Builder, Breaker, Blather, Why.

BuilderI recently gave a brief talk that noted how Let’s Encrypt and cloud-based architectures encourage positive appsec behaviors. Check out the slides and this blog post for a sense of the main points. Shortly thereafter a slew of security and stability events related to HTTPS and cloud services (SHA-1, Cloudbleed, S3 outage) seemed to undercut this thesis. But perhaps only superficially so. Rather than glibly dismiss these points, let’s examine these events from the perspective of risk and engineering — in other words, how complex systems and software are designed and respond to feedback loops.

This post is a stroll through HTTPS and cloud services, following a path of questions and ideas that builders and breakers might use to evaluate security; leaving the blather of empty pronouncements behind. It’s about the importance of critical thinking and seeking the reasons why a decision comes about.

Eventually Encrypted

For more than a decade at least two major hurdles have blocked pervasive HTTPS: Certs and configuration. The first was (and remains) convincing sites to deploy HTTPS at all, tightly coupled with making deployment HTTPS-only instead of mixed with unencrypted HTTP. The second is getting HTTPS deployments to use strong TLS configurations, e.g. TLS 1.2 with default ciphers that support forward secrecy.

For apps that handle credit cards, PCI has been a crucial driver for adopting strong HTTPS. Having a requirement to use transport encryption, backed by financial consequences for failure, has been more successful than either asking nicely, raising awareness at security conferences, or shaming. As a consequence, I suspect the rate of HTTPS adoption has been far faster for in-scope PCI sites than others.

The SSL Labs project could also be a factor in HTTPS adoption. It distilled a comprehensive analysis of a site’s TLS configuration into a simple letter score. The publically-visible results could be used as a shaming tactic, but that’s a weaker strategy for motivating positive change. The failure of shaming, especially as it relates to HTTPS, is partly demonstrated by the too-typical disregard of browser security warnings. (Which is itself a design challenge, not a user failure.)

Importantly, SSL Labs provides an easy way for organizations to consistently monitor and evaluate their sites. This is a step towards providing help for migration to HTTPS-only sites. App owners still bear the burden of fixing errors and misconfigurations, but this tool made it easier to measure and track their progress towards strong TLS.

Effectively Encrypted

Where SSL Labs inspires behavioral change via metrics, the Let’s Encrypt project empowers behavioral change by addressing fundamental challenges faced by app owners.

Let’s Encrypt eases the resource burden of managing HTTPS endpoints. It removes the initial cost of certs (they’re free!) and reduces the ongoing maintenance cost of deploying, rotating, and handling certs by supporting automation with the ACME protocol. Even so, solving the TLS cert problem is orthogonal to solving the TLS configuration problem. A valid Let’s Encrypt cert might still be deployed to an HTTPS service that gets a bad grade from SSL Labs.

A cert signed with SHA-1, for example, will lower its SSL Labs grade. SHA-1 has been known weak for years and discouraged from use, specifically for digital signatures. Having certs that are both free and easy to rotate (i.e. easy to obtain and deploy new ones) makes it easier for sites to migrate off deprecated versions. The ability to react quickly to change, whether security-related or not, is a sign of a mature organization. Automation as made possible via Let’s Encrypt is a great way to improve that ability.

BreakerThe recent work that demonstrated a SHA-1 collision is commendable, but it shouldn’t be the sudden reason you decided to stop using it. If such proof of danger is your sole deciding factor, you shouldn’t be using (or supporting) Flash or most Android phones.

Facebook explained their trade-offs along the way to hardening their TLS configuration and deprecating SHA-1. It was an engineering-driven security decision that evaluated solutions and chose among conflicting optimizations — all informed by measures of risk. Engineering is the key word in this paragraph; it’s how systems get built. Writing down a simple requirement and prototyping something on a single system with a few dozen users is far removed from delivering a service to hundreds of millions of people. WhatsApp’s crypto design fell into a similar discussion of risk-based engineering. Another example of evaluating risk and threat models is this excellent article on messaging app security and privacy.

