Entries Tagged "cloud computing"

Page 1 of 8

Vulnerability Finding Using Machine Learning

Microsoft is training a machine-learning system to find software bugs:

At Microsoft, 47,000 developers generate nearly 30 thousand bugs a month. These items get stored across over 100 AzureDevOps and GitHub repositories. To better label and prioritize bugs at that scale, we couldn’t just apply more people to the problem. However, large volumes of semi-curated data are perfect for machine learning. Since 2001 Microsoft has collected 13 million work items and bugs. We used that data to develop a process and machine learning model that correctly distinguishes between security and non-security bugs 99 percent of the time and accurately identifies the critical, high priority security bugs, 97 percent of the time.

News article.

I wrote about this in 2018:

The problem of finding software vulnerabilities seems well-suited for ML systems. Going through code line by line is just the sort of tedious problem that computers excel at, if we can only teach them what a vulnerability looks like. There are challenges with that, of course, but there is already a healthy amount of academic literature on the topic — and research is continuing. There’s every reason to expect ML systems to get better at this as time goes on, and some reason to expect them to eventually become very good at it.

Finding vulnerabilities can benefit both attackers and defenders, but it’s not a fair fight. When an attacker’s ML system finds a vulnerability in software, the attacker can use it to compromise systems. When a defender’s ML system finds the same vulnerability, he or she can try to patch the system or program network defenses to watch for and block code that tries to exploit it.

But when the same system is in the hands of a software developer who uses it to find the vulnerability before the software is ever released, the developer fixes it so it can never be used in the first place. The ML system will probably be part of his or her software design tools and will automatically find and fix vulnerabilities while the code is still in development.

Fast-forward a decade or so into the future. We might say to each other, “Remember those years when software vulnerabilities were a thing, before ML vulnerability finders were built into every compiler and fixed them before the software was ever released? Wow, those were crazy years.” Not only is this future possible, but I would bet on it.

Getting from here to there will be a dangerous ride, though. Those vulnerability finders will first be unleashed on existing software, giving attackers hundreds if not thousands of vulnerabilities to exploit in real-world attacks. Sure, defenders can use the same systems, but many of today’s Internet of Things (IoT) systems have no engineering teams to write patches and no ability to download and install patches. The result will be hundreds of vulnerabilities that attackers can find and use.

Posted on April 20, 2020 at 6:22 AMView Comments

Security in 2020: Revisited

Ten years ago, I wrote an essay: “Security in 2020.” Well, it’s finally 2020. I think I did pretty well. Here’s what I said back then:

There’s really no such thing as security in the abstract. Security can only be defined in relation to something else. You’re secure from something or against something. In the next 10 years, the traditional definition of IT security — that it protects you from hackers, criminals, and other bad guys — will undergo a radical shift. Instead of protecting you from the bad guys, it will increasingly protect businesses and their business models from you.

Ten years ago, the big conceptual change in IT security was deperimeterization. A wordlike grouping of 18 letters with both a prefix and a suffix, it has to be the ugliest word our industry invented. The concept, though — the dissolution of the strict boundaries between the internal and external network — was both real and important.

There’s more deperimeterization today than there ever was. Customer and partner access, guest access, outsourced e-mail, VPNs; to the extent there is an organizational network boundary, it’s so full of holes that it’s sometimes easier to pretend it isn’t there. The most important change, though, is conceptual. We used to think of a network as a fortress, with the good guys on the inside and the bad guys on the outside, and walls and gates and guards to ensure that only the good guys got inside. Modern networks are more like cities, dynamic and complex entities with many different boundaries within them. The access, authorization, and trust relationships are even more complicated.

Today, two other conceptual changes matter. The first is consumerization. Another ponderous invented word, it’s the idea that consumers get the cool new gadgets first, and demand to do their work on them. Employees already have their laptops configured just the way they like them, and they don’t want another one just for getting through the corporate VPN. They’re already reading their mail on their BlackBerrys or iPads. They already have a home computer, and it’s cooler than the standard issue IT department machine. Network administrators are increasingly losing control over clients.

