Monthly Archives: October 2018

Proactive System Hardening: Continuous Hardening’s Coming of Age

The first article in this series examined configuration hardening—essentially looking at ports, processes and services where security configuration management (SCM) is key. The second article looked at application and version hardening strategies. This third installment will discuss the role of automation in the coming of age of what’s called “continuous hardening.”

Known Vulnerabilities vs. Conditional Vulnerabilities

If I want to harden my systems against “known vulnerabilities”—weaknesses or deficiencies for which there are known common vulnerabilities and exposures (CVEs)—I use a vulnerability management solution. If I need to harden my systems against “conditional vulnerabilities”—weaknesses based on the way they’re configured—I use an SCM solution. But without automation to provide the element of “continuousness” to these efforts, we rapidly find ourselves back at square one.

What is Configuration Drift?

To stick with our house analogy: If I’ve checked the configurations of all my doors and windows, but I have no way to know when the state has changed and I instead rely on periodic inspection by human eyes, a phenomenon known as “configuration drift” invariably occurs.

I open the fire escape window to water the potted hydrangea sitting out there but forget to close it afterward: configuration drift. I enable Telnet to maintain or update a server and then forget to disable it afterward: configuration drift.

The Role of Automation in Continuous System Hardening

A primary weakness of our house analogy is actually useful here, as it shows us the critical need for automation. In real life, most people have one house. But most organizations have hundreds—if not many, many thousands—of servers, desktop systems, laptops and devices. These represent an almost inexhaustible supply of attack surface and potential beachheads. How can we win a war at this scale?

Automation requires us to not only create continuous, ongoing routines to assess states across this vast array of targets, but it also requires us to make allowances for the constantly changing conditions that give meaning and relevance to risk.

In the case of our house, it’s useful to know that, over the last two years, the leafy maple out back has grown a large solid branch that’s close enough to an upstairs bedroom for a tall thief to reach the window. And the inverse is sometimes true: If the old kitchen window was painted shut twenty years ago, who needs to waste time including it in our daily “is it locked” checklist?

This critical need for current “state” information has caused the security community to create more persistent real-time agents, more effective scanning processes that are “aware” of network constraints and ways to avoid “mega scans” in favor of continuous segmented scanning.

Integrating Disparate Security Systems

They’ve also broken down barriers between infosec solutions themselves and addressed another critical requirement for achieving this attribute of “continuousness”: Information security systems must talk to one another. A few simple examples illustrate this need:

  • Vulnerability Management: Vulnerability management (VM) systems are quite good at finding unexpected (and likely unsecured) systems. When one of these is discovered, the VM system can tell the SCM system about the new asset and ask it to perform an on-the-spot configuration assessment.
  • Security Configuration Management: Similarly, SCM systems are evolving intelligent ways to classify assets: by business unit, by system owner, by critical application, and even by the type and criticality of data stored on the system. This helps manage and prioritize their own risks, but when shared with a VM system, this also helps clarify and prioritize remediation efforts.
  • Security Information and Event Management: Both of these systems are being used extensively by SIEM systems as a foundational source of security information: in the first case, correlating known vulnerabilities with detected threats, and in the second case, using sudden configuration changes (“Why is the ‘Telnet should not be enabled‘ test suddenly failing?“) to power real-time threat intelligence models.

SC Magazine summed up these needs in a prescient review of policy management systems—what we’ve called “security configuration management” systems in this article—way back in 2010: “The only reasonable answer to the challenges of compliance, security and configuration management is to automate the tasks.”

The key to continuous system hardening as a goal and a discipline is a willingness to seek out and employ automation wherever possible. Gone are the days when isolated, siloed systems can harden information systems and keep them that way in the face of continuous drift.

Highly interactive solutions that understand the ever-shifting nature of “state” and talk to each other regularly—security configuration and vulnerability management solutions in particular—are the first, best and often the last line of defense.

 

via:  tripwire

Proactively Hardening Systems: Application and Version Hardening

The first article in this series examined configuration hardening, essentially looking at ports, processes and services as the “doors, gates and windows” into a network where security configuration management (SCM) becomes the job of determining which of these gateways should be open, closed, or locked at any given time. Now it’s time to look at application and version hardening.