Exceptional Events

Companies like Cloudflare take a step beyond SSL Labs and Let’s Encrypt by offering a service to handle both certs and configuration for sites. They pioneered techniques like Keyless SSL  in response to their distinctive threat model of handling private keys for multiple entities.

If you look at the Cloudbleed report and immediately think a service like that should be ditched, it’s important to question the reasoning behind such a risk assessment. Rather than make organizations suffer through the burden of building and maintaining HTTPS, they can have a service the establishes a strong default. Adoption of HTTPS is slow enough, and fraught with error, that services like this make sense for many site owners.

Compare this with heartbleed, which also affected TLS sites, could be more targeted, and exposed private keys (among other sensitive data). The cleanup was long, laborious, and haphazard. Cloudbleed had significant potential exposure, although its discovery and remediation likely led to a lower realized risk than heartbleed.

If you’re saying move away from services like that, what in practice are you saying to move towards? Self-hosted systems in a rack in an electrical closet? Systems that will likely degrade over time and, even more likely, never be upgraded to TLS 1.3? That seems ill-advised.Blather

Does the recent Amazon S3 outage raise concern for cloud-based systems? Not to a great degree. Or, at least, not in a new way. If your site was negatively impacted by the downtime, a good use of that time might have been exploring ways to architect fail-over systems or revisit failure modes and service degradation decisions. Sometimes it’s fine to explicitly accept certain failure modes. That’s what engineering and business do against constraints of resource and budget.

Coherently Considered

So, let’s leave a few exercises for the reader, a few open-ended questions on threat modeling and engineering.

Flash has been long rebuked as both a performance hog and security weakness. Like SHA-1, the infosec community has voiced this warning for years. There have even been one or two (maybe more) demonstrated exploits against it. It persists. It’s embedded in Chrome, which you can interpret as a paternalistic effort to sandbox it or (more cynically) an effort to ensure YouTube videos and ad revenue aren’t impacted by an exodus from the plugin — or perhaps somewhere in between.

Browsers have had impactful vulns, many of which carry significant risk and impact as measured by the annual $50K+ rewards from Pwn2Own competitions. The minuscule number of browser vendors carries risk beyond just vulns, affecting influence on standards and protections for privacy. Yet more browsers doesn’t necessarily equate to better security models within browsers.

On the topic of decentralization, how much is good, how much is bad? DNS recommendations go back and forth. We’ve seen huge DDoS against providers, taking out swaths of sites. We’ll see more. But is shifting DNS the right solution, or a matter that misses the underlying threat or cause of such attacks? How much of IoT is new or different (scale?) compared to the swarms of SQL Slammer and Windows XP botnets of yesteryear’s smaller internet population?

Approaching these with ideas around resilience, isolation, authn/authz models, or feedback loops are (just a few) traits of a builder. As much as they might be for a breaker executing attack models against them.

Approaching these by explaining design flaws and identifying implementation errors are (just a few) traits of a breaker. As much as they might be for a builder designing controls and barriers to disrupt attacks against them.

Approaching these by dismissing complexity, designing systems no one would (or could) use, or highlighting irrelevant flaws is often just blather. Infosec has its share of vacuous or overly-ambiguous phrases like military-grade encryption, perfectly secure, artificial intelligence (yeah, I know, blah blah blah Skynet), use-Tor-use-Signal. There’s a place for mockery and snark. This isn’t concern trolling, which is preoccupied with how things are said. This is about the understanding behind what is said — the risk calculation, the threat model, the constraints.

Constructive Closing

PourI believe in supporting people to self-identity along the spectrum of builder and breaker rather than pin them to narrow roles. (A principle applicable to many more important subjects as well.) This about the intellectual reward of tackling challenges faced by builders and breakers alike, and leaving behind the blather of uninformed opinions and empty solutions.

I’ll close with this observation from Carl Sagan (from his book, The Demon-Haunted World): “It is far better to grasp the universe as it really is than to persist in delusion, however satisfying and reassuring.”