This trend will only increase. Consumer devices will become trendier, cheaper, and more integrated; and younger people are already used to using their own stuff on their school networks. It’s a recapitulation of the PC revolution. The centralized computer center concept was shaken by people buying PCs to run VisiCalc; now it’s iPads and Android smartphones.

The second conceptual change comes from cloud computing: our increasing tendency to store our data elsewhere. Call it decentralization: our email, photos, books, music, and documents are stored somewhere, and accessible to us through our consumer devices. The younger you are, the more you expect to get your digital stuff on the closest screen available. This is an important trend, because it signals the end of the hardware and operating system battles we’ve all lived with. Windows vs. Mac doesn’t matter when all you need is a web browser. Computers become temporary; user backup becomes irrelevant. It’s all out there somewhere — and users are increasingly losing control over their data.

During the next 10 years, three new conceptual changes will emerge, two of which we can already see the beginnings of. The first I’ll call deconcentration. The general-purpose computer is dying and being replaced by special-purpose devices. Some of them, like the iPhone, seem general purpose but are strictly controlled by their providers. Others, like Internet-enabled game machines or digital cameras, are truly special purpose. In 10 years, most computers will be small, specialized, and ubiquitous.

Even on what are ostensibly general-purpose devices, we’re seeing more special-purpose applications. Sure, you could use the iPhone’s web browser to access the New York Times website, but it’s much easier to use the NYT’s special iPhone app. As computers become smaller and cheaper, this trend will only continue. It’ll be easier to use special-purpose hardware and software. And companies, wanting more control over their users’ experience, will push this trend.

The second is decustomerization — now I get to invent the really ugly words — the idea that we get more of our IT functionality without any business relation­ship. We’re all part of this trend: every search engine gives away its services in exchange for the ability to advertise. It’s not just Google and Bing; most webmail and social networking sites offer free basic service in exchange for advertising, possibly with premium services for money. Most websites, even useful ones that take the place of client software, are free; they are either run altruistically or to facilitate advertising.

Soon it will be hardware. In 1999, Internet startup FreePC tried to make money by giving away computers in exchange for the ability to monitor users’ surfing and purchasing habits. The company failed, but computers have only gotten cheaper since then. It won’t be long before giving away netbooks in exchange for advertising will be a viable business. Or giving away digital cameras. Already there are companies that give away long-distance minutes in exchange for advertising. Free cell phones aren’t far off. Of course, not all IT hardware will be free. Some of the new cool hardware will cost too much to be free, and there will always be a need for concentrated computing power close to the user — game systems are an obvious example — but those will be the exception. Where the hardware costs too much to just give away, however, we’ll see free or highly subsidized hardware in exchange for locked-in service; that’s already the way cell phones are sold.

This is important because it destroys what’s left of the normal business rela­tionship between IT companies and their users. We’re not Google’s customers; we’re Google’s product that they sell to their customers. It’s a three-way relation­ship: us, the IT service provider, and the advertiser or data buyer. And as these noncustomer IT relationships proliferate, we’ll see more IT companies treating us as products. If I buy a Dell computer, then I’m obviously a Dell customer; but if I get a Dell computer for free in exchange for access to my life, it’s much less obvious whom I’m entering a business relationship with. Facebook’s continual ratcheting down of user privacy in order to satisfy its actual customers­–the advertisers–and enhance its revenue is just a hint of what’s to come.

The third conceptual change I’ve termed depersonization: computing that removes the user, either partially or entirely. Expect to see more software agents: programs that do things on your behalf, such as prioritize your email based on your observed preferences or send you personalized sales announcements based on your past behavior. The “people who liked this also liked” feature on many retail websites is just the beginning. A website that alerts you if a plane ticket to your favorite destination drops below a certain price is simplistic but useful, and some sites already offer this functionality. Ten years won’t be enough time to solve the serious artificial intelligence problems required to fully real­ize intelligent agents, but the agents of that time will be both sophisticated and commonplace, and they’ll need less direct input from you.

