An overview of how to approach security in the cloud, new services and features from AWS, and a glimpse of what we do internally to secure our operations.
Action, Resource, Condition • Encryption • Renders data inaccessible without a key • Authenticated encryption protects data from modification • Easier to tightly control access to a key than the data • Independent controls for keys and data Confidentiality
How much time can your users live with zero access? • Latency of access to primary copy of the data • How much time can your users wait for normal access? Availability
Regions Availability Zones Edge Locations Client-side Data Encryption Server-side Data Encryption Network Traffic Protection Platform, Applications, Identity & Access Management Operating System, Network & Firewall Configuration Customer content Customers Security is a shared responsibility Customers are responsible for their security IN the Cloud AWS is responsible for the security OF the Cloud
it for managing account structure and permissions only • Store its credentials safely • Use AWS Identity and Access Management (IAM) for everything else • Create personal users • Use roles when possible (e.g. EC2 workloads) AWS Account Management
geographically isolated by design Data is not replicated to other AWS regions and doesn’t move unless the customer tell us to do so Customers always own their data, the ability to encrypt it, move it, and delete it DATA OWNERSHIP
AWS Identity and Access Management (IAM) made 10 AWS managed policies available that align with common job functions in organizations • IAM console now helps prevent you from accidentally deleting in-use resources
Administrator • Billing • Database Administrator • Data Scientist • Developer Power User • Network Administrator • System Administrator • Security Auditor • Support User • View-Only User • AWS Identity and Access Management (IAM) made 10 AWS managed policies available that align with common job functions in organizations
Windows Server 2008 R2, Server 2012, and Server 2012 R2 • Assessments complete even if some targeted agents are offline • Filter findings based on severity levels
management strategy Manage and use keys in AWS Key Management Service (AWS KMS) Use service-provided built-in key management Use your own key management system Manage and use keys in AWS CloudHSM
in practice • Abstraction to finite/tractable problems • Counterexample-guided abstraction refinement • Interpolation for guessing inductive invariants To learn more please reference Byron Cook’s session, please see session: SEC401 – Automated Formal Reasoning About AWS Systems
2,400 controls, but multiply that by the 64 services we have, over a period of 6 months that may be 30 million instances of control performance • We collect terabytes and terabytes of logs on our own data
Examples: event type, source IP, principal/AKID, MFA used Use data to rapidly detect and respond to threats • “Walking” credentials • Compromised accounts • Other malicious behavior Detecting anomalies through AWS CloudTrail Logs
daily deployments • Our response: automated auditable deployment and validation environment • How we use it: auditor validation of our preventative and detective change management controls • Benefit: all changes to environment and controlled and documented
91 service and feature launches for re:Invent 2016 • Leading into 2016 re:Invent (Sept-Nov 2016), AWS Security completed 139 pen-tests (equaling 2,357 person days)
the rise… • Workload migrations to software defined environments… • Mass adoption of the public cloud… • Talent migration to progressive cloud companies… • Startups have game-changing tech at their disposal… • Competitive landscape is becoming fierce… • The perimeter is no longer an option… • Security, now more than ever, is an arms race…
• Iteration over perfection • Hunting over reaction • Hmmm → Wait a minute, this sounds like a manifesto…insert shameless plug here: http://www.devsecops.org
the APIs! • Programmatically test environments • Determine state of environment at a specific point in time • Repeatable processes • Scalable operations