Our application universe consists of systems and data and users, each in different orbits. Security should contribute to the gravity that binds them together, not the black hole that tears them apart. Engineering sees the universe as it really is; shed the delusion that one appsec solution in a vacuum is always universal.

Cheap Essential Scenery

Keep Calm and Never MindThis October people who care about being aware of security in the cyberspace of their nation will celebrate the 10th anniversary of National Cyber Security Awareness Month. (Ignore the smug octal-heads claiming preeminence in their 12th anniversary.) Those with a better taste for acronyms will celebrate Security & Privacy Awareness Month.

For the rest of information security professionals it’s just another TUESDAY (That Usual Effort Someone Does All Year).

In any case, expect the month to ooze with lists. Lists of what to do. Lists of user behavior to be reprimanded for. What software to run, what to avoid, what’s secure, what’s insecure. Keep an eye out for inconsistent advice among it all.

Ten years of awareness isn’t the same as 10 years of security. Many attacks described decades ago in places like Phrack and 2600 either still work today or are clear antecedents to modern security issues. (Many of the attitudes haven’t changed, either. But that’s for another article.)

Web vulns like HTML injection and SQL injection have remained fundamentally unchanged across the programming languages that have graced the web. They’ve been so static that the methodologies for exploiting them are sophisticated and mostly automated by now.

Awareness does help, though. Some vulns seem new because of awareness (e.g. CSRF and clickjacking) even though they’ve haunted browsers since the dawn of HTML. Some vulns just seem more vulnerable because there are now hundreds of millions of potential victims whose data slithers and replicates amongst the cyber heavens. We even have entire mobile operating systems designed to host malware. (Or is it the other way around?)

So maybe we should be looking a little more closely at how recommendations age with technology. It’s one thing to build good security practices over time; it’s another to litter our cyberspace with cheap essential scenery.

Here are two web security examples from which a critical eye leads us into a discussion about what’s cheap, what’s essential, and what actually improves security.

Cacheing Can’t Save the Queen

I’ve encountered recommendations that insist a web app should set headers to disable the browser cache when it serves a page with sensitive content. Especially when the page transits HTTP (i.e. an unencrypted channel) as well as over HTTPS.

That kind of thinking is deeply flawed and when offered to developers as a commandment of programming it misleads them about the underlying problem.

If you consider some content sensitive enough to start worrying about its security, you shouldn’t be serving it over HTTP in the first place. Ostensibly, the danger of allowing the browser to cache the content is that someone with access to the browser’s system can pull the page from disk. It’s a lot easier to sniff the unencrypted traffic in the first place. Skipping network-based attacks like sniffing and intermediation to focus on client-side threats due to cacheing ignores important design problems — especially in a world of promiscuous Wi-Fi.

Then you have to figure out what’s sensitive. Sure, a credit card number and password are pretty obvious, but the answer there is to mask the value to avoid putting the raw value into the browser in the first place. For credit cards, show the last 4 digits only. For the password, show a series of eight asterisks in order to hide both its content and length. But what about email? Is a message sensitive? Should it be cached or not? And if you’re going to talk about sensitive content, then you should be thinking of privacy as well. Data security does not equal data privacy.

And if you answered those questions, do you know how to control the browser’s cacheing algorithm? Are you sure? What’s the recommendation? Cache controls are not as straight-forward as they seem. There’s little worth in relying on cache controls to protect your data from attackers who’ve gained access to your system. (You’ve uninstalled Java and Flash, right?)

Browsers used to avoid cacheing any resource over HTTPS. We want sites to use HTTPS everywhere and HSTS whenever possible. Therefore it’s important to allow browsers to cache resources loaded via HTTPS in order to improve performance, page load times, and visitors’ subjective experiences. Handling sensitive content should be approached with more care than just relying on headers. What happens when a developer sets a no-cacheing header, but saves the sensitive content in the browser’s Local Storage API?

