In December 2020, FireEye analysts discovered something that would become one of the biggest security scandals of the decade. The attackers — later identified as the Russian APT29 group — had been moving through SolarWinds’ internal network for months, modifying the source code of the popular Orion platform. Then, through the automatic update mechanism, the compromised software reached thousands of organizations worldwide, including US government agencies, technology giants, and companies in the defense sector. The key takeaway from the post-mortem analysis was painful and unequivocal: had SolarWinds’ internal network been properly segmented, the scope of the compromise would have been dramatically smaller. The attackers had free access to the entire environment because nothing stopped them.
The SolarWinds operation is a model example of what MITRE ATT&CK defines as Lateral Movement — a group of techniques identified as TA0008, encompassing, among others, T1021 (Remote Services), T1550 (Use Alternate Authentication Material), and T1080 (Taint Shared Content). These techniques allow an attacker who has gained a foothold at one point in the network to move further — horizontally, from system to system — until reaching a valuable target. Microsegmentation is the architectural response to this problem. It does not eliminate the possibility of the initial breach, but it drastically limits what an attacker can do once inside.
What is microsegmentation and how does it differ from traditional VLAN segmentation?
Microsegmentation is a method of creating granular, logical security boundaries within a network — down to the level of an individual workload, application, or even a process — combined with defining precise policies that specify which communication is permitted and which is blocked. Unlike traditional perimeter firewalls, which protect the boundary between the internal and external network, microsegmentation operates inside that boundary, controlling so-called east-west traffic.
Traditional VLAN-based segmentation divides the network into zones of varying trust levels: server zone, user zone, DMZ. This approach was groundbreaking in its time, but it has a fundamental limitation: communication within the same VLAN is typically unrestricted. An application server in the production VLAN can freely attempt to connect to the database server, file server, backup system, and management console — because they all belong to the same “trusted zone.” It is precisely this flatness that creates an ideal environment for lateral movement.
Microsegmentation abandons the idea of trusted zones in favor of policies based on workload identity. Each server, container, or virtual machine has precisely defined: what it can communicate with, on which ports and protocols, and in which direction. An application server may communicate with a database on port 5432 TCP, but not with the file server, not with another application server, and not with the management system — even if all of these systems physically reside in the same subnet.
Key architectural difference: VLAN segmentation creates trust zones based on network location. Microsegmentation creates policies based on identity and function — regardless of physical location or IP address.
The practical consequence of this difference is enormous in the context of security. In an environment with traditional VLAN segmentation, an attacker who compromises one server in the production VLAN potentially has access to dozens or hundreds of other servers in the same zone. In an environment with properly implemented microsegmentation, that same attacker is confined within the boundary of a single workload — they cannot move anywhere that is not covered by an explicitly permitted communication path.
Why is lateral movement a critical stage of advanced attacks?
The Lateral Movement tactic (TA0008 in MITRE ATT&CK) describes a set of methods that an attacker uses after gaining initial access to an environment. The goal is not to exploit a single system — the goal is to move through the network toward high-value resources: domain controllers, credential stores, backup systems, ERP infrastructure, and customer databases. Without the ability to move laterally, most advanced attacks end at the Initial Access stage, causing no catastrophic consequences.
The most commonly exploited lateral movement techniques include: Pass-the-Hash (T1550.002) and Pass-the-Ticket (T1550.003), which allow movement through Windows environments without knowing passwords in plaintext — stolen NTLM hashes or Kerberos tickets are sufficient. Remote Services (T1021) covers RDP, SMB, WMI, WinRM — administrative protocols that, in flat networks, are widely accessible between servers. SSH Hijacking (T1563.001) and Remote Desktop Protocol (T1021.001) provide direct interactive access to additional systems.
In organizations without active threat hunting, the median dwell time — the time an attacker spends in the network before detection — is measured in weeks. During this time, a skilled attacker is capable of traversing dozens of systems, building alternative access paths, and identifying the most valuable assets. Every day of dwell time broadens the scope of the compromise.
Lateral movement is particularly dangerous because it lives in the grey zone of normalcy. The protocols used for lateral movement — RDP, SMB, WMI, SSH — are the same protocols that administrators use for legitimate work. Automated detection systems must rely on subtle anomalies: unusual connection hours, unexpected sources, sequences of logins suggesting automated traversal. Without behavioral context and proper telemetry, such signals are easy to miss.
Microsegmentation acts as a barrier at every step of lateral movement. The attacker cannot use a protocol that is not in the policy. They cannot connect to a system with which the compromised workload has no defined relationship. Every attempt to exceed the permitted boundaries generates a detection event — and achieves nothing.
