Posts

Simplifying and Securing Hybrid Clouds

GovLoop | October 26, 2021

President Joe Biden’s executive order (EO) on cybersecurity suggests the cloud will play a pivotal role in the federal government’s future; it urges agencies to maximize the technology’s flexibility and scalability rapidly and securely.

But what can happen if agencies embrace the cloud too rapidly? The answer is haphazard and insecure IT environments. These environments often occur when agencies combine on-premises and cloud-based IT in a hybrid model.

Understanding What’s In My Cloud

Today’s business applications run in an environment that would be unrecognizable to IT professionals 10 years ago. The rise of virtualization and the cloud has finally cut the ties to specific hardware, and all but the most exotic workloads can now be run anywhere — on virtual machines in your physical buildings, or on a cloud vendor of your choice. The underlying cloud technologies are powerful, but with that power comes great responsibility. Security teams struggle to keep up, because the new technologies focus on agility, rapid rate of change, and dynamic response — all of these are positive buzzwords to most people in a business, but all of them are bad news to security. Ask any military commander — defense is far easier when your resources are home in a well-built fort, and far harder when your troops are constantly moving, shifting location into unfamiliar terrain.

It’s not all doom and gloom, however. Cloud innovation takes away certain legacy risks — after all, you can’t leave an open password on a key router in the middle of your network infrastructure if you don’t control the routers any more! The trouble is that the change to new ways of building and managing modern apps (often referred to as DevOps) closes out some old challenges, but opens just as many new ones. Cloud gives you new kinds of rope, and it’s different from the old rope, but you can still get just as tangled up in the complexities.

Some security fundamentals remain, though. No matter what kind of infrastructure you own or rent, you still need to pursue the basics:

1.    Find all your stuff

2.    Categorize it so you know what’s most important

3.    Harden the individual elements to avoid easy compromise

4.    Map out and run your defenses as a system, so you can be a hard target

The most basic discipline of all is inventory — cyber security experts and industry guidance all agree that you must start there. Inventory in cloud is not like inventory in conventional networks, though, so the same old principle has to be thought about differently in a cloud world.

The good news with the cloud is that each virtual network has a “God of the Cloud” — a central controller, run by the cloud provider that you can talk to via a proprietary API. I call it a “God”, because no endpoints can exist in that small virtual network that the controller did not create. This means you can always find a completely reliable resource for each virtual network — someone who knows the inventory. Problem solved, right? Well, not so fast — it’s certainly very different from legacy on-premises networks, but that’s hardly all there is to it. There are three major problems when talking to each cloud controller — finding the controllers, speaking their language, and keeping up with the changes.

The good news is a cloud account comes with an API you can talk to and get a complete inventory of the assets it knows about. The bad news is your company has many, many accounts. And even once you locate them all, they will speak a proprietary and changing language — the Amazon language for the AWS API is different from Microsoft’s for Azure, or Google’s, or Oracle’s. You need a network linguist to make sense of it all, and pull together a single view of your clouds — in all flavors. And since security is central by its nature (because it needs to look at the complete picture), that means security has the unenviable task of needing to speak all the languages — fluently — at once. This is hard, but it’s a great job for automated software.

Equally, the rate of change in the cloud is something automated software can tackle far more effectively than humans can. Cloud assets have ugly names — often just a long stream of gibberish assigned by a robot, to make it easy for other robots. You’ll need your own robot interpreter to even identify one asset, let alone track it as it moves and changes. The nature of the cloud is highly dynamic — instances are spun up and killed on demand, and they move far faster than, say, a classic vulnerability scanner can keep up with. If you want to see your final as-built infrastructure (and you need to, since this is what your adversary is looking at too), you need software to keep up with all the changes, track the assets, and untangle the myriad ways that cloud assets are marked. There are tags, there are labels, there are unique ID’s, and there are security groups. Every vendor has subtly different rules, and just to add to the confusion cloud vendors don’t even agree on what a cloud network should be called, but they all offer the same idea.

At the end of the day, security is about adapting and keeping up, as the pace of change keeps speeding up. Cloud is just the latest evolution, where names change, details shift, but the core principles remain — first and strongest of all is inventory. This is why we at RedSeal build software to automate all the communication and mapping, so that you can visually scan your cloud footprint, understand your security posture, and make optimal moves to increase your security and reduce your risk.