Similarly, connecting objects to the Internet will soon be cheap enough to be viable. There’s already considerable research into Internet-enabled medical devices, smart power grids that communicate with smart phones, and networked automobiles. Nike sneakers can already communicate with your iPhone. Your phone already tells the network where you are. Internet-enabled appliances are already in limited use, but soon they will be the norm. Businesses will acquire smart HVAC units, smart elevators, and smart inventory systems. And, as short-range communications — like RFID and Bluetooth — become cheaper, everything becomes smart.

The “Internet of things” won’t need you to communicate. The smart appliances in your smart home will talk directly to the power company. Your smart car will talk to road sensors and, eventually, other cars. Your clothes will talk to your dry cleaner. Your phone will talk to vending machines; they already do in some countries. The ramifications of this are hard to imagine; it’s likely to be weirder and less orderly than the contemporary press describes it. But certainly smart objects will be talking about you, and you probably won’t have much control over what they’re saying.

One old trend: deperimeterization. Two current trends: consumerization and decentralization. Three future trends: deconcentration, decustomerization, and depersonization. That’s IT in 2020 — it’s not under your control, it’s doing things without your knowledge and consent, and it’s not necessarily acting in your best interests. And this is how things will be when they’re working as they’re intended to work; I haven’t even started talking about the bad guys yet.

That’s because IT security in 2020 will be less about protecting you from traditional bad guys, and more about protecting corporate business models from you. Deperimeterization assumes everyone is untrusted until proven otherwise. Consumerization requires networks to assume all user devices are untrustworthy until proven otherwise. Decentralization and deconcentration won’t work if you’re able to hack the devices to run unauthorized software or access unauthorized data. Deconsumerization won’t be viable unless you’re unable to bypass the ads, or whatever the vendor uses to monetize you. And depersonization requires the autonomous devices to be, well, autonomous.

In 2020 — 10 years from now — Moore’s Law predicts that computers will be 100 times more powerful. That’ll change things in ways we can’t know, but we do know that human nature never changes. Cory Doctorow rightly pointed out that all complex ecosystems have parasites. Society’s traditional parasites are criminals, but a broader definition makes more sense here. As we users lose control of those systems and IT providers gain control for their own purposes, the definition of “parasite” will shift. Whether they’re criminals trying to drain your bank account, movie watchers trying to bypass whatever copy protection studios are using to protect their profits, or Facebook users trying to use the service without giving up their privacy or being forced to watch ads, parasites will continue to try to take advantage of IT systems. They’ll exist, just as they always have existed, and — like today — security is going to have a hard time keeping up with them.

Welcome to the future. Companies will use technical security measures, backed up by legal security measures, to protect their business models. And unless you’re a model user, the parasite will be you.

My only real complaint with the essay is that I used “decentralization” in a nonstandard manner, and didn’t explain it well. I meant that our personal data will become decentralized; instead of it all being on our own computers, it will be on the computers of various cloud providers. But that causes a massive centralization of all of our data. I should have explicitly called out the risks of that.

Otherwise, I’m happy with what I wrote ten years ago.

Posted on February 7, 2020 at 12:50 PMView Comments

Apple Abandoned Plans for Encrypted iCloud Backup after FBI Complained

This is new from Reuters:

More than two years ago, Apple told the FBI that it planned to offer users end-to-end encryption when storing their phone data on iCloud, according to one current and three former FBI officials and one current and one former Apple employee.

Under that plan, primarily designed to thwart hackers, Apple would no longer have a key to unlock the encrypted data, meaning it would not be able to turn material over to authorities in a readable form even under court order.

In private talks with Apple soon after, representatives of the FBI’s cyber crime agents and its operational technology division objected to the plan, arguing it would deny them the most effective means for gaining evidence against iPhone-using suspects, the government sources said.

