Domain 3 - Security Architecture and Engineering Flashcards
Zero Trust Security
○ Addresses limitations of legacy network perimeter-based security model
○ Treats user identity as the control plane
Assumes compromise/ breach in verifying every request, no entity is trusted by default, verify identity, manage device, manage apps, protect data
Secure defaults (secure design principles)
default configuration reflects a restrictive and conservative enforcement of security policy, taken from NIST 800-160
Fail securely (secure design principles)
default configuration reflects a restrictive and conservative enforcement of security policy, taken from NIST 800-160
Trust but verify (secure design principles)
depends on initial authentication process to gain access to the internal “secured” environment then relied on generic access control methods, taken from NIST 800-160, given way to zero trust
Privacy by design (secure design principle)
making privacy and integral part of every system, tech, policy, and design process
§ Proactive and not reactive approach
§ Privacy as the default setting
§ Privacy embedded into design, not added later
§ Privacy should be positive-sum approach, not zero-sum, needs of everybody are met
§ End to end full lifecycle data protection
§ Visibility and transparency, i.e. privacy policy explaining what company does with data
Keep privacy user-centric, i.e. in GDPR the customer can request data and tell them to “forget” them
Security-aaS
cloud provider concept in which security is provided to an org through and online entity
internet of things (IoT)
class of devices connected to internet in order to provide automation, remote control, or AI processing in home or business, plugs, thermostats, speakers, etc
SIEM
sec information and event management, collects data from many sources, provides real time monitoring, traffic analysis of potential attacks, often use AI, ML and threat intel
SOAR
sec orchestration automation and response, threat specific playbooks, response may be fully automated or single click, domain 8, delivered with SIEM typically
Microservices
fine grained services with a discrete function, more modern version of SOA to cloud computing, run on Docker/ kubernetes
SOA (service oriented architecture)
creation of discrete services that may be accessed by users in black box fashion (don’t know whats going on under the hood)
code level vulnerabilities
should be identified early in development lifecycle via static code analysis and dynamic testing to identify deficiencies before release
containerization
lightweight, granular, portable way to package apps for multiple platforms, reduces overhead of server virtualization by enabling containerized apps to run on a shared OS kernel, containers don’t have their own OS, sharing OS of host, test focuses on devops security (container level) and application level security (Authentication and Authorization)
API
set of exposed interfaces allow for program interaction between services, REST uses HTTPS for web comms to offer API end points, all comms between client and server should be encrypted and access limited with API keys, storage/ dist/ transmission of access keys should be done in secure fashion
Embedded Systems
tech component of an IOT device, full computer system embedded inside a larger system, ie printers/ drones/ semi autonomous vehicles, consider authentication practices to ensure they meet security best practices (avoid implied trust)
high performance computing
alternative to client/ server computing model for intensive operations with large data sets, for problems that require large-scale parallel processing, SETI project where individuals can volunteer their compute time i.e. grid computing
grid computing
employs a centralized controller that makes computing assignments to grid members, secure the grid controller
edge computing
some compute operations require processing activities to occur locally not in the cloud, common in IOT scenarios like agricultural, science/ space, military, ie watering plants in a field by sensing moisture, kiosk in a drug store, consider encryption, spoofing protection and authentication
fog computing
places gateway devices in the field to collect and correlate data centrally at the edge, version of edge computing
Serverless (Function as a service FaaS)
different that PaaS, more granular, less decisions around service tier and scale, azure functions and AWS Lambda
IaaS
CSP: networking, storage, servers, virtualization
You: OS, middleware, runtime, data, apps
PaaS
CSP: networking, storage, servers, virtualization, OS, middleware, runtime
You: data, apps
SaaS
CSP: networking, storage, servers, virtualization, OS, middleware, runtime, data, apps
You: n/a
Public cloud
everything runs on CSPs hardware, advantages include scalability, agility, pay as you go, no maintenance, low skills