tests on 2nd January,2008. When we tried to locate http://www.g3pcs.com IP address 208.109.181.237 the three sites using for triangulation were Eugene_UnitedStates_PL 12.5ms Seattle_UnitedStates_PL 15.2ms and TRIUMF_Vancouver 19.1 ms. The Speed database showed that the site was in Ephrata, WA (Latitude:47.32 Longitude: -119.55). Although there was some Error distance but we believed the Speed Net database to be correct because Eugene,Seattle and Vancouver are close to Ephrata.
Latter on (4th March,2008) we conducted some new tests. This time three different sites responded with min RTT. SanDiego_UnitedStates_PL 13.337ms, LosAngeles_UnitedStates_PL 13.696ms and Riverside_UnitedStates_PL 14.342ms. The was located somewhere in California state. This is very interesting as we think that the site has moved since then. When we used the geoiptool to locate the site it gave Scottsdale, Arizonan(Latitude: 33.6 Longitude -111.89). We have a Planet Lab landmark in Phoenix, Arizona but that did not responded. If that site would have responded we would have got better results. A landmark in Tuscon, AZ responded with 23ms.
So our conclusion is that the site has moved between 2nd January,2008 and 4th March,2008. Hostip.info and geobytes.com says the host is in Brooklyn, NY which is obviously wrong. The whois information from (www.ip-lookup.net) suggests the site is in Scottsdale, AZ which we think is correct. here is the whois information:
OrgName: GoDaddy.com, Inc. OrgID: GODAD Address: 14455 N Hayden Road Address: Suite 226 City: Scottsdale StateProv: AZ PostalCode: 85260 Country: US |
Locating PingER Sites
Using TULIP2 we located all PingER sites in North America. The results can be found here We conducted Tests both using alpha=100km/ms and 40km/ms Currently we are comparing the two results to see, where and how much improvement we are getting.
Alpha Comparison ( 100 km/ms vs 40 km/ms)
A series of three tests were performed to compare the value of 100 vs 40 for the value of alpha.
Landmarks |
Frequency |
|
|
|
|
|
|
Chicago_UnitedStates_PL |
34 |
||
NewYork_UnitedStates_PL |
31 |
||
FNAL,Chicago |
26 |
|
|
Philadelphia_UnitedStates_PL |
26 |
||
Bloomington_UnitedStates_PL |
22 |
||
Fairfax_UnitedStates_PL |
11 |
||
Washington_UnitedStates_PL |
11 |
Previously the TULIP algorithm located a host only giving latitude and longitude as integers without any floating points, as a result a lot of precision was lost. Therefore the TULIP algorithm was improved to give floating point precision. In order to test the new accuracy some tests were conducted.
After careful selection of landmarks (removing non-responding landmarks and writing scripts to dynamically generate landmarks) on 20th Aug,1008 we obtained much better results. These results are compared to the old results in order to see the improvements. The results can be download here Here are some quick observations:
1. Menlo Park Landmark (Stanford) is giving wrong result (rtt 2ms) which is ruining the results. We need to disable this landmark since we already have landmark at SLAC.
2. There are some major improvements in results due to careful selection of landmarks.
3. the median error is 170km, 25% 100km, 75% 362km & IQR 262km. This is including the numbers in red.
4. Need to come up with a method to avoid bad results for replicated servers.
Hosts that are on satellite links have very large delays. They cannot be located using the RTTs from a set of landmarks. Usually they have RTTs > 500ms. Most of the Central Asian countries (including Afghanistan) and some African countries use satellite links. Data from hosts in Kazakhistan, Burundi, Madagascar, Uganda, Afghanistan and Nigeria can be found here. The three best Landmarks chosen gives an RTT greater than 500ms. If the RTT from the three best landmarks is greater than 500ms then our algorithm returns the host as satellite link and does not perform trilateration. This figure may be changed latter.