Implementing Continuous Security through DevOps

DevOps can be defined as a cultural change in work or a set of technology practices. It focuses on building a quality code, enabling automated testing based on a culture of continuous improvement eventually leading to improved stability and throughput and moving new features to customers quickly. Although DevOps is not a particular set of tools, but implementing DevOps practices leads to using certain tool sets.

-      The Version control systems track changes to the files and allow collaboration between teams. This leads to comparisons and merging of versions and rollbacks of issues. 

-      The Configuration management system automates new systems, enforces consistent application installation and configuration of system and application across different classes of servers. 

-      The ‘infrastructure as  code’ can be versioned and tested, ensuring that identical configurations are in place. This improves the odds that the software that was tested fine in the staging system will be fine in the production system as well. 

-      The automated system which helps to move through the build, deploy, test and release phase in the key component. 

Apart from the efficiency provided by DevOps, it can also create challenges in the implementation of security control infrastructures. Deployment through cloud reduces the control over the visibility of hardware and network layers. Further, it also complicates the tracking of hardware assets over time. DevOps blurs the limitations of the developers and operations and their duties are not well segregated.

Security teams need to be engaged early in the DevOps process in order to ensure continuous deployment. The following critical controls may be enforced in order to reduce potential security risks.

1. Inventory of Authorized and Unauthorized Devices: In DevOps environment, the very idea of ‘devices’ and servers is obscured in layers of deployed containers and virtual machines. Hence, cloud provider portals and APIs can provide verification of automated inventory along with cloud assets as well.
2. Inventory of Authorized and Unauthorized Software: The common practices of DevOps limit the servers to an approved list of the installed software. The configuration management tools can also be used to restrict to block software versions with known vulnerabilities.
3. Secure configurations for Hardware and Software on Servers: Installing and running only the required softwares, keeping them updated and configured is one of the best ways minimizing attacks. Once the configurations for OS and applications are developed, DevOps greatly simplifies the process of syncing these configurations throughout the system.
4. Continuous Assessment of vulnerability and Remediation: Keeping up with new vulnerabilities is a challenge. However, the DevOps environment provides a strong foundation for testing new patches. Deployment related security scans verify that all updates are addressed and reaches all intended targets.
5. Security of Application Software: The DevOps automated deployment pipeline performs  the code review, static analysis and web application scanning before the new software is deployed in production. Further, security testing such as security related functionality, vulnerability scanning and application security scans can be run in parallel to acceptance testing within the staging environment.
6. Controlled use of Administrative privileges: In a DevOps model, the code itself acts as a privileged user. Administrative privileges are used by configuration management that install new software, make configuration changes as per events and alerts. These credential secrets must only be used by orchestration systems and should not be made available to any unauthorized access.

As the DevOps movement matures,  security can no longer be considered as an afterthought, rather the best security practices need to be implemented into the DevOps evolution. Recently, there has been a considerable development of tools which help in securing DevOps environment. They range from repository firewalls, new application scanners, security functional test infrastructures to new SSH management solutions.

Conclusion

DevOps practices come with both advantages as well as disadvantages when Critical control based control infrastructure is implemented. The DevOps team needs to implement security early in the process to ensure a seamless deployment. The new security tools that is enforced for DevOps provides a new level of visibility and automation for security control implementations.

IoT Devices – Areas of Security concern

The Internet of Things(IoT) is actually much bigger than anyone realizes. IoT revolves around machine to machine communication. It is built on networks of sensors gathering data. These are virtual, mobile with instantaneous connectivity. The real value of the Internet Of Things is in the collection of data and leveraging it. The cloud based applications help in the interpretation and transmission of data from these sensors. 

Of all the current technology trends going on today, the Internet Of Things is the biggest one and hence it will also be the biggest one to give the most disruption. With IP and personal data accessible through connected devices,  hackers have the potential to bring an organization or  even a government to a standstill. Hence, we need to start with building new security approaches. Continuous monitoring, threat mitigation, secure operating systems are a few approaches to start with. 

Some of the aspects for SecurityTesting of IoT devices are as follows:

 -          Web Interface security: Most of our devices now have an inbuilt Web Server. Hence, cover the following points in this regard:

- Clear default login credentials while the initial setup

- Ensure complex passwords

- Check for Cross Side Scripting

- Check for SQL Injection

- Check for vulnerabilities of Cross Site Request Forgery(CSRF)

-          Authorization and Authentication: Authentication is much weaker in smart devices. Often limited to four digit codes. Check for weak passwords during the initial installation, use client side Java codes, send for authentication without using HTTPS transports or ask for no password at all. 