When Apple spoke privately to the FBI about its work on phone security the following year, the end-to-end encryption plan had been dropped, according to the six sources. Reuters could not determine why exactly Apple dropped the plan.

EDITED TO ADD (2/13): Android has enrypted backups.

Posted on January 23, 2020 at 6:10 AMView Comments

SpiderOak's Warrant Canary Died

BoingBoing has the story.

I have never quite trusted the idea of a warrant canary. But here it seems to have worked. (Presumably, if SpiderOak wanted to replace the warrant canary with a transparency report, they would have written something explaining their decision. To have it simply disappear is what we would expect if SpiderOak were being forced to comply with a US government request for personal data.)

EDITED TO ADD (8/9): SpiderOak has posted an explanation claiming that the warrant canary did not die — it just changed.

That’s obviously false, because it did die. And a change is the functional equivalent — that’s how they work. So either they have received a National Security Letter and now have to pretend they did not, or they completely misunderstood what a warrant canary is and how it works. No one knows.

I have never fully trusted warrant canaries — this EFF post explains why — and this is an illustration.

Posted on August 8, 2018 at 9:37 AMView Comments

Russian Censorship of Telegram

Internet censors have a new strategy in their bid to block applications and websites: pressuring the large cloud providers that host them. These providers have concerns that are much broader than the targets of censorship efforts, so they have the choice of either standing up to the censors or capitulating in order to maximize their business. Today’s Internet largely reflects the dominance of a handful of companies behind the cloud services, search engines and mobile platforms that underpin the technology landscape. This new centralization radically tips the balance between those who want to censor parts of the Internet and those trying to evade censorship. When the profitable answer is for a software giant to acquiesce to censors’ demands, how long can Internet freedom last?

The recent battle between the Russian government and the Telegram messaging app illustrates one way this might play out. Russia has been trying to block Telegram since April, when a Moscow court banned it after the company refused to give Russian authorities access to user messages. Telegram, which is widely used in Russia, works on both iPhone and Android, and there are Windows and Mac desktop versions available. The app offers optional end-to-end encryption, meaning that all messages are encrypted on the sender’s phone and decrypted on the receiver’s phone; no part of the network can eavesdrop on the messages.

Since then, Telegram has been playing cat-and-mouse with the Russian telecom regulator Roskomnadzor by varying the IP address the app uses to communicate. Because Telegram isn’t a fixed website, it doesn’t need a fixed IP address. Telegram bought tens of thousands of IP addresses and has been quickly rotating through them, staying a step ahead of censors. Cleverly, this tactic is invisible to users. The app never sees the change, or the entire list of IP addresses, and the censor has no clear way to block them all.

A week after the court ban, Roskomnadzor countered with an unprecedented move of its own: blocking 19 million IP addresses, many on Amazon Web Services and Google Cloud. The collateral damage was widespread: The action inadvertently broke many other web services that use those platforms, and Roskomnadzor scaled back after it became clear that its action had affected services critical for Russian business. Even so, the censor is still blocking millions of IP addresses.

More recently, Russia has been pressuring Apple not to offer the Telegram app in its iPhone App Store. As of this writing, Apple has not complied, and the company has allowed Telegram to download a critical software update to iPhone users (after what the app’s founder called a delay last month). Roskomnadzor could further pressure Apple, though, including by threatening to turn off its entire iPhone app business in Russia.

Telegram might seem a weird app for Russia to focus on. Those of us who work in security don’t recommend the program, primarily because of the nature of its cryptographic protocols. In general, proprietary cryptography has numerous fatal security flaws. We generally recommend Signal for secure SMS messaging, or, if having that program on your computer is somehow incriminating, WhatsApp. (More than 1.5 billion people worldwide use WhatsApp.) What Telegram has going for it is that it works really well on lousy networks. That’s why it is so popular in places like Iran and Afghanistan. (Iran is also trying to ban the app.)