It is worth understanding that microsegmentation is not a panacea for all lateral movement techniques. If policies are too loose — for example, allowing communication between all application servers “because it is easier” — the attacker will find an allowed path. The effectiveness of microsegmentation is directly proportional to the precision and completeness of the defined policies.
How to design microsegmentation policies — from communication mapping to rules?
Designing microsegmentation policies is the most difficult and most important phase of the entire deployment. The most common mistake we observe in audits of environments with declared “microsegmentation” is policies designed on the basis of intuition or application documentation — rather than observation of actual network traffic. Applications often communicate in ways that are not documented: legacy dependencies, communication with monitoring systems, connections to external APIs, unexpected paths between components.
The starting point is mapping communication flows — the discovery phase. For at least 2–4 weeks (and ideally longer, to cover the cycle of monthly batch processes), we collect full network traffic telemetry: NetFlow/IPFIX from switches, firewall logs, telemetry from host agents. Tools such as Illumio Map Intelligence, Guardicore Centra (Akamai), or open-source solutions based on eBPF (e.g., Cilium Hubble) visualize these flows, creating an automatic dependency map.
On the dependency map, we identify workload groups — groups that naturally communicate with each other. Application servers of a given ERP system, database servers serving specific applications, backup servers, management consoles, monitoring systems. Each of these groups has its own communication profile: which ports, which directions, which sources and destinations. This profile becomes the basis for defining policies.
Microsegmentation policies are best designed according to the “deny all, permit by exception” principle — block all communication by default, then add explicit permits only for documented and justified flows. For each flow, we specify: source (workload or group label), destination (workload or group label), protocol (TCP/UDP), port or port range, direction (unidirectional or bidirectional).
A key architectural decision is the choice of workload labeling model. An approach based on IP addresses is static and fragile — in dynamic cloud environments, IP addresses change with every container restart. Modern microsegmentation systems use labels or tags: a workload has labels such as env:production, app:erp, tier:database — and policies are written in terms of these labels, not specific addresses. Such a policy is resilient to infrastructure changes and readable for security architects.
The Goldilocks principle in microsegmentation policy design: policies that are too loose do not restrict lateral movement; policies that are too restrictive disrupt production. The goal is policies granular enough to stop an attacker, but precise enough not to block any legitimate communication.
What technical approaches exist — host-based, network-based, hypervisor-based?
There are three main technical approaches to implementing microsegmentation, differing in the point of policy enforcement, granularity of control, and deployment requirements. In practice, the best environments combine elements of more than one approach.
The host-based approach involves installing a lightweight software agent directly on each protected system — physical server, virtual machine, or container. The agent implements firewall rules at the network stack level of the operating system (iptables/nftables on Linux, Windows Filtering Platform on Windows) and enforces policies locally, regardless of network topology. Key advantages include the highest granularity (policies can be tied to specific processes, not just ports), independence from network infrastructure, and excellent fit for cloud environments where traditional network devices have no visibility into communication between virtual machines. Representative products: Illumio Core, Akamai Guardicore, ColorTokens Xshield.
The network-based approach implements microsegmentation through network devices — next-generation firewalls (NGFW) or virtual switches. Traffic between segments is physically routed through a central Policy Enforcement Point. The advantage is that no agents need to be installed on protected systems, which simplifies deployment in environments with legacy systems. The disadvantage is dependency on network topology and lower granularity — policies primarily operate on IP addresses and ports, without application context. Representative products: Cisco ACI, Juniper Contrail Security, Fortinet FortiGate with security zones.
The hypervisor-based approach (or fabric-based) implements segmentation at the hypervisor or software-defined networking (SDN) level. Similar mechanisms are offered by AWS Security Groups (which are technically hypervisor-based firewalls) and Azure Network Security Groups. This approach is ideal for technologically uniform environments (pure pure cloud), but loses relevance in hybrid and multi-cloud environments.
| Criterion | Host-based | Network-based | Hypervisor-based |
|---|---|---|---|
| Granularity | Highest (process level) | Medium (IP/port) | High (VM/workload) |
| Cloud visibility | Full | Limited | Native (per platform) |
| Agent requirement | Yes (agent on host) | No | No |
| Legacy support | Yes (most OS) | Yes | Limited (requires hypervisor) |
| Container environments | Yes (eBPF, sidecar) | Limited | Limited |
For hybrid environments — with on-premises data centers, cloud, and container environments — the host-based approach is recommended as the most consistent. An agent works the same way on a physical server in a data center, on a virtual machine in AWS, and on a Kubernetes node in Azure AKS. Policy management is centralized, and enforcement is distributed and independent of network infrastructure.
