Modern Security Operations aka Secure DevOps - All Day DevOps 2017

Transcript

1. October 24, 2017 Modern Security Operations aka Secure DevOps Madhu Akula Automation Ninja @ Appsecco
2. October 24, 2017 About Me • Automation Ninja at Appsecco • Interested in Security, DevOps and Cloud • Speaker & Trainer : Defcon, All Day DevOps, DevSecCon, c0c0n, null, etc. • Discovered security vulnerabilities in Google, Microsoft, Yahoo, Adobe, etc. • Never ending learner • Follow me (or) Tweet to me @madhuakula
3. October 24, 2017 Modern Security Operations • To improve collaboration between Developers, Operations and Security • Applying security into each phase of DevOps lifecycle • Practice of developing and deploying safer software sooner • Building secure defaults and following best practices • Proactive monitoring & defence • Performing redteam activities, before real attacks happen • Learning & sharing with community
4. October 24, 2017 What is DevOps? There are many definitions for this term. I personally follow CAMS by Damon Edwards and John Willis ‘Implementing a culture of sharing between Development and Operations’ • Culture • Automation • Measurement • Sharing
5. October 24, 2017 DevOps lifecycle Test Monitor Deploy Code Plan
6. October 24, 2017 Let’s talk about some DevOops highlights
7. October 24, 2017 Security Misconfiguration Source: https://www.upguard.com/breaches/cloud-leak-accenture
8. October 24, 2017 Components with known security vulnerabilities Source: https://github.com/blog/2447-a-more-connected-universe
9. October 24, 2017 Insecure Defaults Source: https://hackernoon.com/capturing-all-the-flags-in-bsidessf-ctf-by-pwning-our-infrastructure-3570b99b4dd0
10. October 24, 2017 Secret keys in public github Source: https://www.theregister.co.uk/2015/01/06/dev_blunder_shows_github_crawling_with_keyslurping_bots
11. October 24, 2017 Unauthorised access Source: https://www.shodan.io/report/nlrw9g59
12. October 24, 2017
13. October 24, 2017 DevSecOps
14. October 24, 2017 What is DevSecops?
15. October 24, 2017 DevSecOps moto “The purpose and intent of DevSecOps is to build on the mindset that “everyone is responsible for security” with the goal of safely distributing security decisions at speed and scale to those who hold the highest level of context without sacrificing the safety required.” Source: http://www.devsecops.org/blog/2015/2/15/what-is-devsecops
16. October 24, 2017 Let’s fit security into DevOps lifecycle
17. October 24, 2017 Thinking about security from the outset “Companies that consider security from the start assess their options and make reasonable choices based on the nature of their business and the sensitivity of the information involved. Threats to data may transform over time, but the fundamentals of sound security remain constant.” - Federal Trade Commision Source: https://www.ftc.gov/system/files/documents/plain-language/pdf0205-startwithsecurity.pdf
18. October 24, 2017 Planning • The most important phase of development is planning • Involve all parties (Dev, Sec, Ops) from the beginning, it will enable everyone to understand and speed up things without compromising quality & security • Build things with the mindset of secure defaults with built-in security
19. October 24, 2017 Planning • This applies to all teams ◦ Developers need to think about secure coding best practices, using secure libraries and keeping up to date with latest vulnerabilities. ◦ Operations teams need to be aware of technology specific security configurations, best practices and hardening guidelines. ◦ Security teams have to understand the workflow, create suitable standards and apply them throughout the lifecycle
20. October 24, 2017 • Version control gives the power of moving traditional operations to modern DevOps shops • Managing things will be super-easy and efficient • This way everything can be audited, tracked and can be rolled back if required Version control
21. October 24, 2017 • Some of the things that can be version controlled include ◦ Documentation, knowledge bases, etc ◦ Developers’ code ◦ Op’s configurations and playbooks ◦ Custom scripts and snippets ◦ Many more... Version control
22. October 24, 2017 • Infrastructure as code aims to make operations more efficient and remove human errors • By doing this, we can achieve ◦ Version controlled and codified versions of secure infrastructure ◦ We can perform continuous integration with the deployment process ◦ We can improve the inventory by building Configuration Management DataBases (CMDB) Infrastructure as code
23. October 24, 2017 • This requires process and tools ◦ Identifying the all manual repetitive tasks and structuring them for automation ◦ With tools like Ansible, Chef, Terraform, etc. • We can validate our infrastructure as code against security & compliance • We can create security playbooks for hardening & patching Infrastructure as code
24. October 24, 2017 Ansible playbook snippet for MySQL hardening - name: Secures the MySQL root user mysql_user: user: root password: "{{ mysql_root_password }}" host: "{{ item }}" login_password: "{{ mysql_root_password }}" login_user: root with_items: - 127.0.0.1 - localhost - ::1 - "{{ ansible_fqdn }}" - name: Removes the MySQL test database mysql_db: db: test state: absent login_password: "{{ mysql_root_password }}" login_user: root
25. October 24, 2017 • The practice of integrating work frequently, which requires quick verification to process next steps using automated build processes • In this phase you include your test cases and security checks, performing them before going to production • It allows us to integrate existing tool sets using web hooks and plugins into the build process Continuous everything
