Gilad Parann-Nissany Founder and CEO Porticor

Cloud computing security issues are constantly a top concern for IT leaders migrating to the cloud. There are many issues related to data security in the cloud and more than one approach to cloud security. Focusing on Infrastructure as a Service cloud security, there are five issues that repeatedly concern customers that are resolved by implementing the best practices discussed at a high level below. Cloud security best practices are technology related, but also focuses on P3: Process, People, Products.

1.     Choose your cloud wisely
Infrastructure clouds come in many variations. Some are big (like Amazon Web Services, Microsoft Azure, Google, or HP) and some are smaller but more focused on specific needs such as addressing compliance (Firehost and Layered Technologies are two examples, but there are many more).

A cross-platform concern in every Infrastructure as a Service (IaaS) deployment is that data security is a shared responsibility. When shortlisting cloud providers, make sure specific certification such as ISO 27001 or SOC3 are in place. If you have specific regulatory concerns such as HIPAA safe harbor or PCI DSS compliance, ensure your cloud provider can support these specific requirements. Ask to speak with customers with a similar use case and similar size (or bigger). Learn from their lessons and make a decision accordingly. 

2.     Encrypt your data
As we’ve written before, encryption becomes your virtual walls in a cloud deployment. In your datacenter, your physical servers are protected by the 4 walls and the tight access security policy. In a shared cloud infrastructure, those measures are basically nonexistent.

This is where data encryption steps in. Data encryption allows you to segregate and isolate your environment from other companies (or adversaries) running on the same infrastructure. Data encryption in the cloud takes multiple forms: some more secure than others.

Freeware encryption tools might seem attractive at first, but they have two major issues:

1. They don’t scale well
2. In many cases, the encryption key is stored on the virtual disk along with the encrypted data, which renders the entire solution useless.

In a compliance use case, these drawbacks might pose serious issues.

Research and test more than one encryption solution, and learn carefully about the security posture of the encryption provider.

3.     Focus on encryption keys
Although it sounds strange at first, cloud encryption can be easily achieved. The challenge lies=s not with the actual encryption, but with the encryption keys.

The Heartbleed bug, discovered a few weeks ago, exposed a weakness in Open SSL’s SSL/TLS protocol, allowing anyone on the Internet to read the memory of the systems protected by the vulnerable versions of the OpenSSL software. As a result, many encrypted servers in the cloud unknowingly exposed their encryption key, which resided in memory of a server impacted by heartbleed.

To mitigate with such sophisticated attack vectors, the encryption keys should be secured throughout their life cycle: while in the key management system and while in use in the cloud. Emerging technologies such as split-key encryption and homomorphic key management can be used to avoid such attacks.

When researching encryption and key management technologies, look for cloud-enabled innovative technologies, and verify how encryption keys are managed and secure throughout the life cycle of the key usage.

4.     Automate cloud security as much as possible
One of the clouds’ most important benefits is its ability to automatically scale an infrastructure environment up or down, in a single geography or across multiple geographies. When it comes to cloud security, the paradigm shifts. You’ll hear experts telling you about the need to “keep security under your control,” and that security automation means sacrificing trust.

While true for traditional security systems, new – cloud based – security solutions do enable automation using secure, RESTful API tools.

Automating data security actually reduces risk and configuration mistakes. Assuming the software vendor can prove automation is done securely, it is a best practice for IaaS cloud security.

5.     Train your employees
Implementing the latest and greatest security toys is fun. Training employees may not seem as exciting. Yet, in many cases, a trained employee will be more efficient in stopping an attack than most technologies. (Art Gilliland gave a great pitch on “defense in depth” during RSA 2014 - see the video here). 

A common modern attack pattern would start by identifying and infiltrating privileged users’ accounts (such as database administrator, or system administrators), and once access is gained, getting to end user data stored on those databases becomes a much simpler task for the attacker.

By educating users on risks and security best practices, the access of the attacker can be avoided. In addition, the cloud brings additional potential attack vectors, such as disk snapshots, or identity theft to the online portal, managing all servers. When training employees, always keep cloud in mind, together with your business tools and processes.

Cloud Security Best Practices Lead to Successful Deployments
There are many considerations surrounding cloud security. These best practices do not eliminate the risks or remove the need to always be kept abreast of the latest development. They do, however, ensure that your cloud will be more secure and enable you to comply with laws and industry regulations while taking advantage of the many business benefits of the cloud. Much of this, when explored from such a high level seems like common sense, and yet, the news is filled with stories of companies large and small who failed to properly manage their encryption keys or permitted their employees or vendors to enable access by attackers (a la recent breaches at Ebay and Target). Implementing these best practices will ensure that yours isn’t listed among the news-making breaches.