Unintended Consequences of Website Blocking
When website blocking becomes the natural alternative used to limit undesired behaviors, one should be aware of the implications of such measures. These implications might range from the involuntary and intermittent disconnection of a third-party website to completely disabling the networks of operators in a given area or even in an entire country. In any case, it should be noted that the consequences of such actions may typically go unnoticed by clients and/or users both within and outside the region.
It is common to observe restrictions on the use of and/or access to the Internet, implemented either by blocking certain websites or domain names or by limiting the use of content, all of which causes a fragmentation of the Internet. As a result, these actions create technical complexities that affect users who access different services in content networks.
When we speak about network blocking, we tend to think about “filtering” users’ ability to use and/or access the Internet. However, the two are not the same, as the term “blocking” typically refers to preventing access to all targeted resources, while “filtering” refers to preventing access to specific elements of such resources.
There are many documents describing the techniques that can be used for blocking based on protocol and IP address, deep packet inspection, URL, search platforms, DNS, etc. However, each blocking or filtering results in actions that affect other services or users at local and/or regional level.
Technical Implications of Website Blocking
While it is technically possible to block a website from a network, as the number of networks and operators involved increases, this causes unintended consequences. This becomes even more complex if the networks and/or operators are distributed among different territories or jurisdictions.
It is commonly thought that a website can be blocked based on its (domain) name, its IP address, or a combination of the two (domain name and related IP address). Reality, however, is significantly more complex, as a website or web page is comprised of multiple objects, many of which can use different domain names and/or be hosted at different IP addresses.
Just as a website can be composed of multiple objects hosted at different IP addresses, the reverse situation is also possible, i.e., IP addresses can individually host multiple websites. This is typically the case of companies known as Hosting Providers and Content Distribution Networks (CDN). In such cases, blocking the IP address of a Hosting Provider or CDN may result in blocking hundreds or thousands of websites that have nothing to do with the initial purpose of the blocking measure.
When considering blocking a website, an Internet service provider should take into account the factors mentioned above.
IP address blocking techniques entail risks to the operator’s own network: in case of blocking addresses that absorb large volumes of traffic, the devices on the blocking operator’s network can affect the operation of their own network and those of other operators interacting with them.
The same is true when attempting to block a website for networks involving more than one operator. In this case, entire networks belonging to entities that favor proper Internet operation at global level might be blocked, both in the affected territory as well as within the network of those operators involved. Services such as public and private DNS (servers that translate domain names to IP addresses, necessary for global Internet operation) might be compromised or completely blocked.
Another consideration worth noting is the level of effectiveness that site blocking based on IP addresses or domain names might achieve. Ironically, most organizations that maintain their websites lawfully would need several days to recover from a block, yet precisely the organizations that are constantly in court for alleged infringements of intellectual property rights or other illegal behaviors are better prepared and can recover from such blocks in a matter of hours (or sometimes even in a matter of minutes). In addition, once they their IP addresses or main domain names have been blocked, they can easily move to another datacenter or a different provider in another jurisdiction and/or quickly secure another domain name hosted at a different IP address, thus bypassing the original block of their website. Nevertheless, by ordering or executing a block without having the proper expertise, the blocking party will probably leave a trail of unintended side effects that will affect legitimate users across the Internet.
Non-Technical Implications of Website Blocking
The technical implications of website blocking are not the only implications that should be considered. As mentioned earlier, attempting to block a website based on its IP address can result in the inadvertent blocking of hundreds or thousands of other, completely unrelated websites, potentially infringing on constitutional precepts or high-level laws that promote the protection of freedom of expression and the right to information.
We should consider that these websites may be hosting not only commercial initiatives, but also the websites of government agencies (the president’s office, the tax collection office, the right to information, etc.), non-profit organizations and others that will probably have a greater impact among citizens.
We should also take into account the fact that networks are not necessarily bound to a specific jurisdiction. This means that the block that an operator is ordered to implement may not apply to the entire intended territory (best-case scenario) or that it may exceed the targeted territory and leave the website unavailable to thousands of Internet users who have the right to access the content.
LACNIC believes that website blocking should in no case be the result of precautionary measures in a judicial process. As already noted, the unintended technical implications can be catastrophic for the operation of a network or a group of networks, so the decision to block a website should be the result of a due process that identifies blocking as the only feasible course of action once all other possibilities have been exhausted and all implications have been analyzed. In addition to the risks mentioned above, blocking a website as a precautionary measure (sometimes at the request of the plaintiff in case of an alleged violation of their intellectual property rights) might result in the indefinite extension of the court proceedings, as the plaintiff obtained the resulting thing from a non-finalized dispute, leaving the particular group of IP addresses unavailable to Internet users and therefore harming legitimate network operators and third-party websites reachable under the same IP address (or IP addresses).
Likewise, unlike measures that are the result of due process of law, requests to implement blocking as a precautionary measure ordered by an administrative authority may undermine certain human and/or civil rights, as has been the case in several jurisdictions where major corporate brand holders have abused these mechanisms before the administrative authorities (trademark or copyright) and had them delete websites criticizing one of their brands or products, alleging the infringement of trademarks or copyrights when what they were actually seeking was to avoid criticism of one of their products or services.
Finally, the non-technical consequences of blocking a website based on its domain name lie mainly in the inefficiency of this measure. We already observed that a website can use different domain names. This means that the same website might answer to different names, in which case blocking one of its names will not make the website unavailable under the others. The above requires an in-depth analysis to identify every name to which a particular website might respond, thus increasing the chances that one or more of those names will be completely unrelated to the alleged offense.
Blocking websites indefinitely based on their domain names (URL) or IP addresses is inappropriate under any circumstance. If this is the only measure used to discourage or prevent illegal behavior, the block should specify a specific period of time after which it will expire.
The repeated implementation of these blocking mechanisms creates a restrictive environment that might not be in line with national strategies for promoting the digital economy, innovation and entrepreneurship, with the resulting waste of economic resources: while on the one hand these strategies are being promoted, on the other they are being discouraged by unintentionally blocking legitimate initiatives.
Blocking or filtering Internet content can be inadmissible given that, in addition to directly and indirectly affecting the use of and/or access to the Internet, such measures do not provide an actual solution to the problem they are attempting to solve.
Blocking or filtering does not guarantee that the content or any other type of illegal activity will be eliminated. On the contrary, the content and its uses may be replicated on different platforms and the measure may be “bypassed” using techniques or tools available on the Internet.
We believe that policy makers in our countries should take into consideration the open and interoperable nature of the Internet when considering measures of this magnitude, and that they should make sure that their actions are in line with their national digital strategies.
The Latin American and Caribbean Internet Addresses Registry (LACNIC) considers that any measure that seeks to regulate the use of the Internet should consider broad criteria that promote innovation and the development of new uses and services for the network, criteria that should be defined using a multistakeholder model, not by the direct application of intrusive regulations that will result in its restriction.
- Internet Society Perspectives on Internet Content Blocking: An Overview (https://www.internetsociety.org/es/resources/doc/2017/internet-content-blocking/)
- RFC 7754 – Technical Considerations for Internet Service Blocking and Filtering (https://tools.ietf.org/html/rfc7754 )