What the Russian government doesn’t like about Telegram is its anonymous broadcast feature­ — channel capability and chats — ­which makes it an effective platform for political debate and citizen journalism. The Russians might not like that Telegram is encrypted, but odds are good that they can simply break the encryption. Telegram’s role in facilitating uncontrolled journalism is the real issue.

Iran attempts to block Telegram have been more successful than Russia’s, less because Iran’s censorship technology is more sophisticated but because Telegram is not willing to go as far to defend Iranian users. The reasons are not rooted in business decisions. Simply put, Telegram is a Russian product and the designers are more motivated to poke Russia in the eye. Pavel Durov, Telegram’s founder, has pledged millions of dollars to help fight Russian censorship.

For the moment, Russia has lost. But this battle is far from over. Russia could easily come back with more targeted pressure on Google, Amazon and Apple. A year earlier, Zello used the same trick Telegram is using to evade Russian censors. Then, Roskomnadzor threatened to block all of Amazon Web Services and Google Cloud; and in that instance, both companies forced Zello to stop its IP-hopping censorship-evasion tactic.

Russia could also further develop its censorship infrastructure. If its capabilities were as finely honed as China’s, it would be able to more effectively block Telegram from operating. Right now, Russia can block only specific IP addresses, which is too coarse a tool for this issue. Telegram’s voice capabilities in Russia are significantly degraded, however, probably because high-capacity IP addresses are easier to block.

Whatever its current frustrations, Russia might well win in the long term. By demonstrating its willingness to suffer the temporary collateral damage of blocking major cloud providers, it prompted cloud providers to block another and more effective anti-censorship tactic, or at least accelerated the process. In April, Google and Amazon banned­ — and technically blocked­ — the practice of “domain fronting,” a trick anti-censorship tools use to get around Internet censors by pretending to be other kinds of traffic. Developers would use popular websites as a proxy, routing traffic to their own servers through another website­ — in this case Google.com­ — to fool censors into believing the traffic was intended for Google.com. The anonymous web-browsing tool Tor has used domain fronting since 2014. Signal, since 2016. Eliminating the capability is a boon to censors worldwide.

Tech giants have gotten embroiled in censorship battles for years. Sometimes they fight and sometimes they fold, but until now there have always been options. What this particular fight highlights is that Internet freedom is increasingly in the hands of the world’s largest Internet companies. And while freedom may have its advocates — ­the American Civil Liberties Union has tweeted its support for those companies, and some 12,000 people in Moscow protested against the Telegram ban­ — actions such as disallowing domain fronting illustrate that getting the big tech companies to sacrifice their near-term commercial interests will be an uphill battle. Apple has already removed anti-censorship apps from its Chinese app store.

In 1993, John Gilmore famously said that “The Internet interprets censorship as damage and routes around it.” That was technically true when he said it but only because the routing structure of the Internet was so distributed. As centralization increases, the Internet loses that robustness, and censorship by governments and companies becomes easier.

This essay previously appeared on Lawfare.com.

Posted on June 13, 2018 at 6:55 AMView Comments

The Effects of the Spectre and Meltdown Vulnerabilities

On January 3, the world learned about a series of major security vulnerabilities in modern microprocessors. Called Spectre and Meltdown, these vulnerabilities were discovered by several different researchers last summer, disclosed to the microprocessors’ manufacturers, and patched­ — at least to the extent possible.

This news isn’t really any different from the usual endless stream of security vulnerabilities and patches, but it’s also a harbinger of the sorts of security problems we’re going to be seeing in the coming years. These are vulnerabilities in computer hardware, not software. They affect virtually all high-end microprocessors produced in the last 20 years. Patching them requires large-scale coordination across the industry, and in some cases drastically affects the performance of the computers. And sometimes patching isn’t possible; the vulnerability will remain until the computer is discarded.

Spectre and Meltdown aren’t anomalies. They represent a new area to look for vulnerabilities and a new avenue of attack. They’re the future of security­ — and it doesn’t look good for the defenders.

