Automatic updating, if done right, can help eliminate the threat of known security vulnerabilities before attackers start exploiting the flaws. Done wrong, however, the updating process itself becomes an efficient way for attackers to install their code on the victim’s system.
One security researcher has found that at least a hundred programs use an update process that puts their users at risk. How? A computer on the same network as the target machine–think public wireless network–intercepts a message requesting the most recent software update, replies that there is a more recent version available, and then provides malicious code that will be installed through the update process, explains Itzik Kotler, security-operations-center team leader for security firm Radware.
“Every security guru will tell you that you have to patch, have a firewall, and have your antivirus updated,” Kotler says. “However, if [someone] attack[s] the update channel, none of those protections will stop [him] from putting [harmful] code on the system.”
The problem is that many programs use a simple Web request to the software developers’ server, through the hypertext transfer protocol (HTTP), to check for an update. Without encryption, a malicious attacker on the same network can see the request and immediately reply to it, far faster–in Internet time–than a server out on the Web. The attack convinces the software running on the victim’s machine that the attacker’s computer is the legitimate update server, Kotler says.
“I came to the conclusion that the majority of the applications–we have over 100 now–download a file through a simple HTTP request to the vendor Web site,” he says.
The issue affects some major applications, including popular instant messaging and document software, according to Kotler, who asked that the names of the software not be divulged. Among the applications whose update feature does not have the problem: Microsoft’s Office. Microsoft, which has focused on locking down its software since it announced the Trustworthy Computing Initiative in 2002, uses encryption to secure its update requests.
Thinking about the security of the update system is uncommon, Kotler says. Software developers typically believe that sending an unencrypted request through the Internet is secure.
“You can’t say that they have neglected anything or done anything wrong,” he says. “The assumption that the infrastructure is secure is a very natural one for many people.”
While the attacker needs to be on the same network as the victim for the initial infection, after that, the malicious program could use the same technique to infect anyone that checks for updates in the presence of a compromised machine, Kotler says.
“I can basically create an airborne attack,” he says.
The attack can be blunted by making sure that programs do not update on an untrusted network. Security-conscious users should also make sure that all programs notify them when updating.
This new data poisoning tool lets artists fight back against generative AI
The tool, called Nightshade, messes up training data in ways that could cause serious damage to image-generating AI models.
The Biggest Questions: What is death?
New neuroscience is challenging our understanding of the dying process—bringing opportunities for the living.
Rogue superintelligence and merging with machines: Inside the mind of OpenAI’s chief scientist
An exclusive conversation with Ilya Sutskever on his fears for the future of AI and why they’ve made him change the focus of his life’s work.
How to fix the internet
If we want online discourse to improve, we need to move beyond the big platforms.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.