HttpOnly Is Pretty Vacant

Web apps litter our browsers with all sorts of cookies. This is how some companies get billions of dollars. Developers sprinkle all sorts of security measures on cookies to make them more palatable to privacy- and security-minded users. (And weaken efforts like Do Not Track, which is how some companies keep billions of dollars.)

The HttpOnly attribute was proposed in an era when security documentation about HTML injection attacks (a.k.a. cross-site scripting, XSS) incessantly repeated the formula of attackers inserting <img> tags whose src attributes leaked victims’ document.cookie values to servers under the attackers’ control. It’s not wrong to point out such an exploit method. However, as Stephen King repeated throughout the Dark Tower series, “The world has moved on.” Exploits don’t need to be cross-site, they don’t need <script> tags in the payload, and they surely don’t need a document.cookie to be effective.

If your discussion of cookie security starts and ends with HttpOnly and Secure attributes, then you’re missing the broader challenge of designing good authorization, authentication, and session handling mechanisms. If the discussion involves using the path attribute as a security constraint, then you shouldn’t be talking about cookies or security at all.

HttpOnly is a cheap attribute to throw on a cookie. It doesn’t prevent sniffing — use HTTPS everywhere for that (notice the repetition here?). It doesn’t really prevent attacks, just a single kind of exploit technique. Content Security Policy is a far more essential countermeasure. Let’s start raising awareness about that instead.

Problems

Easy security measures aren’t useless. Prepared statements are easy to use and pretty soundly defeat SQL injection; developers just choose to remain ignorant of them.

This month be extra wary of cheap security scenery and stale recommendations that haven’t kept up with the modern web. Ask questions. Look for tell-tale signs like they

  • fail to clearly articulate a problem with regard to a security or privacy control (e.g. ambiguity in what the weakness is or what an attack would look like)
  • fail to consider the capabilities of an attack (e.g. filtering script and alert to prevent HTML injection)
  • do not provide clear resolutions or do not provide enough details to make an informed decision (e.g. can’t be implemented)
  • provide contradictory choices of resolution (e.g. counter a sniffing attack by applying input validation)

Oh well, we couldn’t avoid a list forever.

Never mind that. I’ll be back with more examples of good and bad. I can’t wait for this month to end, but that’s because Halloween is my favorite holiday. We should be thinking about security every month, every day. Just like the song says, Everyday is Halloween.

30% of the 2010 OWASP Top 10 not common, only 1 not hard to detect

Letter O One curious point about the new 2010 OWASP Top 10 Application Security Risks is that 30% of them aren’t even common. The “Weakness Prevalence” for each of Insecure Cryptographic Storage (A7), Failure to Restrict URL Access (A8), and Unvalidated Redirects and Forwards (A10) is rated uncommon. That doesn’t mean that an uncommon risk can’t be a critical one; these three points highlight the challenge of producing an unbiased list and conveying risk.

Risk is complex and difficult to quantify. The OWASP Top 10 includes a What’s My Risk? section that provides guidance on how to interpret the list. The list is influenced by the experience of people who perform penetration tests, code reviews, or conduct research on web security.

The Top 10 rates Insecure Cryptographic Storage (A7) with an uncommon prevalence and difficult to detect. One of the reasons it’s hard to detect is that back-end storage schemes can’t be reviewed by blackbox scanners (i.e. web scanners) nor can source code scanners point out these problems other than by indicating misuse of a language’s crypto functions. So, one interpretation is that insecure crypto is uncommon only because more people haven’t discovered or revealed such problems. Yet not salting a password hash is one of the most egregious mistakes a web app developer can make and one of the easier problems to fix — the practice of salting password hashes has been around since Unix epoch time was in single digits.

