The Doppler Quarterly Fall 2019 | Page 46

to deliver than those coming from the current human con- sultancy method. Similarly, a radiologist who needs 13 years of high-level education to read and diagnose diseases from X-Ray images may soon be replaced by someone with far less formal education who can leverage ML technology. For companies, the question is shifting from should they adopt these new technologies, to how they can adopt them more quickly. The public cloud offerings from AWS, Azure and Google offer great platforms for a company to start its technology transition. There is in fact a rush to this transi- tion, as evidenced by the massive growth in public cloud migration. This, however, has led to new potential security problems resulting from improperly planning and executing the cloud migration. Major Data Breaches When it comes to cybersecurity, there are several areas that need attention. One that stands out is the ability to protect your cloud identity and access management (IAM) system. The IAM system is a strong control and can protect your data and resources when properly managed. If improperly managed, there can be severe data losses. Here are a few examples of data breaches caused mainly by poor IAM management: • 2013: The NSA/Snowden data breach, in which an insider who had full access to many internal systems, shared top secret data with other organizations. A lengthy forensic investigation ended without a solid conclusion due to the lack of auditing records during the events. Snowden would have had a hard time executing the breach if the NSA had enforced audit- ing controls and proper IAM access control with multi-factor authentication (MFA). • 2014: Codespaces.com went out of business, when its cloud services and data hosted on AWS was hijacked and destroyed by the hacker because of an unfulfilled ransom request. This is another case of no MFA enforcement. • 2017: The Equifax data breach, in which attackers accessed a database that contained unencrypted cre- dentials that they then used to access other internal databases, resulting in a leak of the records of an 44 | THE DOPPLER | FALL 2019 estimated 147 million people. This is also a case of privileged accounts without MFA. Enabling MFA could have prevented the hacker from accessing other systems by merely using stolen passwords. • 2019: Facebook’s 540 million-user data breach, in which user accounts were stored in an open access S3 bucket by third parties who had access to the data. If Facebook had followed AWS S3 bucket policy for access control best practices, the breach would never have happened. AWS policy stipulates the S3 bucket should not have open access to everyone. An S3 bucket that stores data must have a policy that limits access only to its rightful owners. Golden Keys to Cloud Kingdoms One of the benefits of the cloud is its ability to help your organization deploy your cloud applications quickly, and store data redundantly. Agility is the goal, but to achieve this, your application, operations and security teams may need the next level of training and preparation to adapt to the nature of cloud services. However, typical early cloud adopters tended to start with a proof of concept (PoC) by someone in the organization. The PoC would then turn into something larger, with applications and data. But there would be no formal enterprise process to strategize and assign proper access control policies to the resources in the cloud. It is often the case that a new cloud deployment is fully owned by a small group of people, if not just one or two, who have full access to all cloud resources. These peo- ple hold the golden keys to your cloud kingdoms. DevOps and DevSecOps DevOps and DevSecOps are the new models of rapid devel- opment and innovation in the cloud. These models demand that software developers either transform into full-stack developers or work very closely with operations (DevOps) or security and operations staff (DevSecOps). The end goal is that developers and their supporting operations and security counterparts will join forces to create some auto- mation scripts to quickly and securely deploy applications, along with infrastructure and tools, in the cloud. Sometimes the new cloud-aware developers prefer to work and control their cloud applications and infrastructure independently