ONI testing results
ONI tested for content filtering on five ISPs—
Adanet, AzerOnline, AzEuroTel, AZnet, and
BakInternNet—as well as several end-use locations (such as cybercafés). The results indicate
that only AZnet engages in filtering, and that filtering is both limited and backed by an explicit filtering policy. AZnet blocks obscene and erotic
content and gay and lesbian sites, as well as certain hacker and dating sites. Some sites with
commercial content, such as gambling and drug
use, are also inaccessible to users. One religious
site was also blocked. Filtering on AZnet is
explicit, with the user receiving a blockpage indicating that access is blocked as it violates the
network’s Accepted Usage Policy. AZnet uses
Symantec Gateway Security on a backbone network for virus protection and blocking selected
sites. No other instances of persistent filtering
were detected on the other ISPs tested, and ONI
did not detect the presence of commercial filtering software at any other ISP.
However, ONI did detect the selective blocking of Web sites in early 2007 during protests
against the government’s unpopular decision to
raise consumer prices of basic utilities. The
affected sites were temporarily inaccessible from
ten Azeri ISPs, and the editor of one of the sites
was detained by the police for a few hours, but
later released without charge. 9
At the cybercafé level, many owners impose
restrictions that prevent users from downloading
large attachments and visiting certain pornographic sites. But these policies are not universal,
and they are implemented at the discretion of the
cybercafé owner.
At the enterprise level, most employers limit
access to the Internet through the use of intelligent firewalls that restrict downloads of files with
certain extensions (.mp3, .avi, .mpg, .mov, and so
on), as well as access to storage file servers and
to the servers of instant communication (ICQ,
MSN, Skype, and so on).
The ongoing cyber war between Azeri and
Armenian hackers has also caused disruptions to
some Web sites and ISPs. In early 2007, five
Armenian sites were inaccessible. Users viewed
a defaced Web page commenting on the political
affiliation of the Nagorno-Karabakh region. 10 At
the same time the Web site of the Azerbaijani
National TV Channel was taken down. 11 Since
most of the allegedly inaccessible sites contained oppositional political content, there are
allegations that the Azerbaijani government was
involved in the attacks. However, ONI testing
could not confirm these suspicions. ONI did not
test for political issues related to the proclaimed
independence of the Nagorno-Karabakh region.
Conclusion
Azerbaijan’s Internet remains for the most part
“free and open” as a result of the government’s
strong interest in developing the country into an
“ICT hub” for the region. With the exception of
AZnet (which has a declared filtering policy) and
the discretion of certain cybercafé owners, ONI
did not detect the presence of any systematic
policy of Internet filtering. Instances of just-in-time
filtering appear to result from “informal” requests
by state officials to ISP operators, and these were
limited in duration and scope. Moreover, public
pressure led to a swift reversal of these policies.
That said, the filtering requests appear to have
occurred extrajudicially. Given the prospect of
increased use of the Internet by Azerbaijani
opposition groups and the government’s sensitivity to opposition, we may expect to see some
attempts to regulate Internet content and further
instances of just-in-time filtering affecting opposition Web sites.
NOTES
1. See Human Rights Watch, Azerbaijan: Presidential
Elections 2003, http://www.hrw.org/backgrounder/
eca/azerbaijan/index.htm (accessed April 4, 2007).
252
Country Summaries
ONI tested for content filtering on five ISPs—
Adanet, AzerOnline, AzEuroTel, AZnet, and
BakInternNet—as well as several end-use locations (such as cybercafés). The results indicate
that only AZnet engages in filtering, and that filtering is both limited and backed by an explicit filtering policy. AZnet blocks obscene and erotic
content and gay and lesbian sites, as well as certain hacker and dating sites. Some sites with
commercial content, such as gambling and drug
use, are also inaccessible to users. One religious
site was also blocked. Filtering on AZnet is
explicit, with the user receiving a blockpage indicating that access is blocked as it violates the
network’s Accepted Usage Policy. AZnet uses
Symantec Gateway Security on a backbone network for virus protection and blocking selected
sites. No other instances of persistent filtering
were detected on the other ISPs tested, and ONI
did not detect the presence of commercial filtering software at any other ISP.
However, ONI did detect the selective blocking of Web sites in early 2007 during protests
against the government’s unpopular decision to
raise consumer prices of basic utilities. The
affected sites were temporarily inaccessible from
ten Azeri ISPs, and the editor of one of the sites
was detained by the police for a few hours, but
later released without charge. 9
At the cybercafé level, many owners impose
restrictions that prevent users from downloading
large attachments and visiting certain pornographic sites. But these policies are not universal,
and they are implemented at the discretion of the
cybercafé owner.
At the enterprise level, most employers limit
access to the Internet through the use of intelligent firewalls that restrict downloads of files with
certain extensions (.mp3, .avi, .mpg, .mov, and so
on), as well as access to storage file servers and
to the servers of instant communication (ICQ,
MSN, Skype, and so on).
The ongoing cyber war between Azeri and
Armenian hackers has also caused disruptions to
some Web sites and ISPs. In early 2007, five
Armenian sites were inaccessible. Users viewed
a defaced Web page commenting on the political
affiliation of the Nagorno-Karabakh region. 10 At
the same time the Web site of the Azerbaijani
National TV Channel was taken down. 11 Since
most of the allegedly inaccessible sites contained oppositional political content, there are
allegations that the Azerbaijani government was
involved in the attacks. However, ONI testing
could not confirm these suspicions. ONI did not
test for political issues related to the proclaimed
independence of the Nagorno-Karabakh region.
Conclusion
Azerbaijan’s Internet remains for the most part
“free and open” as a result of the government’s
strong interest in developing the country into an
“ICT hub” for the region. With the exception of
AZnet (which has a declared filtering policy) and
the discretion of certain cybercafé owners, ONI
did not detect the presence of any systematic
policy of Internet filtering. Instances of just-in-time
filtering appear to result from “informal” requests
by state officials to ISP operators, and these were
limited in duration and scope. Moreover, public
pressure led to a swift reversal of these policies.
That said, the filtering requests appear to have
occurred extrajudicially. Given the prospect of
increased use of the Internet by Azerbaijani
opposition groups and the government’s sensitivity to opposition, we may expect to see some
attempts to regulate Internet content and further
instances of just-in-time filtering affecting opposition Web sites.
NOTES
1. See Human Rights Watch, Azerbaijan: Presidential
Elections 2003, http://www.hrw.org/backgrounder/
eca/azerbaijan/index.htm (accessed April 4, 2007).
252
Country Summaries