Modern computers do lots of things at the same time. Your computer and your phone simultaneously run several applications — ­or apps. Your browser has several windows open. A cloud computer runs applications for many different computers. All of those applications need to be isolated from each other. For security, one application isn’t supposed to be able to peek at what another one is doing, except in very controlled circumstances. Otherwise, a malicious advertisement on a website you’re visiting could eavesdrop on your banking details, or the cloud service purchased by some foreign intelligence organization could eavesdrop on every other cloud customer, and so on. The companies that write browsers, operating systems, and cloud infrastructure spend a lot of time making sure this isolation works.

Both Spectre and Meltdown break that isolation, deep down at the microprocessor level, by exploiting performance optimizations that have been implemented for the past decade or so. Basically, microprocessors have become so fast that they spend a lot of time waiting for data to move in and out of memory. To increase performance, these processors guess what data they’re going to receive and execute instructions based on that. If the guess turns out to be correct, it’s a performance win. If it’s wrong, the microprocessors throw away what they’ve done without losing any time. This feature is called speculative execution.

Spectre and Meltdown attack speculative execution in different ways. Meltdown is more of a conventional vulnerability; the designers of the speculative-execution process made a mistake, so they just needed to fix it. Spectre is worse; it’s a flaw in the very concept of speculative execution. There’s no way to patch that vulnerability; the chips need to be redesigned in such a way as to eliminate it.

Since the announcement, manufacturers have been rolling out patches to these vulnerabilities to the extent possible. Operating systems have been patched so that attackers can’t make use of the vulnerabilities. Web browsers have been patched. Chips have been patched. From the user’s perspective, these are routine fixes. But several aspects of these vulnerabilities illustrate the sorts of security problems we’re only going to be seeing more of.

First, attacks against hardware, as opposed to software, will become more common. Last fall, vulnerabilities were discovered in Intel’s Management Engine, a remote-administration feature on its microprocessors. Like Spectre and Meltdown, they affected how the chips operate. Looking for vulnerabilities on computer chips is new. Now that researchers know this is a fruitful area to explore, security researchers, foreign intelligence agencies, and criminals will be on the hunt.

Second, because microprocessors are fundamental parts of computers, patching requires coordination between many companies. Even when manufacturers like Intel and AMD can write a patch for a vulnerability, computer makers and application vendors still have to customize and push the patch out to the users. This makes it much harder to keep vulnerabilities secret while patches are being written. Spectre and Meltdown were announced prematurely because details were leaking and rumors were swirling. Situations like this give malicious actors more opportunity to attack systems before they’re guarded.

Third, these vulnerabilities will affect computers’ functionality. In some cases, the patches for Spectre and Meltdown result in significant reductions in speed. The press initially reported 30%, but that only seems true for certain servers running in the cloud. For your personal computer or phone, the performance hit from the patch is minimal. But as more vulnerabilities are discovered in hardware, patches will affect performance in noticeable ways.

And then there are the unpatchable vulnerabilities. For decades, the computer industry has kept things secure by finding vulnerabilities in fielded products and quickly patching them. Now there are cases where that doesn’t work. Sometimes it’s because computers are in cheap products that don’t have a patch mechanism, like many of the DVRs and webcams that are vulnerable to the Mirai (and other) botnets — ­groups of Internet-connected devices sabotaged for coordinated digital attacks. Sometimes it’s because a computer chip’s functionality is so core to a computer’s design that patching it effectively means turning the computer off. This, too, is becoming more common.

Increasingly, everything is a computer: not just your laptop and phone, but your car, your appliances, your medical devices, and global infrastructure. These computers are and always will be vulnerable, but Spectre and Meltdown represent a new class of vulnerability. Unpatchable vulnerabilities in the deepest recesses of the world’s computer hardware is the new normal. It’s going to leave us all much more vulnerable in the future.