What is System Hardening?

If configuration hardening settings are “conditional,” meaning they must find and keep that balance between security and productivity, then hardening against known vulnerabilities in applications and versions is much more black-and-white.

If an exploit path has been found in an operating system or application, the vendor rushes to create a patch or upgrade that removes the vulnerability. “Hardening” in this sense means “making sure the holes are known and that the most current security patches are deployed.”

One Way Hackers Exploit Known Vulnerabilities

To go back to our “secure house” analogy from the previous article in this series for a moment, imagine that the house I’m protecting has three external doors and that they all use Secure-A-Door Model 800 high-strength locks.

But a tester at the Secure-A-Door factory (or worse, a professional burglar) has just discovered an interesting thing: If you slide a credit card along the door jamb at 15 degrees while pulling up on the handle, the Secure-A-Door 800 pops open like a Coke can.

One of the most famous examples of this exploitation began in 2008. That’s when the makers of the Conficker worm discovered and exploited an underlying weakness in Port 445 of the Windows operating system.

The worm created a remote procedure call that dropped a DLL on the system, unloaded two distinct packets for data and code, and hid itself in a remote thread to make itself at home. (It was infinitely more complex and clever than that, but you get the idea.)

In effect, the worm popped the Secure-A-Door Model 800, let itself in, repaired the lock, installed a new phone line to listen for orders, and sat in a comfy chair waiting for instructions. It was able to leverage the internet, could register new domain names in which to hide, and created an extensive botnet that by 2010 had infected, according to Panda Security, as many as 18 million PCs—6 percent of the world’s PC population at the time.

Common Vulnerabilities and Exposures (CVEs)

This type of design failure or exploit is usually repaired by a patch. In the case of Conficker, Windows Security bulletin MS08-067 made the danger known to the worldwide Microsoft community and introduced a patch to prevent easy violation of Port 445.

The MS bulletin was in turn translated by the Common Vulnerabilities and Exposures site as CVE-2008-4250 and given a Common Vulnerability Scoring System (CVSS) rating of 10—the most severe rating possible.

Vulnerability Management

Vulnerability management (VM) systems, unlike SCM systems that check to see that doors and gates and windows are locked, do their part in system hardening differently. They make sure the proper patch levels are maintained and that any available defenses have been utilized. Using our analogy, we’d be conducting the following checks:

  • Proactively discovering whether I have any Secure-A-Door Model 800 locks installed
  • If I do, reporting on whether they’re the corrected “B” version made after October 2012
  • Verifying that any “bad” ones I have are only on inside doors and don’t serve as a primary defense

VM systems enable continuous hardening by making sure that CVE-2008-4250—and its many thousands of friends—are understood, mitigated, and more-or-less unexploitable when the right steps are taken.

More mature solutions provide an ongoing assessment of overall risk based on whether these vulnerabilities are mitigated or ignored.

 

via:  tripwire

California passes law that bans default passwords in connected devices

Good news!

California has passed a law banning default passwords like “admin,” “123456” and the old classic “password” in all new consumer electronics starting in 2020.

Every new gadget built in the state from routers to smart home tech will have to come with “reasonable” security features out of the box. The law specifically calls for each device to come with a preprogrammed password “unique to each device.”

It also mandates that any new device “contains a security feature that requires a user to generate a new means of authentication before access is granted to the device for the first time,” forcing users to change the unique password to something new as soon as it’s switched on for the first time.

For years, botnets have utilized the power of badly secured connected devices to pummel sites with huge amounts of internet traffic — so-called distributed denial-of-service (DDoS) attacks. Botnets typically rely on default passwords that are hardcoded into devices when they’re built that aren’t later changed by the user. Malware breaks into the devices using publicly available default passwords, hijacks the device and ensnares the device into conducting cyberattacks without the user’s knowledge.

Two years ago, the notorious Mirai botnet dragged thousands of devices together to target Dyn, a networking company that provides domain name service to major sites. By knocking Dyn offline, other sites that relied on its services were also inaccessible — like Twitter, Spotify and SoundCloud.

