Policy & Governance Pillar
Finding the right balance of security and business enablement is a crucial requirement for any Zero Trust strategy. The primary category to help achieve this balance is the Policy & Governance pillar of Cisco’s Zero Trust Model. With the Policy & Governance pillar, an organization may establish how tightly the entire organization is governed, how long information is retained, how the organization will recover in an emergency, and how important sets of data are managed from group to group. Organizations also need to focus on their industry, their regulations, the organization and its business goals, and their customers’ risk tolerance levels. The Policy & Governance pillar focuses on key factors that must be established to enable the Zero Trust journey.
Change Control
It is necessary for many organizations to have change management services. Many use the Information Technology Infrastructure Library (ITIL) change management process. ITIL is an accepted approach to managing Information Technology services to support and enable organizations. ITIL enables organizations to deliver services. Frameworks such as ITIL help establish architectures, processes, tools, metrics, documentation, technology services, and configuration management practices.
Changes must be coordinated, managed, and details disseminated to relevant parties. A unique characteristic of Zero Trust means that changes will occur end to end within the environment, so special care must be paid to ensure smooth forward progress. As a critical part of the change process, testing provides the ability to ensure that production deployments in support of Zero Trust can be accomplished in a timely and effective fashion.
Data Governance
It’s critical to classify data and to understand where it is stored and how it is monitored for compliance to organizational policies. Some examples of data classifications are personally identifiable information (PII), Electronic Protected Health Information (ePHI), Payment Card Information (PCI), Restricted Intellectual Property, and Classified Information. Data governance also includes a well-defined and maintained configuration management database that contains where all data stores are located, who owns them within the organization, along with data classification, labeling and storage, and access requirements.
Data Retention
Data retention is dictated based on organizational and regulatory requirements. After an incident, the ability to determine the cause of an outage or breach is critical information that must be retained both for restoration of service as well as audit purposes. Data retention must consider data at rest, how long the data must be stored, and when the data should be purged to limit organizational liability. The legal and compliance teams of the organization manage policy requirements on what data an organization must retain and for how long.
Quality of Service (QoS)
Quality of service, including the marking and prioritization of key traffic in times of micro or long-term congestion, is a key component of availability to ensure that control plane traffic continues to flow to ensure Zero Trust capabilities function as intended. QoS provides for preferential treatment of traffic to meet defined policy requirements to ensure that critical functions necessary for security and business functions continue without undue impairment. Without this safeguard in place, organizations run the risk of congestion on the network having unpredictable impacts to traffic and the solutions that rely on that traffic.
Redundancy
Redundancy is necessary to maintain availability and is part of a Zero Trust strategy. Critical components of the ecosystem are required to be duplicated by many frameworks, standards, regulations, and laws. Redundancy can have multiple aspects: control plane redundancy is necessary for the functioning of capabilities, whereas data plane redundancy is necessary to ensure that business functions continue unimpeded.
Replication
Replication involves the duplication and the encryption of key data stores to backup storage arrays and offline storage backups, which provide a restoration path in the event of partial or complete loss of an environment due to ransomware. Software automation is necessary to ensure that the proper environments are replicated to proper locations. Without replication automation, errors are inevitable, and critical data stores may be overwritten, creating a large-scale outage requiring full restoration of one or many databases.
Auditors, regulators, or governing bodies routinely validate these controls. The key point to note is the regulations, standards, or laws are the minimum of replication that should be in place for the organization. Protecting an organization’s data is its top concern. Without protective controls—that is, encryption and locations around these replicated data stores—there can be no data integrity, confidentiality, and in the end availability; therefore, a gap in Zero Trust is created.
Business Continuity
Confidentiality, integrity, and availability are the foundation of all security programs and are necessary in a Zero Trust strategy. Business continuity relies on a well-executed Zero Trust strategy. The development of business continuity teams and business continuity documentation that can be accessed by the critical teams in the event of a crisis is a cornerstone to business continuity. Please note that a business continuity plan (BCP) should always be protective of human life first, in all cases. Ensure teams are safe at the outset of plan activation and throughout the event. A well-developed communication plan will assist in locating and checking in on those associated with the organization. The plan should also be protective of what is shared publicly to provide a level of protection to recovery efforts.
The second most important step is maintaining the integrity of data in the middle of responding to a business continuity event. Maintaining data integrity may seem trivial to some of those responding to a critical event, but that is exactly when an attacker will attack. Ensure “temporary” controls or measures do not expose the organization to issues with data integrity along with availability. Some ransomware attacks may activate the business continuity plan and be the cause of an organization-level outage. Restricted or intellectual property may be at risk.
Work out these scenarios in advance and partner with your nearest fusion center and other government entities to respond to these critical types of events. Tabletop exercises may expose gaps, but putting teams through BCP drills reveals how prepared teams are to respond and may uncover shortcuts that could expose the organization’s critical data stores.
Disaster Recovery (DR)
Typically, a disaster recovery event is activated as soon as a problem has been detected, but many times the business continuity plan (BCP) should be activated. After the BCP team assesses the situation, recovery efforts are officially started. The DR plan may include many of the same contacts from a leadership perspective as the BCP does, but the DR plan focuses on recovering a solution, a set of solutions, or the critical infrastructure of the organization.
The scope of any DR event may be assessed and categorized as minor, or it could go more broadly. At first, the event may impact only one aspect of the business or even one solution, but this is where teams should not have tunnel vision and should consider other systems and environments that could be also impacted. Activating the proper process and notifying the right level of leadership is a function of the business continuity plan based on impact and risk. It is important that proper criteria have been established for DR planning, primarily the criticality of the system to the organization and impact upon daily functioning and therefore the acceptable limits of data loss and recovery time. This is normally classified into two categories, recovery point objective (RPO) and recovery time objective (RTO). RPO defines the amount of time acceptable for transactional data loss. Stated another way, RPO is the amount of data or work that will be unrecoverable after a system failure. RTO, on the other hand, is the amount of time it takes to restore the system and data back to normal. These are minimum variables that should be defined for each system where it is determined that DR capability is required.
DR plans go hand in hand with the business continuity plan. With proper controls as defined in the “Policy & Governance” section of the book, disaster recovery should be achievable and complete. Development and testing of a DR plan are part of the standup procedures for new environments. Each environment must define a method of recovery prior to “production go-live” events so the definition of what constitutes a successful recovery can be worked through and can be checked off as complete during an actual DR event or during a DR test. If the plan is not created after the application has been purchased, many installation requirements are forgotten, neglected, or known by only a handful of individual team members. Testing of the DR plan is required for both new and old ecosystems. The adage still holds true: “If there is no testing, there is no DR plan.” Adding to that, without a business continuity plan and a disaster recovery plan, there cannot be a valid and implemented Zero Trust strategy.
Risk Classification
Risk classification helps inform multiple other capabilities such as data governance, business continuity, and redundancy. This includes classifying the risk for data as well as capabilities. For data, risk must be assessed to understand the criticality of the data to the organization. For capabilities, risk must be classified to understand what impact to the organization may occur if that capability ceases to operate as expected.
Risk classification structures should be developed with compliance and legal teams to ensure that the business is protected and to ensure business continuity. Having a Zero Trust mindset as these classifications are developed or updated will go a long way to provide greater protections and controls put in place, while at the same time enabling the business.