It’s also interesting that insecure crypto is the only one on the list that’s rated difficult to detect. On the other hand, Cross-Site Scripting (A2) is “very widespread” and “easy” to detect. But maybe it’s very widespread because it’s so trivial to find. People might simply focus on searching for vulns that require minimal tools and skill to discover. Alternately, XSS might be very widespread because it’s not easy to find in a way that scales with development or keeps up with the complexity of web apps. (Of course, this assumes that devs are looking for it in the first place.) XSS tends to be easy for manual testing, whereas scanners need to effectively crawl an app before they can go through their own detection techniques.

Broken Authentication and Session Management (A3) covers brute force attacks against login pages. It’s an item whose risk is too-often demonstrated by real-world attacks. In 2010 the Apache Foundation suffered an attack that relied on brute forcing a password. (Apache’s network had a similar password-driven intrusion in 2001. These events are mentioned because of the clarity of their postmortems, not to insinuate that the org inherently insecure.) In 2009 Twitter provided happiness to another password guesser.

Knowing the password to an account is the best way to pilfer a user’s data and gain unauthorized access to a site. The only markers for an attacker using valid credentials are behavioral patterns – time of day the account was accessed, duration of activity, geographic source of the connection, etc. The attacker doesn’t have to use any malicious characters or techniques that could trigger XSS or SQL injection countermeasures.

The impact of a compromised password is similar to how CSRF (A5) works. The nature of a CRSF attack is to force a victim’s browser to make a request to the target app using the context of the victim’s authenticated session in order to perform some action within the app. For example, a CSRF attack might change the victim’s password to a value chosen by the attacker, or update the victim’s email to one owned by the attacker.

By design, browsers make many requests without direct interaction from a user (such as loading images, CSS, JavaScript, and iframes). CSRF requires the victim’s browser to visit a booby-trapped page, but doesn’t require the victim to click anything. The target web app neither sees the attacker’s traffic nor even suspects the attacker’s activity because all of the interaction occurs between the victim’s browser and the app.

CSRF serves as a good example of the changing nature of the web security industry. CSRF vulnerabilities have existed as long as the web. The attack takes advantage of the fundamental nature of HTML and HTTP whereby browsers automatically load certain types of resources. Importantly, the attack just needs to build a request. It doesn’t need to read the response, hence it isn’t inhibited by the Same Origin Policy.

CSRF hopped on the Top 10 list’s revision in 2007, three years after the list’s first appearance. It’s doubtful that CSRF vulnerabilities were any more or less prevalent over that three year period (or even the before 2000). Its inclusion was a nod to having a better understanding and appreciation of the risk associated with the vuln. And it’s a risk that’s likely to increase when the pool of victims can be measured in the hundreds of millions rather than the hundreds of thousands.

This vuln also highlights an observation bias of security researchers. Now that CSRF is in vogue people start to look for it everywhere. Security conferences get more presentations about advanced ways to exploit the vulnerability, even though real-world attackers seem fine with the returns on guessing passwords, seeding web pages with malware, and phishing. Take a look at HTML injection (what everyone else calls XSS). Injecting script tags into a web page via an unchecked parameter dates back to the beginning of the web.

Before you shrug off this discussion of CSRF as hand waving with comments like, “But I could hack site Foo by doing Bar and then make lots of money,” consider what you’re arguing: A knowledgeable or dedicated attacker will find a useful exploit. Risk may include many components, including Threat Agents (to use the Top 10’s term). Risk increases under a targeted attack — someone actively looking to compromise the app or its users’ data. If you want to add an “Exploitability” metric to your risk calculation, keep in mind that ease of exploitability is often related to the threat agent and tends to be a step function: It might be hard to figure out the exploit in the first place, but anyone can run a 42-line Python script that automates the attack.

This is why the Top 10 list should be a starting point to defining security practices for your web site, but it shouldn’t be the end of the road. Even the OWASP site admonishes readers to use the list for awareness rather than policy. So, if you’ve been worried about information leakage and improper error handling since 2007 don’t think the problem has disappeared because it’s not on the list in 2010. And if you don’t think anyone cares about the logic problems within your site…well, just hope they haven’t been reading about them somewhere else.