In the weeks leading up to and following Zimbabwe’s disputed 2013 election, Zimbabweans were hit by significant Internet-based attacks. Because the incident was not widely reported, it did not gain traction at all in the Internet Freedom Community. Yet the incident was one of a kind to be documented during an African election. It adversely affected Zimbabweans’ rights to stay informed including by accessing first-hand information on the elections to inform civic action and response to the election irregularities.
This also had repercussions on the transparency and outcome of the election since those who were monitoring them on online platforms were deprived of necessary information to base their reports on. As part of the project Sub Saharan Africa Cyber Threat Modelling, I propose that as Zimbabwe prepares for the 2018 elections, civil society actors in Zimbabwe and those who support their digital security and integrity projects should use the 2013 incident to undertake a proper threat model that takes account of DDOS attacks.
This will coincide with the Zimbabwe CSOs’ launch of the2018 Election Situation Room on 27 June 2018- an initiative that seeks to coordinate their activities & enhance citizen monitoring & participation in electoral processes. Unlike other attack vectors that only affect information confidentiality and integrity, a DDOS goes after the availability of a system or a network. The nature of its attack is like having your home flooded- without warning; attackers can upend the availability of information during an election. When it hits a network, a long time can pass before detection and mitigation. In an ever-expanding field of adversaries and other attack vectors, DDOS is still often difficult to attribute as it can often be orchestrated remotely.
Around July 30, 2013, while working for the Zimbabwe Human Rights Forum, I woke up to realise that most of the real-time content of the website I managed had been compromised through deliberate defacement and selective data erasure. As I tried to locate the content, the site went offline. I fiddled with the network until a U.S. Congress Researcher, who had been following our blogs, alerted me to the DDOS attack directed at our web host Greennet and web hosts of other critical websites such as Electionride.com and Nehanda Radio.
The incident included two massive distributed denial of service (DDoS) on Greennet to disrupt the Forum’s activities, which in turn caused collateral damage to other sites like that of Privacy International. Despite the difficulty of the attack source attribution, experts believed that either a government entity or a private organisation was responsible for the attack given both its nature and magnitude: 100Gbps attack that used DNS reflection rather than an unsophisticated botnet to attempt to overwhelm its servers.
What is a DDOS attack?
Confidentiality, integrity, and availability are the fundamentals of information assurance. Organisations often rely on the so-called CIA (Confidentiality, Integrity, and Availability) triad to benchmark and evaluate their information security. For instance, the data defacement and erasure on the web pages of the Zimbabwe Human Rights Forum affected the integrity of the data and therefore its reliability. However, a DDoS does not go after the confidentiality or integrity of the CIA model. It’s meant to go after the A, the availability of a system or a network.
A Distributed Denial of Service (DDoS) attack is an attempt made to take a website or online service offline. Attackers use a variety of ways to do this, but they all are designed to overwhelm the site with traffic from multiple sources.
In a DDoS attack, the traffic flooding the site can come from hundreds or thousands of sources, which makes it near-impossible to stop the attack simply by blocking a single IP address. They can be distributed by infected computers via botnets or coordinated. Sites also struggle to differentiate between a legitimate user and attack traffic.
A DDoS attack differs from a Denial of Service (DoS) attack, which typically uses a single computer and connection to flood a system or site.
Zimbabwe experienced a Domain Name System (DNS) reflection attack. This kind of attack spoofs the target’s IP address(es) in DNS requests, causing DNS servers to amplify the volume of data focused on the data centre under attack.
Unlike a malware in the class of worms, a DDOS could generally be classified in the virus category in its mode of attack. Like a DDOS, a virus generally refers to a malicious program that self-replicates but requires some user interaction to be initiated. In this case, the virus/bot has a malicious payload (instruction) that it is meant to execute.
Here is an example by my friend Jonathan Weismann at Rochester Institute of Technology:
If Harry the hacker sends ten, one hundred or even one thousand pictures to an important web server, nothing will happen.
However, if Harry the hacker puts a program on ten thousand user machines and they each are instructed to place programs on thousands of other machines, when the time comes, Harry the hacker will give the kill signal and all machines known as zombies in this botnet, robot network, will be sending traffic to a poor victim’s server that will come to a grinding halt.
Attribution challenge and Recurrence
Cyber-attacks similar to the Zimbabwean one are difficult to attribute to any particular adversary unless such adversaries leave forensic footprints for. In this case, its sophistication and nature seemed to suggest it was done by a nation-state or a well-resourced corporate adversary. We harvested clues that located the attackers in China, for instance, our mailing system geolocation analytics leading to the attack showed an unusual increased Chinese subscription and readership of our newsletter. Also, our mailing system- ‘MailChimp’ system was directly linked to the website meaning the newsletter readers also accessed web content.
We cannot predict recurrence during the 2018 election or in future with any degree of certainty because information controls are often applied in highly dynamic ways often responding to events on the ground displaying wide-ranging motives.
There has been an accelerated, dynamic and complex pace of events in Zimbabwe since the November 2017 power transfer. The country’s diversified international business partners potentially open up and diversify the vendors in the market for computer espionage and surveillance in addition to the so-called Huawei problem. Whereas China, also a major investor in Zimbabwe, continues to top the charts with its nation-sponsored attack activity, aspects of lesser-known nation-states and benign entities give cause for concern as they can hide in the darker parts of the internet. A good example was the hacking into the Zimbabwe Government websites. The attack vectors are expanding to include the use of social media to influence the opinions and actions of large populations.
The Zimbabwe case study and other recent attacks such as on the DNS Company Dyn shed a few lessons.
DDOS attacks happen very fast, and hard to detect yet their consequence can be devastating. There can be a long time lapse between an attack, detection and mitigation. One needs a faster, more immediate means of threat detection to prevent severe damage. There’s little an organisation can do to prevent threats which may be the result of larger geopolitical forces but one can substantially reduce the likelihood of the adversaries’ chances to succeed by reducing own vulnerability, and in turn, own risk. This may include taking technical measures but also a holistic approach. For example, albeit on a different subject, Citizen Lab Research on targeted malware attacks reveal that the technical sophistication of [attacks] may be fairly low, with more effort placed on social engineering.
In our case the following non-prescriptive steps could have helped mitigate the impact of the DDOS attack:
· Web content back up, including hosting a blog hosted on a separate platform where we could re-direct our readers.
· Improving our firewall and password combinations as it appears the adversary gained entry onto our website dashboard to wipe out content.
· Closely paying attention to the tell-tale signs such as the increase in the number of China-based subscribers.
· Establishing a good relationship with the web-host and sharing concerns during key political events to enable their technical team to be prepared.
· Draft an organisational DDOS attack playbook. This document sets out the systematic procedure to be followed in case of a DDOS attack. It helps ensure that organisational staff responds to the attack in an organised manner.