Analysis of the blocking of archive.org in India between 2017-07-20 and 2017-08-30, using data from OONI, the Open Observatory of Network Interference.
Related links:
David Fifield <david@bamsoftware.com>
Gray dots are normal (unblocked) replies. Colored dots indicate some kind of error.
Consistent with reports in the blog post, only blocking of HTTP was detected, not HTTPS. Excluding one ambiguous measurement on 2017-08-03, the first measurement with evident blocking is at 2017-08-05 09:35:47.
These autonomous systems exhibited blocking (and also had an unblocked measurement prior to 2017-08-05):
AS24560 | AIRTELBROADBAND-AS-AP Bharti Airtel Ltd., Telemedia Services | (Netsweeper block page) |
AS45609 | BHARTI-MOBILITY-AS-AP Bharti Airtel Ltd. AS for GPRS Service | (connection reset) |
AS55836 | RELIANCEJIO-IN Reliance Jio Infocomm Limited | (Netsweeper block page) |
These autonomous systems showed no HTTP blocking after 2017-08-05:
AS17488 | HATHWAY-NET-AP Hathway IP Over Cable Internet |
AS17803 | BSES-AS-AP BSES TeleCom Limited |
AS133982 | EXCITEL-AS-IN Excitel Broadband Private Limited |
These autonomous systems had no HTTP measurements after 2017-08-05, so their blocking status is unknown. I am not sure why AS4134 appears once in the data (that AS should be in China, not India).
AS4134 | CHINANET-BACKBONE No.31,Jin-rong Street |
AS9829 | BSNL-NIB National Internet Backbone |
AS45820 | TTSL-MEISISP Tata Teleservices ISP AS |
AS55410 | VODAFONE-NET-AS-AP C48 Okhla Industrial Estate, New Delhi-110020 |
AS55824 | NKN-CORE-NW NKN Core Network |
In AS24560, requests for http://archive.org received an explicit block page:
<meta name="viewport" content="width=device-width,initial-scale=1.0,maximum-scale=1.0"/><style>body{margin:0px;padding:0px;}iframe{width:100%;height:100%}</style><iframe src="http://www.airtel.in/dot/?dpid=1&dpruleid=3&cat=107&dplanguage=-&url=http%3a%2f%2f207%2e241%2e224%2e2%2f" width="100%" height="100%" frameborder=0></iframe>
The page at the embedded URL, http://www.airtel.in/dot/?dpid=1&dpruleid=3&cat=107&dplanguage=-&url=http%3a%2f%2f207%2e241%2e224%2e2%2f [archive], says "Your requested URL has been blocked as per the directions received from Department of Telecommunications, Government of India. Please contact administrator for more information."
The format of the URL shows that the censorship device is made by Netsweeper, a company that is notorious for aiding censorship. Incidentally, essentially the same blocking [archive] has been documented on Airtel going back at least to July 2016.
The block page includes advertisements for ISP-affiliated movies, music, and games.
The ads only appear if you are using a mobile browser
(specifically
[archive],
if the browser User-Agent matches
Android|webOS|iPhone|iPad|iPod|BlackBerry|IEMobile|Opera Mini
).
The page also includes some kind of Akamai-hosted performance-measuring script [archive] based on episodes.
Raw OONI measurements:
In AS45609, a request for http://archive.org was terminated with a TCP connection reset.
Raw OONI measurements:
In AS55836, requests for http://archive.org received a Netsweeper block page, similar to the one in AS24560:
<meta name="viewport" content="width=device-width,initial-scale=1.0,maximum-scale=1.0"/><style>body{margin:0px;padding:0px;}iframe{width:100%;height:100%}</style><iframe src="http://49.44.18.34:8080/webadmin/deny/index.php?dpid=1&dpruleid=3&cat=107&dplanguage=-&url=http%3a%2f%2farchive%2eorg%2f" width="100%" height="100%" frameborder=0></iframe>
The URL it embeds, http://49.44.18.34:8080/webadmin/deny/index.php?dpid=1&dpruleid=3&cat=107&dplanguage=-&url=http%3a%2f%2farchive%2eorg%2f, timed out when I tried to access it.
A earlier measurement on 2017-08-03 is also marked "http-failure" because of a timeout. But the simultaneous control measurement also had an error (HTTP status code 503), so it could be a temporary server failure, not blocking.
Raw OONI measurements:
Another two weeks later, we find that blocking Three days later, new measurements show that blocking has ceased in AS24560 and AS55836. In AS45609, there have been no new measurements since the blocked one at 2017-08-08 15:07:14, so it is unknown whether blocking continues.
Raw OONI measurements:
Three days later, new measurements show that blocking has ceased in AS45609, but after a general break in blocking, there is new blocking AS17488 and AS17813, which were previously unblocked.
AS17488 | HATHWAY-NET-AP Hathway IP Over Cable Internet | (block page) |
AS17813 | MTNL-AP Mahanagar Telephone Nigam Limited | (block page) |
In AS17488, a request for http://archive.org got a block page (the same as in AS17813):
<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01//EN"> <html> <head <title> </title> </head> <body> <h0> <font color="black"> <font size="4.5"> “This website/URL has been blocked as per instructions from Department of Telecommunications of India.” </font> </h0> </body> </html>
Raw OONI measurements:
In AS17813, a request for http://archive.org got a block page (the same as in AS17488):
<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01//EN"> <html> <head <title> </title> </head> <body> <h0> <font color="black"> <font size="4.5"> “This website/URL has been blocked as per instructions from Department of Telecommunications of India.” </font> </h0> </body> </html>
Raw OONI measurements:
Prerequisites:
Download recent OONI reports from India:
ooni-sync -directory reports.in -xz since=2017-07-20 until=2017-08-30 test_name=web_connectivity probe_cc=IN
Filter reports into a CSV file:
(echo date,report_id,country,asn,url,blocked,blocking_type; xz -dc reports.in/*.json.xz | jq -j -f makecsv.jq --unbuffered) | tee in.csv
Generate a graph (blocking-in.png):
Rscript blocking-in.R