-          Network Services: The IoT devices mostly use insecure services like Telnet, FTP, TFTP etc. Penetration testing tools like Nessus, OpenVAS can check the use of such dangerous services.

-          Privacy Concerns: Three areas of concern that need to be covered here are

-  Ensuring collection of minimal data.

- Ensuring encryption of data

- Ensuring protection of data.

-          Transport Encryption: Failure in Transport Encryption exposes all the data and credentials at the same level of risk as an insecure web application. Hence, this aspect needs to be covered for complete security testing.

-          Mobile Interface: IoT devices can also act as wireless access devices. Here security specialists lack in a concrete security checking methodology. 

-          Cloud Interface: Most of the IoTdevices need to connect to the cloud server. These web services, may carry certain vulnerabilities. Hence, focus should be put on situations such as username harvesting, no lockout after brute force guessing attempts etc.

-          Security Configuration: This generally involves features such as password enforcement, data encryption and access through different levels. One additional aspect is to check the multiple user level access(full administrator/root permissions) of the operating system in use. Privilege Escalation attacks need to be attempted if they exist in the device.

-          Software Security: Two main threats to data sent over the network is that it could be changed and sensitive data can be intercepted. To cover these two contexts of insecurity ensure cryptographic signature for all updates, use of only the HTTPS ports and a cryptographic identity of the server provided.

-          Physical Security: Five things to over for ensuring physical security of these devices are encryption of stored data, physical protection of the USB ports, ease of dissembling and removing of unnecessary ports and ease of storage media removal.

Conclusion:

According to recent studies, majority of the IoT devices have security vulnerabilities. With millions of new smart devices, hardware endpoints, innumerable lines of coding and more complex infrastructure to cope up with the load, an extensive set of challenges has been created. Instead of researching on testing techniques, a clear mandatory emphasis on security from day one is a better approach, especially when dealing with such immature miniature technologies.  

Security Testing in Software Development Lify Cycle.

Security Testing ensures the confidentiality of the data. It can be started right from the requirement gathering phase of the Software Development Life Cycle. This can verify the fact that any intentional or unintentional actions are not able to stop or delay the system. The complexity of security testing is different in different phases. 

1. Requirement Gathering:In this phase, we need to determine the level of security that the system may require while we assess the requirements. The level of security can vary through government level, bank level, e-commerce level, simple information level confidentiality, etc. Another consideration is the effort that can be invested with respect to security testing.
2. Design Phase: Here, a detailed plan for Security testing needs to be created. This involves listing of the testing tools, manual approaches and the list of vulnerabilities that will be covered.

The different types of vulnerabilities can be defined as 

-      Data confidentiality

-      Authentication

-      Authorization

-      Integrity

-      Availability

-      Non-repudiation

The different types of techniques that can be applied are:

-      SQL Command Injection

-      Cross Site Scripting

-      Session Management and broken authentication

-      Insecure direct object references

-      Exposure of sensitive data

-      Security misconfiguration

-      Function level access control

-      Cross Site Request Forgery

-      Invalid redirects

-      Using vulnerable components.

- Development Phase: Here tools for static analysis can be used for security testing. This tool helps in identifying the security loop holes based on patterns and rules. Testing in this phase is the most recommended, as it involves unit level testing. Here, each code block is tested for potential bugs. These bugs if left unchecked, result in a bigger error at the later stages. Examples of such bugs 

-      Exception Handling

-      Ignored Return Values

-      Buffer Overruns

-      Complicated Functions

- Integration Phase: As the code blocks are integrated one by one to form the whole system, integration testing is performed. This is also known as black box testing. These tests are performed to eliminate security vulnerabilities before deployment. These uncovers errors that were missed out or were nonexistent in the unit level testing. It also detects configuration errors, potential security inconsistencies with third party applications and the overall system. 

- System Testing Phase: Here vulnerability scanners are used extensively. They have a predefined set of security rules, adhering to which they crawl through the entire application finally creating a comprehensive report on the system vulnerabilities. It clearly defines the cause, severity and the solution for each case.  However, the scanner should be used with a manual check as it may generate many false positives due to its pre-defined set of rules and no knowledge of the business logic.

- Deployment Phase: Here, PenetrationTesting is performed where a hacking attack is simulated on the system. This is performed to uncover all the remaining security threats. Performing penetration tests avoid network costs as well as regulatory requirements. 

- Support/Maintenance Phase: All the above testing were pre-production testing. After deployment, post production testing need to be performed. This can be done more efficiently by using a Software Security Assurance Program.

Conclusion:

Security Testing is not something which can/needs to be only performed in the production environment. It should be performed through every phase of the SDLC. As it covers all the security issues right from initiation, it also reduces the cost, time and effort for future fixings.