Mirai was a relatively rudimentary, albeit powerful botnet that relied on default passwords. This law is a step in the right direction to prevent these kinds of botnets, but falls short on wider security issues.

Other, more advanced botnets don’t need to guess a password because they instead exploit known vulnerabilities in Internet of Things devices — like smart bulbs, alarms and home electronics.

As noted by others, the law as signed does not mandate device makers to update their software when bugs are found. The big device makers, like Amazon, Apple and Google, do update their software, but many of the lesser-known brands do not.

Still, as it stands, the law is better than nothing — even if there’s room for improvement in the future.

 

via:  techcrunch

Google+ Shutting Down After Bug Leaks Info of 500k Accounts

Google has announced that they are closing the consumer functionality of Google+ due lack of adoption and an API bug that leaked the personal information of up to 500,000 Google+ accounts.

While no evidence was found that indicates this bug was ever misused, it was determined that the complexity of protecting and operating a social network like Google+ was not a worthwhile endeavor when so few users actually used the service for any length of time.

“This review crystallized what we’ve known for a while: that while our engineering teams have put a lot of effort and dedication into building Google+ over the years, it has not achieved broad consumer or developer adoption, and has seen limited user interaction with apps,” stated a blog post by Google regarding the Google+ closure. “The consumer version of Google+ currently has low usage and engagement: 90 percent of Google+ user sessions are less than five seconds.”

The consumer functionality of Google+ will be closing over a 10 month period, while Google transitions the product to be used internally by the Enterprise.

API bug caused data leak

After performing a code review of the Google+ APIs, called Project Strobe, Google stated they discovered a bug that could leak the private information of Google+ accounts. This bug could allow a user’s installed apps to utilize the API and access non-public information belonging to that user’s friends. The non-public information that was accessible includes an account holder’s name, email address, occupation, gender and age.

Underlining this, as part of our Project Strobe audit, we discovered a bug in one of the Google+ People APIs:

  • Users can grant access to their Profile data, and the public Profile information of their friends, to Google+ apps, via the API.
  • The bug meant that apps also had access to Profile fields that were shared with the user, but not marked as public. 
  • This data is limited to static, optional Google+ Profile fields including name, email address, occupation, gender and age. (See the full list on our developer site.) It does not include any other data you may have posted or connected to Google+ or any other service, like Google+ posts, messages, Google account data, phone numbers or G Suite content.
  • We discovered and immediately patched this bug in March 2018. We believe it occurred after launch as a result of the API’s interaction with a subsequent Google+ code change.

As Google only keeps two weeks of API logs for its Google+ service, it was impossible for them to determine if the bug was ever misused. They were able to determine that the bug was not misused during the two weeks that they had log data.

Google knew about leak in May but did not disclose

According to a report by the Wall Street Journal, the bug in the Google+ API existed between 2015 and March 2018, which was when Google discovered and fixed the bug. According to their reporting, an internal committee at Google decided not to disclose the bug even though they were not 100% sure that it was not abused.

The Wall Street Journal, reported that they have reviewed a memo prepared by Google’s legal and policy staff, which indicated that disclosing the data breach could lead to scrutiny by government regulatory agencies.

“disclosing the incident would likely trigger “immediate regulatory interest” and invite comparisons to Facebook’s leak of user information to data firm Cambridge Analytica.”

In a statement, a Google Spokesperson said that their Privacy & Data Protection Office felt it was not necessary to disclose as it did not meet the threshold that would warrant it.

“Every year, we send millions of notifications to users about privacy and security bugs and issues. Whenever user data may have been affected, we go beyond our legal requirements and apply several criteria focused on our users in determining whether to provide notice.

Our Privacy & Data Protection Office reviewed this issue, looking at the type of data involved, whether we could accurately identify the users to inform, whether there was any evidence of misuse, and whether there were any actions a developer or user could take in response. None of these thresholds were met in this instance.

The review did highlight the significant challenges in creating and maintaining a successful Google+ that meets consumers’ expectations. Given these challenges and the very low usage of the consumer version of Google+, we decided to sunset the consumer version of Google+.” – Google Spokesperson.

 

via:  bleepingcomputer