Peeling Back the Layers of the Cloud Security Onion
I had the chance to sit down with Lee Slaughter, Skytap’s Manager of Information Security, to discuss security in the cloud, the efforts that make up a thorough due diligence when choosing a cloud provider, and how to keep up with external and internal threats to your business.
Noel: Last year, Charles Babcock wrote a piece for InformationWeek on the list of the top 9 worst cloud security threats that was put out by the Cloud Security Alliance. As I read through them, I noticed that the threats were predominantly related to malicious insiders and outsiders, data breaches, denial of service attacks, etc.
One of the listed threats (#8) was “insufficient due diligence.” That’s…obviously a really important one. What are some of the ways that an organization looking for a cloud services provider can make sure they do their own due diligence, and are there any areas that are often or easier overlooked?
Lee: Due diligence of a cloud vendor should be an extension of the security risk assessment program that is internal to your organization. First, you must scope and classify the vendor’s service(s) according to the value of the asset you’re looking to outsource. The value of the asset directly correlates to the amount of scrutiny that is placed on the assessment.
For example, you’re looking for cloud vendors that could host your SaaS web site. Let’s say that this site is mission critical for your organization because it delivers your product to your customers. It’s so important that you’re requiring a 99.9% uptime on that site, because you can calculate the unbelievable dollar amount that is lost when the site goes down, even for minutes. What assurances can those vendors give you that they can live up to the 99.9% uptime SLA that they signed up to provide?
To ensure you’re properly assessing a vendor, use an industry standard framework. You most likely have some sort of standard that you certify or align to, so you can use the same for this purpose. ISO27k or NIST are very commonly used frameworks. Ask your vendor if they certify or align to these, and ensure their security certifications are scoped to protect the things relevant to you and your systems or data that reside in their cloud. If the vendor only aligns to these standards, ask for more details. Other factors that determine level of scrutiny include the laws and/or regulations that you are obligated to follow. Make sure you’re not violating any of these obligations by putting this asset in the cloud.
Noel: This list of threats, like most warnings regarding the cloud, brings up both the good and the bad of “Shadow IT.” You’ve mentioned that “there’s a balance between enabling your employees to be happy and agile, and locking things down as much as possible.” This seems like another area where a thorough due diligence period has to be taken seriously, but what else have you seen be effective in combatting Shadow IT from getting out of control/visibility.
Lee: Security management/staff require a good understanding of the business and should have enough visibility into the organization to know what product/services will be offered in the future. Questions such as, “Does a law/regulation require us to do this in a certain way?” or “What happens if someone gets unauthorized access to this system?” should be asked. Determine the threshold for control of a system or data based on its value. If the value of an asset is at the top of your scale, then your controls around that system/data should leave no room for shadow IT.
Noel: I think one of the reasons that people fear or are reluctant to embrace cloud technologies is the widespread front page news that hacks/breaches/leaks tend to make. We’ve seen the impacts that these attacks and/or oversights can have, and a recent study showed that “15% of logins for business apps used by organizations had been breached by hackers.” Where would you say the fault often lies in these breaches? The cloud provider is likely quick to be blamed when the app itself was hacked, but what are some other potential, or even likely reasons that should be considered when investigating the security hole or failure?
Lee: As for fault in the 15% of breached logins, that depends. Cloud companies, who have been hacked and login credentials were stolen in usable form, are 100% at fault. By “usable form”, I mean credentials that were NOT hashed and salted. Security should be practiced in many layers, often called the “security onion.” If a company has a breach in outer area(s), the exposure should be limited. Many companies have what I call the M&M model of security…the hard coating that protects the outside, but nice and soft in the middle…meaning they don’t have those necessary internal controls for prevention or detection of a breach.
Other reasons for breached logins would have nothing to do with the vendor at all. Unsuspecting users quite often fall for phishing scams that result in the bad guys getting the login credentials. Additionally, with so many apps and sites that require usernames and passwords, users typically re-use passwords and/or usernames. If one vendor with failing security controls allows an attacker to obtain those credentials, there’s a good chance that those same credentials can be used for other apps/sites. This is where a password manager can come in handy to enable the use of different passwords for all of your apps but without having to remember all of them.
Noel: That same study said, “A quarter of all files in cloud storage apps are shared with one or more people outside of the organization.” This reminds me of one of the areas of Skytap that I’ve always thought was really cool – the ability to only share certain data with teams that truly need it.
For instance, not exposing production data and environments with test/QA teams, instead, providing them with relevant mock test data. The production data never even enters the cloud. What ways can security teams ensure that controls are not just established, but monitored and maintained in regards to sensitive data?
Lee: This goes back to the risk assessment I discussed earlier. When certain data or systems are deemed to be of the highest sensitivity, you must rely on the features available through your cloud vendor to ensure you’re addressing every possible risk that you identify during the assessment.
To ensure that controls are monitored and continuously improved, you must create a security program around those controls. This requires that you have a series of defined standards and procedures on the right way to do things. A defined, mature, control leaves no room for ambiguity in the steps that are taken to ensure security of data or systems. The controls should be monitored for any deviation of expected outcome and tweaked for improvement as necessary.
Noel: You often hear or read about how hackers are often “one step ahead” of software security, but that seems like somewhat of a negative assumption. How are you, and Skytap’s other security-minded pros, utilizing the latest technology or even just forward-thinking strategies to combat this assumption that the cloud will always be risky business?
Lee: I’m actually in that camp as well. Hackers are and will always be one step ahead. Hackers get their foot in the door by tricking users and/or using a vulnerability to gain access. There will ALWAYS be at least one user that clicks that malicious link within an email and there will ALWAYS be zero-day vulnerabilities that are not foreseen or preventable. The security industry is shifting to doing more than the old ways of doing things – harden the outside and buy technology and implement processes for prevention of breaches. The shift is now to how fast can you detect and react to a breach. We all still need to do the things we’ve been doing, while increasing our focus on incident detection and response.