This essay previously appeared on TheAtlantic.com.

Posted on January 26, 2018 at 6:12 AMView Comments

Yet Another FBI Proposal for Insecure Communications

Deputy Attorney General Rosenstein has given talks where he proposes that tech companies decrease their communications and device security for the benefit of the FBI. In a recent talk, his idea is that tech companies just save a copy of the plaintext:

Law enforcement can also partner with private industry to address a problem we call “Going Dark.” Technology increasingly frustrates traditional law enforcement efforts to collect evidence needed to protect public safety and solve crime. For example, many instant-messaging services now encrypt messages by default. The prevent the police from reading those messages, even if an impartial judge approves their interception.

The problem is especially critical because electronic evidence is necessary for both the investigation of a cyber incident and the prosecution of the perpetrator. If we cannot access data even with lawful process, we are unable to do our job. Our ability to secure systems and prosecute criminals depends on our ability to gather evidence.

I encourage you to carefully consider your company’s interests and how you can work cooperatively with us. Although encryption can help secure your data, it may also prevent law enforcement agencies from protecting your data.

Encryption serves a valuable purpose. It is a foundational element of data security and essential to safeguarding data against cyber-attacks. It is critical to the growth and flourishing of the digital economy, and we support it. I support strong and responsible encryption.

I simply maintain that companies should retain the capability to provide the government unencrypted copies of communications and data stored on devices, when a court orders them to do so.

Responsible encryption is effective secure encryption, coupled with access capabilities. We know encryption can include safeguards. For example, there are systems that include central management of security keys and operating system updates; scanning of content, like your e-mails, for advertising purposes; simulcast of messages to multiple destinations at once; and key recovery when a user forgets the password to decrypt a laptop. No one calls any of those functions a “backdoor.” In fact, those very capabilities are marketed and sought out.

I do not believe that the government should mandate a specific means of ensuring access. The government does not need to micromanage the engineering.

The question is whether to require a particular goal: When a court issues a search warrant or wiretap order to collect evidence of crime, the company should be able to help. The government does not need to hold the key.

Rosenstein is right that many services like Gmail naturally keep plaintext in the cloud. This is something we pointed out in our 2016 paper: “Don’t Panic.” But forcing companies to build an alternate means to access the plaintext that the user can’t control is an enormous vulnerability.

Posted on January 11, 2018 at 7:05 AMView Comments

Spectre and Meltdown Attacks against Microprocessors

The security of pretty much every computer on the planet has just gotten a lot worse, and the only real solution — which of course is not a solution — is to throw them all away and buy new ones.

On Wednesday, researchers just announced a series of major security vulnerabilities in the microprocessors at the heart of the world’s computers for the past 15-20 years. They’ve been named Spectre and Meltdown, and they have to do with manipulating different ways processors optimize performance by rearranging the order of instructions or performing different instructions in parallel. An attacker who controls one process on a system can use the vulnerabilities to steal secrets elsewhere on the computer. (The research papers are here and here.)

This means that a malicious app on your phone could steal data from your other apps. Or a malicious program on your computer — maybe one running in a browser window from that sketchy site you’re visiting, or as a result of a phishing attack — can steal data elsewhere on your machine. Cloud services, which often share machines amongst several customers, are especially vulnerable. This affects corporate applications running on cloud infrastructure, and end-user cloud applications like Google Drive. Someone can run a process in the cloud and steal data from every other user on the same hardware.

Information about these flaws has been secretly circulating amongst the major IT companies for months as they researched the ramifications and coordinated updates. The details were supposed to be released next week, but the story broke early and everyone is scrambling. By now all the major cloud vendors have patched their systems against the vulnerabilities that can be patched against.

“Throw it away and buy a new one” is ridiculous security advice, but it’s what US-CERT recommends. It is also unworkable. The problem is that there isn’t anything to buy that isn’t vulnerable. Pretty much every major processor made in the past 20 years is vulnerable to some flavor of these vulnerabilities. Patching against Meltdown can degrade performance by almost a third. And there’s no patch for Spectre; the microprocessors have to be redesigned to prevent the attack, and that will take years. (Here’s a running list of who’s patched what.)