For more information, check out our overview of RedSeal Stratus Maps and Inventory capabilities to learn more about how you can Map Your AWS Infrastructure Including Connectivity Paths.

Five Steps to Improve your Multi-Cloud Security

In 2021, the COVID-19 pandemic had a dramatic impact on how and where we do business. For many enterprises, the “where” became the cloud – immediately. This rapid adoption of the cloud – in most cases multiple clouds – created a rapid increase in security issues. Suddenly, enterprises had new cloud security requirements they needed to understand and deploy without the benefit of time to learn. The complexity continued to increase, and this triggered new security issues with potentially costly consequences. These included:

  • Data leakage/exfiltration – Unauthorized movement of sensitive data from inside the enterprise to outside can be accidental or deliberate. Often the discovery that data has been leaked occurs days, weeks, or months later, and can result in a damaged brand, lost customer trust, and fines.
  • Ransomware – Enterprises can pay thousands to millions of dollars to access encrypted data and systems in order to restore operations. Additionally they can be extorted to pay for the recovery of stolen sensitive information.  If they refuse to pay,  enterprises can lose days or weeks of revenue trying to recover their systems, and risk having sensitive data posted on the internet.
  • Non-compliance – Enterprises not adhering to mandatory regulations (PCI-DSS, CMMC, HIPAA) or voluntary cybersecurity frameworks (NIST, GDPR) can incur costly penalties and potential shutdowns that limit their ability to conduct business. Customer relationships may be damaged by the perception that security isn’t a priority.
  • Team collaboration/staffing shortages – DevOps is highly distributed across the enterprise and many teams acknowledge the lack of cloud platform security expertise. Cloud security practices should encourage significant collaboration that leverages both internal and external expertise.

To maintain cloud security and reduce–if not totally eliminate–the impact of these serious security issues, enterprises need a proven cybersecurity framework to address these issue directly.

Steps to strengthen your cloud security

Cloud environments are dynamic and constantly evolving. These 5 steps provide a proven framework to improve your enterprise’s cloud security using a technology driven approach, even in a multi-cloud environment.

  1. Visualize/maintain an accurate inventory of compute, storage and network functions
    Security teams often lack visibility across multi-cloud and hybrid environments. Cloud environments are often managed in disparate consoles in tabular forms. Security teams need to understand controls that filter traffic, including cloud native controls (network security groups and NACLs), and third-party infrastructure (SASE, SD-WAN and third-party firewalls). A single solution that provides a detailed visual representation of the multi-cloud environment is critical.
  2. Continuously monitor for exposed resources
    It is important to understand which cloud resources are publicly accessible or Internet-facing. Unintentional exposure of resources to the Internet is a major cause of cloud breaches. This includes any data resources like AWS S3 buckets or AWS EC2 instances. Security teams need to easily identify and report on exposed resources, and then provide remediation options that include changes to security groups or firewall policy.
  3. Continuously validate against industry best practices
    There are many industry best practice frameworks that can be used to validate cloud security. CIS Benchmarks and Cloud Security Alliance are two of these frameworks. Security teams should continuously validate adherence to best practices and quickly remediate findings to eliminate misconfigurations and avoid excessive permissions.
  4. Validate policies – segmentation within/across clouds and corporate mandates
    Many security teams create segmentation policies to minimize attack service and reduce the risk of lateral movement. Examples may be segmenting one Cloud Service Provider from another (AWS cannot talk to Azure) or segmenting access across accounts in the same CSP. Both segmentation and corporate policies should be continuously monitored for violations and provide detailed information that enables rapid remediation.
  5. Conduct comprehensive vulnerability prioritization
    All vulnerability management solutions provide a severity score, but more comprehensive prioritization can occur by identifying which vulnerabilities in the cloud are Internet-facing (including the downstream impact of these vulnerabilities).

Implementing success

While the risks grew for many enterprises this past year as they rapidly moved to the cloud, several have dodged the bullet. RedSeal has helped many successfully adopt a strong security framework and gained actionable insights into their cloud environments. These insights were often an eye-opener.

