Guest post by AWS Community Hero Mark Nunnikhoven. Mark is the Vice President of Cloud Research at long-time APN Advanced Technology Partner Trend Micro. In addition to helping educate the AWS community about modern security and privacy, he has spearheaded Trend Micro’s launch-day support of most of the AWS security services and attended every AWS re:Invent!
Security is a pillar of the AWS Well-Architected Framework. It’s critical to the success of any workload. But it’s also often misunderstood. It’s steeped in jargon and talked about in terms of threats and fear. This has led to security getting a bad reputation. It’s often thought of as a roadblock and something to put up with.
Nothing could be further from the truth.
At its heart, cybersecurity is simple. It’s a set of processes and controls that work to make sure that whatever I’ve built works as intended… and only as intended. How do I make that happen in the AWS Cloud?
It all starts with the shared responsibility model. The model defines the line where responsibility for day-to-day operations shifts from AWS to me, the user. AWS provides the security of the cloud and I am responsible for security in the cloud. For each type of service, more and more of my responsibilities shift to AWS.
My tinfoil hat would be taken away if I didn’t mention that everyone needs to verify that AWS is holding up their end of the deal (#protip: they are and at world-class levels). This is where AWS Artifact enters the picture. It is an easy way to download the evidence that AWS is fulfilling their responsibilities under the model.
But what about my responsibilities under the model? AWS offers help there in the form of various services under the Security, Identity, & Compliance category.
The trick is understanding how all of these security services fit together to help me meet my responsibilities. Based on conversations I’ve had around the world and helping teach these services at various AWS Summits, I’ve found that grouping them into five subcategories makes things clearer: authorization, protected stores, authentication, enforcement, and visibility.
A few of these categories are already well understood.
- Authentication services help me identify my users.
- Authorization services allow me to determine what they—and other services—are allowed to do and under what conditions.
- Protected stores allow me to encrypt sensitive data and regulate access to it.
Two subcategories aren’t as well understood: enforcement and visibility. I use the services in these categories daily in my security practice and they are vital to ensuring that my apps are working as intended.
Teams struggle with how to get the most out of enforcement controls and it can be difficult to understand how to piece these together into a workable security practice. Most of these controls detect issues, essentially raising their hand when something might be wrong. To protect my deployments, I need a process to handle those detections.
By remembering the goal of ensuring that whatever I build works as intended and only as intended, I can better frame how each of these services helps me.
AWS CloudTrail logs nearly every API action in an account but mining those logs for suspicious activity is difficult. Enter Amazon GuardDuty. It continuously scours CloudTrail logs—as well as Amazon VPC flow logs and DNS logs—for threats and suspicious activity at the AWS account level.
Amazon EC2 instances have the biggest potential for security challenges as they are running a full operating system and applications written by various third parties. All that complexity added up to over 13,000 reported vulnerabilities last year. Amazon Inspector runs on-demand assessments of your instances and raises findings related to the operating system and installed applications that include recommended mitigations.
Despite starting from a locked-down state, teams often make mistakes and sometimes accidentally expose sensitive data in an Amazon S3 bucket. Amazon Macie continuously scans targeted buckets looking for sensitive information and misconfigurations. This augments additional protections like S3 Block Public Access and Trusted Advisor checks.
Each of these services support the work teams do in hardening configurations and writing quality code. They are designed to help highlight areas of concern for taking action. The challenge is prioritizing those actions.
Prioritization is where the visibility services step in. As previously mentioned, AWS Artifact provides visibility into AWS’ activities under the shared responsibility model. The new AWS Security Hub helps me understand the data generated by the other AWS security, identity, and compliance services along with data generated by key APN Partner solutions.
The goal of AWS Security Hub is to be the first stop for any security activity. All data sent to the hub is normalized in the Amazon Finding Format, which includes a standardized severity rating. This provides context for each findings and helps me determine which actions to take first.
This prioritized list of findings quickly translates in a set of responses to undertake. At first, these might be manual responses but as with anything in the AWS Cloud, automation is the key to success.
Using AWS Lambda to react to AWS Security Hub findings is a wildly successful and simple way of modernizing an approach to security. This automated workflow sits atop a pyramid of security controls:
• Core AWS security services and APN Partner solutions at the bottom
• The AWS Security Hub providing visibility in the middle
• Automation as the crown jewel on top
In this post, I described my high-level approach to security success in the AWS Cloud. This aligns directly with the AWS Well-Architected Framework and thousands of customer success stories. When you understand the shared responsibility model and the value of each service, you’re well on your way to demystifying security and building better in the AWS Cloud.