26. October 24, 2017 • Ensures that the quality of the code and configurations remains the same by using automated test cases and validation checks • This requires defining the steps each team needs to perform to speed up the delivery process without compromising security Continuous everything
27. October 24, 2017 • Deployment is the phase where things are made live; aka production • Using a standard baseline-OS and containers, which reduces the level of security risk • Hardening configuration and environments with best practice and against known vulnerabilities (Using your security playbooks) Secure deployments
28. October 24, 2017 • Managing secrets and data is a key part while deploying to production, use secure communication channels and storage like Vault • Verify deployments by running security scans against them for misconfigurations • Also using modern tool-set like Moby project, LinuxKit, etc. for docker containers Secure deployments
29. October 24, 2017 • To make an important decisions (or) to troubleshoot things, monitoring is the place to start • Monitoring needs to apply to every phase of the DevOps lifecycle • Health checks of applications & infrastructure to know how things are going • Security monitoring of applications, servers, network devices Proactive monitoring & alerting
30. October 24, 2017 • Alerting based on thresholds and attack anomalies • Fine-tuning and improving the alerting system gives more control • Automating actions against known repetitive alerts can be efficient, but take care Proactive monitoring & alerting
31. October 24, 2017 • Define baseline security ◦ Test against it ◦ And run tests continuously • Drive testing from the DevOps pipeline • Never deploy sub-standard code • Requires tests to be passed in order to deploy into production • Empower DevOps teams to fix issues • Apply feedback loops Test driven security Source: https://docs.google.com/presentation/d/1H0Ym0bJ4TgVz7BRkoeun7ndGuTIteVyFGDDxguXmff0
32. October 24, 2017 • This requires you to have proactive monitoring in place, which includes building centralised logging and monitoring systems • Build your defences from an offensive mindset and start by focusing on your critical infrastructure • Enable DevOps teams to better understand and identify what security attacks look like by red teaming (we can also use this log data to train defence systems) Attack driven defence Source: https://www.slideshare.net/zanelackey/attackdriven-defense
33. October 24, 2017 • Identify patterns and anomalies for alerting and take action against them using automated defence • Apply data science and machine learning techniques for data sets • Build defence systems with real attack data and defend like an attacker Attack driven defence Source: https://www.slideshare.net/zanelackey/attackdriven-defense
34. October 24, 2017 • Clear communication enables us to be more productive • Collaboration between teams makes things faster. It should start from outset! • Break requirements into actionable items and assign them to respective teams • Eliminate the barriers between Devs, Ops and Security teams and work towards a DevSecOps approach (everyone is responsible for security) Communication & collaboration
35. October 24, 2017 • Use task and project management tools for collaboration, this will help showcase dependencies between teams • Spread awareness of different roles and skills by conducting social events; learning by lunch, etc. Communication & collaboration
36. October 24, 2017 Training people Training developers and operations about how attackers work, by using vulnerable labs and applications, will give them a better understanding OWASP Vulnerable Web Applications Directory Project
37. October 24, 2017 • We must learn from each other, the best way to do this is sharing with others ◦ For example, security teams can write a playbook to harden infrastructure to meet the policies and standards rather pointing out that it’s an ops issue • Rather than working as a big teams, we can mix the different teams into smaller groups and work together to achieve great results Culture & innovation
38. October 24, 2017 • Simplicity, documentation and clear communication is a win-win • When things go wrong transparency and open contributions is vital • Attending conferences and meetups and being part of the community helps us to know how the world is doing things differently to us Culture & innovation
39. October 24, 2017 • Fail fast and early, so there is less cost and damage for the business • Maintain secure backups and validate the restore process • Test for resiliency and recoverability using tools like chaos monkey and security monkey • Conduct internal hackathons and bug bounty programs • Perform redteam activities, simulate how real attacks happen Think about failures, before they occur
40. October 24, 2017 Wardly maps for DevSecOps Source: https://github.com/devsecops/wardley-maps
41. October 24, 2017 Demo Time Code to Production
42. October 24, 2017 https://www.youtube.com/watch?v=y9Usd0Q2Il0
43. October 24, 2017 What did we see?
44. October 24, 2017 Takeaways • Everyone is responsible for security (Dev + Sec + Ops) • Clear communication, active collaboration is key to success • Build with secure defaults mindset • Test driven development & Attack driven defence • Hack your applications, infra, etc. like real attackers • Keep learning and sharing
45. October 24, 2017 References • https://www.devsecops.org • http://www.devseccon.com/devsecops-whitepaper • https://pages.cloudpassage.com/sans-a-devsecops-playboo k.html • https://devops-security-checklist.sqreen.io
46. October 24, 2017 Session Title Your Name Your Title Your Company Your @TwitterHandle
47. October 24, 2017 Session Title Your Name Your Title Your Company Your @TwitterHandle
48. October 24, 2017 bit.ly/addo-slack Find me on slack, right now!
49. October 24, 2017 Thank You @madhuakula | @appseccouk