How to deploy microsegmentation without disrupting production — observation mode?
The absence of a gradual, controlled deployment is the most common cause of microsegmentation project failures. Teams that attempt to enable traffic blocking immediately, without a prior observation phase, inevitably disrupt critical business processes — and end up featured in industry publications describing why “microsegmentation is dangerous.” Microsegmentation is not dangerous. What is dangerous is deployment without adequate preparation.
The observation phase — often called monitoring mode, visibility mode, or learn mode — is absolutely essential. In this mode, the microsegmentation system collects full network traffic telemetry and builds an automatic dependency map, but enforces no blocking policies. This typically lasts from 2 to 6 weeks, depending on the complexity of the environment and the cycle of business processes.
During the observation phase, we answer key questions: what network flows actually exist, which of them are expected and documented, which are unexpected and require explanation (potential anomalies or undocumented dependencies), and what ports and protocols are used by each application. Based on this, we build a policy model — a set of rules we intend to deploy.
The next step is a pilot in a non-production environment. A test or staging environment allows policy validation without risk to production. If an application operates correctly in the test environment with microsegmentation policies enabled, we have a high degree of confidence that it will also work in production.
Production deployment proceeds in stages — application by application, zone by zone. We start with low-criticality systems that have well-documented architectures. After each stage, we confirm operational stability before moving forward. In the first week after enabling blocking, we monitor denied traffic with heightened vigilance — any block that should not have occurred is a signal to correct the policy.
The golden rule of microsegmentation deployment: never enable blocking mode without at least two weeks of prior observation in that specific environment. Application documentation lies. Actual network traffic tells the truth.
An important element of safe deployment is defining a rollback path. If something critical stops working after enabling blocking, a fast procedure for restoring observation mode must exist. In practice, this means no one should enable blocking right before a weekend, on a Friday afternoon, without an on-call engineer and without a rollback procedure.
How does microsegmentation support zero trust architecture?
Zero Trust is a security philosophy based on the principle of “never trust, always verify” — no entity (user, device, workload) is trusted by default, regardless of its network location. NIST Special Publication 800-207, which is the de facto standard definition of Zero Trust Architecture (ZTA), lists microsegmentation as one of seven key implementation pillars.
In the context of zero trust, traditional network segmentation protects the perimeter, but assumes trust within that boundary. Zero Trust rejects this assumption and demands verification of every communication session — regardless of whether it originates from outside or from a trusted internal system. Microsegmentation is the technical mechanism that makes this verification possible and enforceable for east-west traffic (between workloads inside the network).
Every microsegmentation policy is effectively a declaration of trust: “workload A is authorized to communicate with workload B on protocol X.” Everything that is not explicitly permitted is rejected. This is a direct implementation of the Least Privilege principle at the network layer — a workload has no network access to anything it does not need to perform its function.
Microsegmentation strengthens zero trust by reducing the “blast radius” in the event of a workload compromise. In the zero trust model, we assume that a security breach will occur (assume breach). The question is not “if” we will be attacked, but “how much damage can an attacker cause once already inside.” Proper microsegmentation limits the answer to: very little, because they cannot move beyond what the policies permit.
Integration of microsegmentation with other zero trust pillars creates a synergy effect. By combining microsegmentation with strong identity (IAM, MFA), attribute-based access control (ABAC), continuous device posture verification (EDR), and behavioral monitoring (UEBA), we build an architecture in which every access decision — both user-to-application and application-to-application — is made dynamically based on current context.
Microsegmentation without identity context is half a solution. The ideal microsegmentation architecture combines workload identity (who is sending) with user or process identity (on whose behalf) and device posture (whether it is compliant with policy). Only this three-dimensional verification achieves true zero trust at the network layer.
It is worth emphasizing that zero trust is not a product to buy, but an architecture to build. Microsegmentation is one of the more important building blocks of this architecture — particularly in organizations where a significant portion of traffic is generated by machine-to-machine communication, not by users.
Which microsegmentation tools dominate the market?
The microsegmentation tools market is today mature and diverse. The choice of the right solution depends on the environment (on-premises, cloud, hybrid), existing infrastructure, granularity requirements, and team resources. We present the key players and their characteristics.
Illumio Core is one of the leaders in the host-based segment. The architecture based on Virtual Enforcement Node (VEN) — a lightweight agent installed on each system — provides full visibility and policy enforcement independent of network infrastructure. The Illumio Policy Compute Engine (PCE) automatically generates policy recommendations based on observed traffic. The solution performs well in hybrid and multi-cloud environments.
