Commentary Sophos Security Insights
The More Things Change
Today's malware is more complex than ever, yet it's still based on three basic hacks
I've been working in the tech field for a bit over 15 years now. It's amazing to see how the industry and the technology has changed during that time. In my first job out of grad school, I taught corporate employees how to use Netscape Communicator on their Pentium II desktop PCs, which had just been migrated to Windows NT 4.0. I was the only person in my family with a cell phone -- a large, clunky object with a telescoping antenna and a belt holster. And IBM had just announced a new hard drive for notebook PCs that broke barriers by providing 6.4 GB of storage, three times the average in those days.
Malware has changed a lot, too, since 1998. Back then, floppy disks were just giving way to email as the infection vector of choice. Money was rarely a motivating factor for malware authors and distributors. The stereotype of the nerdy, young man wreaking havoc from his mother's basement probably wasn't too far off in many cases. The volume of new malware was so low that security experts could name and analyze each new sample.
Things look very different now, of course. Our multicore, always-connected devices can be infected via the network, email, SMS, USB, or the Web. Social engineering has advanced well beyond asking users to open a picture of Anna Kournikova naked (though variations of that trick still work). Behind the scenes, malware has "matured from a cottage industry to a Henry Ford style production line funded by organised crime," as my colleague Peter Szabo put it. Analyzing every piece of malware now would be impossible, with several new samples arriving each second of the day. Simon Reed, who runs SophosLabs, describes his team's work as a big data processing and mining operation.
With all the new technology and the rapid growth of "mass market" cybercrime, it may be easy to overlook one constant: Malware depends on finding a way to install or run on its target without the user's informed consent. And, in 15 years in the industry, I've only seen three fundamental ways for that to happen: exploiting a vulnerability, compromising user credentials, and/or tricking the user. That's it. An entire generation's worth of malware -- tens of millions of variants -- reduced to three simple hacks.
Fortunately, as security professionals, we already know how to defend against these three hacks, even if we don't always give them the attention they deserve. We stop exploits by building or buying more secure software, patching vulnerabilities as they arise, and implementing configurations that balance usability and security. We protect user credentials by implementing multifactor authentication, encouraging or enforcing the creation of strong and unique passwords, and securing the credentials in transit and at rest. Users are human, so they'll always be fallible, but security awareness and education -- emphasizing the why, not just the how -- can go a long way in reducing susceptibility to social engineering.
It's easy to describe these defenses, but implementing them properly, consistently, and completely is much harder. Security products help by providing visibility and by leveraging automation and vendors' expertise. They also fill the inevitable gaps in an organization's defenses, detecting threats that slip through. As such, security tools have had to evolve as the threats have evolved. Firewalls have given way to UTMs, antivirus software has developed into multilayer endpoint protection, and Web and email filters have helped users make fewer and better decisions about what to download or open. They may not be perfect, but they're better than their predecessors, and they're a heck of a lot better than nothing.
So, yes, a lot has changed in 15 years. All things considered, I'd rather not go back to my Pentium II and my dumb brick of a cell phone, even if security was simpler then. But going back to the basics of malware protection, with a little help from today's technology? Well, that doesn't seem like a bad idea at all.