Dark Reading is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them.Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

Cloud

9/17/2019
02:00 PM
Peter Smith
Peter Smith
Commentary
Connect Directly
Twitter
LinkedIn
RSS
E-Mail vvv
50%
50%

5 Common Cloud Configuration Mistakes

It's a joint responsibility to keep data safe in the cloud. Here's what cloud customers must do to keep their end of the bargain.

There's no doubt that the cloud can improve certain aspects of security. After all, clouds have enormous economies of scale that provide customers with dedicated security teams and technologies that aren't feasible for the vast majority of organizations. That's the good news. The bad news happens when customers don't properly configure and secure their own workloads and buckets in a cloud environment.

Consider the recent Capital One data breach, where a hacker exploited a misconfigured cloud firewall to gain access to data on 100 million card customers and applicants in one of the largest breaches in history. There are thousands of potential cloud misconfiguration mistakes such as the one that felled Capital One. But, as with many things, most misconfiguration errors can be classified into a handful of distinct categories. Here are five of the most common areas where cloud misconfiguration attacks happen.

Mistake 1: Storage Access
When it comes to storage buckets, many cloud users think that "authenticated users" covers only those who are already authenticated within their organization or the relevant application. This is, unfortunately, not the case. "Authenticated users" refers to anyone with Amazon Web Services (AWS) authentication, which is effectively any AWS customer. Because of this misunderstanding, and the resulting misconfiguration of the control settings, storage objects can end up fully exposed to public access. Be especially careful when setting storage object access to ensure that only those within your organization who need access have it.

Mistake 2: "Secrets" Management

This configuration mistake can be especially damaging to an organization. It's critical to ensure that secrets such as passwords, API keys, admin credentials and encryption keys are secured. I've seen them openly available in badly configured cloud buckets, compromised servers, open GitHub repositories, and even in HTML code. It's the equivalent of leaving the key to your home's deadbolt taped to the front door.

The solution is to maintain an inventory of all secrets that you use in the cloud, and regularly check to see how each is secured. Otherwise, malicious actors could easily access all your data. Worse, they can take control of your cloud resources to do irreparable damage. Equally as important is the use of a secrets management system. Services such as AWS Secrets Manager, AWS Parameter Store, Azure Key Vault, and Hashicorp Vault are some examples of robust and scalable secrets management tools.

Mistake 3: Disabled Logging and Monitoring
It's surprising how many organizations don't enable, configure, or even review the logs and telemetry data that public clouds provide, which in many cases can be extremely sophisticated. Someone on your enterprise cloud team should have the responsibility for regularly reviewing this data and flagging security-related events.

This advice isn't limited to infrastructure-as-a-service public clouds. Storage-as-a-service vendors often provide similar information, which again, needs to be regularly reviewed. An update bulletin or maintenance alert could have serious security implications for your organization, but it won't do you any good if no one is paying attention.

Mistake 4: Overly Permissive Access to Hosts, Containers and Virtual Machines
Would you directly connect a physical or virtual server in your data center to the Internet without a filter or firewall to protect it? Of course not. But people do almost exactly that in the cloud all the time. A few examples we've recently seen include:

  • Etcd (port 2379) for Kubernetes clusters exposed to the public Internet
  • Legacy ports and protocols such as FTP enabled on cloud hosts
  • Legacy ports and protocols such as rsh, rexec, and telnet in physical servers that have been virtualized and migrated to the cloud.

Make sure you secure important ports and disable — or at the very least lock down — older, insecure protocols in the cloud, just as you would in your on-premises data center.

Mistake 5: Lack of Validation
This final cloud mistake is a meta-issue: We often see organizations fail to create and implement systems to identify misconfigurations as they occur. Whether it's an internal resource or an outside auditor, someone must be responsible for regularly verifying that services and permissions are properly configured and applied. Set a schedule to ensure this occurs like clockwork, because as the environment changes, mistakes are inevitable. You'll also need to establish a rigorous process to periodically audit cloud configurations. Otherwise, you risk leaving a security flaw in place that malicious actors can exploit.

The cloud has the potential to be a secure place for data and workloads, but only if cloud customers live up to their side of its dual responsibility model. By keeping these common mistakes in mind and setting up a system to catch them as quickly as they happen, you can be sure that your digital assets in the cloud will be safe. 

Related Dark Reading Content:

Check out The Edge, Dark Reading's new section for features, threat data, and in-depth perspectives. Today's top story: "How a PIA Can CYA."

Peter Smith, Edgewise Founder and CEO, is a serial entrepreneur who built and deployed Harvard University's first NAC system before it became a security category. Peter brings a security practitioner's perspective to Edgewise with more than ten years of expertise as an ... View Full Bio
 

Recommended Reading:

Comment  | 
Print  | 
More Insights
Comments
Threaded  |  Newest First  |  Oldest First
Why Vulnerable Code Is Shipped Knowingly
Chris Eng, Chief Research Officer, Veracode,  11/30/2020
Inside North Korea's Rapid Evolution to Cyber Superpower
Kelly Sheridan, Staff Editor, Dark Reading,  12/1/2020
Register for Dark Reading Newsletters
White Papers
Video
Cartoon Contest
Write a Caption, Win an Amazon Gift Card! Click Here
Latest Comment: This comment is waiting for review by our moderators.
Current Issue
2021 Top Enterprise IT Trends
We've identified the key trends that are poised to impact the IT landscape in 2021. Find out why they're important and how they will affect you today!
Flash Poll
Assessing Cybersecurity Risk in Todays Enterprises
Assessing Cybersecurity Risk in Todays Enterprises
COVID-19 has created a new IT paradigm in the enterprise and a new level of cybersecurity risk. This report offers a look at how enterprises are assessing and managing cyber-risk under the new normal.
Twitter Feed
Dark Reading - Bug Report
Bug Report
Enterprise Vulnerabilities
From DHS/US-CERT's National Vulnerability Database
CVE-2020-25465
PUBLISHED: 2020-12-04
Null Pointer Dereference. in xObjectBindingFromExpression at moddable/xs/sources/xsSyntaxical.c:3419 in Moddable SDK before OS200908 causes a denial of service (SEGV).
CVE-2020-25461
PUBLISHED: 2020-12-04
Invalid Memory Access in the fxProxyGetter function in moddable/xs/sources/xsProxy.c in Moddable SDK before OS200908 causes a denial of service (SEGV).
CVE-2020-25462
PUBLISHED: 2020-12-04
Heap buffer overflow in the fxCheckArrowFunction function at moddable/xs/sources/xsSyntaxical.c:3562 in Moddable SDK before OS200903.
CVE-2020-25463
PUBLISHED: 2020-12-04
Invalid Memory Access in fxUTF8Decode at moddable/xs/sources/xsCommon.c:916 in Moddable SDK before OS200908 causes a denial of service (SEGV).
CVE-2020-25464
PUBLISHED: 2020-12-04
Heap buffer overflow at moddable/xs/sources/xsDebug.c in Moddable SDK before before 20200903. The top stack frame is only partially initialized because the stack overflowed while creating the frame. This leads to a crash in the code sending the stack frame to the debugger.