This is bad, but expect it more and more. Several trends are converging in a way that makes our current system of patching security vulnerabilities harder to implement.

The first is that these vulnerabilities affect embedded computers in consumer devices. Unlike our computers and phones, these systems are designed and produced at a lower profit margin with less engineering expertise. There aren’t security teams on call to write patches, and there often aren’t mechanisms to push patches onto the devices. We’re already seeing this with home routers, digital video recorders, and webcams. The vulnerability that allowed them to be taken over by the Mirai botnet last August simply can’t be fixed.

The second is that some of the patches require updating the computer’s firmware. This is much harder to walk consumers through, and is more likely to permanently brick the device if something goes wrong. It also requires more coordination. In November, Intel released a firmware update to fix a vulnerability in its Management Engine (ME): another flaw in its microprocessors. But it couldn’t get that update directly to users; it had to work with the individual hardware companies, and some of them just weren’t capable of getting the update to their customers.

We’re already seeing this. Some patches require users to disable the computer’s password, which means organizations can’t automate the patch. Some antivirus software blocks the patch, or — worse — crashes the computer. This results in a three-step process: patch your antivirus software, patch your operating system, and then patch the computer’s firmware.

The final reason is the nature of these vulnerabilities themselves. These aren’t normal software vulnerabilities, where a patch fixes the problem and everyone can move on. These vulnerabilities are in the fundamentals of how the microprocessor operates.

It shouldn’t be surprising that microprocessor designers have been building insecure hardware for 20 years. What’s surprising is that it took 20 years to discover it. In their rush to make computers faster, they weren’t thinking about security. They didn’t have the expertise to find these vulnerabilities. And those who did were too busy finding normal software vulnerabilities to examine microprocessors. Security researchers are starting to look more closely at these systems, so expect to hear about more vulnerabilities along these lines.

Spectre and Meltdown are pretty catastrophic vulnerabilities, but they only affect the confidentiality of data. Now that they — and the research into the Intel ME vulnerability — have shown researchers where to look, more is coming — and what they’ll find will be worse than either Spectre or Meltdown. There will be vulnerabilities that will allow attackers to manipulate or delete data across processes, potentially fatal in the computers controlling our cars or implanted medical devices. These will be similarly impossible to fix, and the only strategy will be to throw our devices away and buy new ones.

This isn’t to say you should immediately turn your computers and phones off and not use them for a few years. For the average user, this is just another attack method amongst many. All the major vendors are working on patches and workarounds for the attacks they can mitigate. All the normal security advice still applies: watch for phishing attacks, don’t click on strange e-mail attachments, don’t visit sketchy websites that might run malware on your browser, patch your systems regularly, and generally be careful on the Internet.

You probably won’t notice that performance hit once Meltdown is patched, except maybe in backup programs and networking applications. Embedded systems that do only one task, like your programmable thermostat or the computer in your refrigerator, are unaffected. Small microprocessors that don’t do all of the vulnerable fancy performance tricks are unaffected. Browsers will figure out how to mitigate this in software. Overall, the security of the average Internet-of-Things device is so bad that this attack is in the noise compared to the previously known risks.

It’s a much bigger problem for cloud vendors; the performance hit will be expensive, but I expect that they’ll figure out some clever way of detecting and blocking the attacks. All in all, as bad as Spectre and Meltdown are, I think we got lucky.

But more are coming, and they’ll be worse. 2018 will be the year of microprocessor vulnerabilities, and it’s going to be a wild ride.

Note: A shorter version of this essay previously appeared on CNN.com. My previous blog post on this topic contains additional links.

Posted on January 5, 2018 at 2:22 PMView Comments

1 2 3 8

Sidebar photo of Bruce Schneier by Joe MacInnis.