  • Underestimated VPC[1] inventory in the cloud – A healthcare customer expected “a few VPCs” in their cloud environment. The implementation of RedSeal revealed they had over 200 VPCs. This helped them see their overall cloud footprint and reduced their attack surface.
  • Exposed cloud resources– An enterprise customer incorrectly believed that all of their cloud resources were protected by a third-party firewall. Consequently, many resources were directly exposed to the Internet. RedSeal identified the exposed resources and the misconfigurations before any exploitation occurred.
  • Risky shadow IT – A technology company’s business unit had cloud instances that did not pass the company’s access security mandate. RedSeal identified these resources and helped determine that employees had bypassed process and created unauthorized cloud resources. The company’s shadow IT with respect to cloud security is now under control.
  • Zone-based segmentation as required by PCI-DSS – A payment card provider validated that card holder data was segregated and protected after their cloud migration. They modeled and monitored their segmentation policy, enabling their audit to be completed quickly and confidently.
  • VPC/VNET without subnets or subnets without instances – A healthcare customer discovered 100s of empty VPC/VNET subnets and subnets without instances in their cloud environment. The default configuration: “ANY/ANY” could have been easily exploited by malicious actors and industry best practices indicate they should be deleted or actively monitored.

 

With RedSeal, all these enterprises, and more, have utilized a multi-cloud security methodology that highlights: Visualization/Inventory, Exposure, Industry Best Practices, Policy Validation, and Vulnerability Prioritization. These 5 steps can bring peace of mind to security teams who have had to act quickly and without warning in response to this most unprecedented year.

Learn More

Looking for more details on how 3rd party firewalls may impact your cloud security framework? Download our whitepaper “How Should I Secure My Cloud?

RedSeal’s Cloud Security Solution -Ensure Your Critical Cloud Resources Aren’t Exposed to the Internet

[1] AWS uses the term VPC (Virtual Private Cloud) and Azure uses the term VNet (Virtual Network). Conceptually, they provide the bedrock for provisioning resources and services in the cloud. However, there is variability in implementation.

Avoiding Storms While Transitioning to the Cloud

SIGNAL Magazine | April 9, 2018

By Wayne Lloyd, RedSeal Federal CTO

From an industry perspective there are many advantages to moving aspects of any organization to the cloud. In theory, cloud is more efficient and easier to manage, but organizations like the Defense Department need to make sure they are not bringing along their bad habits and old baggage with them. Legacy networks are hard to understand and have grown out of control in the last few decades. Cloud is as complex as legacy networks, but the difference is who or what is really maintaining them.

Using Pizza To Understand The Cloud

Forbes | April 4, 2018

By Dr. Mike Lloyd, RedSeal CTO

It’s a tech evangelist’s worst nightmare. I was forced to explain something complex to a non-technical audience who would rather be doing almost anything else. I found myself in front of a sales force while they were in a vacation mood — possibly involving alcohol. We reward our sales overachievers with a vacation. It’s mostly focused on celebrating their success, but with some light company business thrown in. I was the speaker for a late afternoon session, on the topic of the cloud — and the next item on the agenda was the bar.

My assigned topic was cloud networks — a topic familiar to all, but still fuzzy, just like real clouds. It’s been several years since the famed survey that showed people thought bad weather was a problem for cloud computing.

How Does the Cloud and Mobility Change Things?

I remember sitting in a data center deep in an IBM facility in the early 1990s typing access control into a Proteon router that we had installed for our first commercial Internet link at that site. The controls were rudimentary, and severely limited access from outside. No one but I could access most of the connected systems, and very few people even knew that they existed. Few cared. Who wanted access from the Internet, anyway?

Fast forward to today when many people carry the Internet in their pocket. Computational and storage resources are available for pennies from many different cloud providers, and virtually everyone walking into an enterprise facility is carrying a powerful computer capable of connecting to both the Internet and any wireless network within the facility.

How does this change the game?

factoring-in-the-cloudFor one thing, it makes the overall attack surface much larger. That surface now includes all of the wireless networks within your network plus all of the various avenues into any of your public or hybrid cloud infrastructures. This means that knowing the attack surface is critical.

For another, the access controls created must take into account this new set of potential attacks, including source addresses–whether spoofed or not–that may include addresses that are legal within the organization.

Taking that entire set into account and following potential resulting access from outside the organization through all potential paths in the network (including any potential access that would result from legal changes to routing based on either load or lost interfaces) is challenging.

Making sure that necessary, business-critical access is open, while also making sure any unnecessary, potentially dangerous access is blocked, is just as challenging.

On top of this work, being sure that you’ve done all of this in the way you intend, that you maintain it over time with clean, current configurations and documentation, and that you are able to report and determine any changes, is one of the core aspects of managing this ever-more-complex situation going forward.