The cloud, once touted as an IT panacea, has a flip side that we see all too often in headlines when malicious actors take advantage of gaps in security. This cannot be repeated enough: Securing data and networks in a cloud environment is very different than doing so on-premises.
Infrastructure elements that were static on-premises are now abstracted to software. Firewalls must be designed to operate in an inherently fluid infrastructure. And in the cloud, you’ll need to focus more on applications, application programming interfaces, and user roles.
Drawing from our experience with commercial and government cloud clients, here are five tips that re-emphasize and expand upon the fundamentals.
1. Control access to cloud management and configuration tools.
As cloud management and configuration tools — cloud service provider (CSP) consoles, command line interfaces, and APIs — provide end users great flexibility and autonomy, strong role-based access control is vital to protect the organization against external and insider threats.
- Authenticate and authorize privileged users with two-factor authentication, digital signatures, and certificates.
- Make training and skills evaluations mandatory before granting cloud roles.
- Strictly separate user and admin credentials, and restrict user access to production systems.
- Standardize processes for account life-cycle management.
2. Encrypt sensitive (and maybe all) data.
In the cloud, data breach and data spillage are inevitable. Protecting data in transit via techniques like encryption becomes more critical as “the very notions of data-in-motion and data-at-rest become blurred.”
Encrypt all sensitive data, and segment it using multiple keys to minimize the impact of a compromised key. Keys should rotate regularly, with strong access control policies.
- Implement encryption for data in transit and at rest.
- Review CSP network encryption. (Not all network traffic between data centers may be natively encrypted.)
- Evaluate both native cloud and third-party encryption solutions.
3. Use automation to minimize human errors due to misconfiguration.
Manual configuration eventually leads to human error and its consequences: inconsistent deployment configurations, inadvertent data spillages, and vulnerabilities to malicious activity. This is a big security risk. According to Gartner research, nearly all — 99% — of “cloud security failures will be the customer’s fault.”
Automation, with pretested and audited configurations, ensures that infrastructure is deployed and configured correctly. We recommend the following:
- Automating infrastructure and platform builds, security testing, security guardrails, and baseline configurations.
- Implementing high-availability configurations to mitigate the risks of unavailable regions or zones due to cloud provider errors.
- Periodically performing configuration checks and routine compromise assessments.
4. Adapt visibility and vulnerability management to manage ephemeral and new types of cloud assets.
With a cloud solution — as data, systems, and responsibilities spread across environments — organizations may lack sufficient visibility for monitoring threats and ensuring compliance. Furthermore, legacy methods involving human vulnerability detection and remediation management may be too slow or cumbersome in the cloud, where incidents move at warp speed across interconnected systems and data.
A move to the cloud spurs many visibility questions:
- Does your infrastructure allow for visibility across the cloud environment? For instance, is there an aggregated system of logs, dashboards, and reports for collecting data from network appliances, cloud infrastructure, operating systems, applications, and so forth?
- Do you understand your cloud vendor’s incident response service-level agreements and how they integrate into your own processes?
Organizations will need to extend vulnerability management tools to container and serverless architectures and adapt these tools for rapid infrastructure deployment cycles and new cloud services. They will also benefit from in-line services such as proxies and network overlays to replicate traffic to “transparent” security services.
5. Implement enhancements throughout the operations life cycle.
Often, the initial cloud implementation is just the beginning. As apps are introduced, IT teams will need to make continual enhancements. And traditional “stovepiped” IT operations can impede the journey to cloud agility.
DevSecOps approach can help. As you take this approach, do the following:
- Implement site reliability engineer functions so ops staff can continue to upgrade the environment.
- Integrate DevSecOps with security staff in development and operations.
- Track new products and services, with a road map for future integration.
With these five fundamentals in mind, organizations can reduce their security and compliance risk as they reap the cloud’s many benefits: cost savings, lower total cost of ownership, and faster time to value.
Chris Christou is a leader in Booz Allen’s secure cloud and IT infrastructure business. He applies expertise in cloud security, cloud infrastructure and platforms, network engineering, and IT infrastructure. Chris has more than 20 years of experience designing, testing, and … View Full Bio