On January 4th the marketing firm Octly discovered that, due to a misconfiguration in their Amazon Web Services (AWS) cloud storage, the private information of 12,000 clients was publicly available for anyone to view. This blunder revealed client names, addresses, phone numbers, email addresses, hashed passwords and more. Cloud security mistakes similar to this have become all too common in recent years and have caused both reputational and financial damage for companies. The now infamous Uber breach occurred when two hackers were able to gain entry to a third-party cloud-based service provider where sensitive driver and customer data was stored. In September 2017 it was revealed that the Pentagon had left 100 GB of classified data from a failed joint intelligence-sharing program—run by the US Army and National Security Agency—publicly accessible for years.
The explosive popularity of cloud computing has resulted in a new attack surface for cybercriminals with troves of valuable, sensitive and often poorly secured data. All of this begs the question: what risks exist for cloud data and how can they be mitigated?
The Risks of Cloud Storage
If a cybercriminal gains access to a company’s cloud storage, there’s a variety of nefarious actions they can take. Depending on the type of data that is stored, proprietary information can be sold to competitors and personally identifiable information (PII) can either be used for fraud or sold on the dark web. Similar to the Uber breach, any of this information can be used in an extortionist attempt to force a ransom payment. If the cloud storage is configured to be publicly writable, cybercriminals will also be able to edit the stored documents. This vulnerability has been dubbed GhostWriter, and it is serious, as code can be altered or data manipulated to cause damage to the company.
The Octly and Pentagon security incidents mentioned above resulted from a failure to secure Amazon Web Services S3 buckets. These storage buckets can be set to private—which require additional verification steps, or public—which can be accessed by anyone who can find the bucket on the web. The simple step of setting cloud storage buckets to private has been frequently overlooked. With the right search parameters, a basic Google search will reveal unsecured cloud storage buckets (see figure 2). The initial discovery of these buckets has resulted in the creation of open-source tools that actively search for unsecured buckets, which has made finding and stealing information a simple task for even low skilled cyber criminals. There is one major reason that so many organizations have failed to set their cloud storage to private: out of sight, out of mind. It is easy to forget about data once it has been stored and doesn’t require active use.
Encrypting data at rest adds a secondary security layer that prevents data from being deciphered even if a cybercriminal gains access to its storage location. In the case of the 2016 Uber breach, customer and driver information that should have been encrypted was left in plain text, allowing the criminals to download the information and demand a ransom for assurance that the stolen information would be deleted.
Ransomware on the cloud is a relatively new idea but MIT has predicted cloud ransomware to be one of the biggest cyber security risks for 2018. With new threats like this appearing, and cloud storage becoming increasingly more integrated into business, an increased focus needs to be placed on cloud security.
Steps to Improve Cloud Security
Some of the largest cloud breaches over the last two years could have been prevented with proper security due diligence. The first step in securing cloud data is understanding what security responsibilities are covered by the cloud service provider and what steps need to be taken separately. Using a cloud service creates a shared security responsibility between the customer company and the cloud provider. If expectations are not clearly defined from the start, security holes may develop.
The next step to establishing strong cloud security is ensuring that stored data is categorized and documented. This should be done retroactively to be sure previously stored data is well documented. As previously stated, out of sight, out of mind. Keeping organized information on data stored in the cloud decreases the chance that private information will be stored in an insecure way. Data that is meant to be private must be stored in a cloud bucket set to private (not public).
Cloud storage buckets should have randomized names. This increases the difficulty of targeted attackers locating specific buckets belonging to the targeted company. It is also prudent to avoid the use of externally facing web portals. Although not always possible, this step will decrease the available attack surface.
Companies should have a regular auditing schedule to review what groups or individuals have access to data stored on the cloud. Once this has been established, permissions should be re-evaluated based on task requirements. It is highly recommended that all administrative accounts for cloud storage require multi-factor authentication, as this implementation will decrease the likelihood of account compromise. Similarly, following security best practices for passwords is highly recommended; this involves using a strong password or passphrase and never reusing the same password for multiple accounts.
Lastly, encryption should be implemented at rest, as well as in transit, for data stored on cloud infrastructure. Encryption is the last line of defence against the sinister characters looking to pilfer data. Keeping sensitive data encrypted will minimize the effect of a breach or leak and ensure that data meant to be private remains private.