Akamai Guardicore Segmentation (formerly Guardicore) offers similar host-based capabilities with an emphasis on detailed visualization and a “breadcrumb trail” — the path of each attack through the environment. A particularly valued element among SOC analysts is the automatic detection of traffic anomalies against the backdrop of defined policies.
-T (NSX for vSphere / NSX Advanced Load Balancer) is the dominant solution for environments. The Distributed Firewall integrated with the hypervisor provides microsegmentation without agents on virtual machines. The Identity-based firewall (IDFW) allows policies based on Active Directory user identity, not just workload IP address.
Cisco Secure Workload (formerly Tetration) is an enterprise-grade platform with advanced behavioral analytics based on big data. Agents collect process telemetry from every workload — not only network flows, but also information about processes, connections, and application behavior.
In container and Kubernetes environments, dedicated solutions include Cilium (open-source, eBPF-based, CNCF project) with Hubble for visualization, Calico (Tigera) with advanced network policy capabilities, and native cloud mechanisms such as AWS Security Groups, Azure Network Security Groups, or GCP VPC Firewall Rules.
For organizations not yet ready for a dedicated microsegmentation platform, a starting point can be the native capabilities of operating systems: Windows Defender Firewall with Advanced Security and GPO, and Linux iptables/nftables with centralized management via Ansible or Puppet. This is a solution with lower visualization capabilities, but with zero licensing cost — and often sufficient for smaller environments.
How to manage policies in a complex hybrid environment?
The complexity of policy management grows exponentially with the scale of the environment. An organization with 50 workloads and 200 policies is manageable; an organization with 5,000 workloads and 50,000 policies — without the right tools and processes — becomes impossible to maintain. Uncontrolled policy growth (policy sprawl) is one of the greatest operational risks of a microsegmentation deployment.
The key principle is managing policies as code (Policy-as-Code). Policies are stored in a git repository, versioned, reviewed through peer review, and deployed via CI/CD pipelines. Every policy change is auditable: who made the change, when, and on what basis. In practice, this means that deploying a microsegmentation policy requires a ticket in the ticketing system, a code change, review and approval by the security owner, and only then automated deployment through the pipeline. This eliminates “temporary” rules that permanently remain in the configuration.
Managing the policy lifecycle requires regular reviews. Each quarter it is worth reviewing policies to ask: do the workloads referenced in the policy still exist? Are the communication flows still active? Are any policies unnecessarily broad and can they be narrowed? Enterprise-class microsegmentation tools (Illumio, Guardicore) offer “unused rules” dashboards — rules for which no traffic has been observed in the past 30 or 90 days. These are candidates for removal or review.
In hybrid environments — with an on-premises data center, AWS, Azure, and GCP — the challenge is consistency of policies across different platforms. Each cloud has its own network policy model. Solutions such as Illumio or Guardicore build an abstraction above these differences: policies are defined once, in a unified model, and translated into the native mechanisms of each platform. This is the so-called single pane of glass for network security — one management panel for the entire hybrid environment.
Integration with CMDB and infrastructure management tools (Terraform, Ansible, ServiceNow) enables automatic tagging of new workloads and their assignment to appropriate policy groups at the moment of provisioning. A new production server for an ERP application automatically receives the tags env:prod and app:erp, which immediately subjects it to the appropriate policies — without manual intervention. This is so-called security-by-default: new systems are protected from the first second of their existence.
The most difficult phase to manage is the “middle” stage of microsegmentation deployment — when part of the environment is already segmented and part is not. Maintaining two parallel security models for an extended period increases complexity and the risk of errors. The strategic goal is full environment coverage or the deliberate exclusion of legacy systems from scope.
What does a microsegmentation deployment roadmap look like?
Deploying microsegmentation is a multi-month or multi-year project, depending on the scale and complexity of the environment. The table below presents a typical phased structure for a medium-sized organization.
| Phase | Duration | Key Activities | Outcome |
|---|---|---|---|
| Phase 0: Inventory and Assessment | 2–4 weeks | Mapping all workloads, environment assessment (cloud, on-prem, hybrid), selecting a microsegmentation tool, defining scope | Complete environment inventory, selected technology, approved KPIs |
| Phase 1: Tool Deployment and Discovery | 4–8 weeks | Agent installation or SDN configuration, enabling observation mode, collecting traffic telemetry, building dependency map | Full network flow visibility, automatic application dependency map |
| Phase 2: Policy Design | 4–6 weeks | Dependency map analysis, defining workload groups and labels, designing policies (“deny all + permit by exception”), review by application owners | Approved microsegmentation policies for the entire scope |
| Phase 3: Pilot in Test Environment | 2–4 weeks | Deploying policies on staging/dev, application testing, policy correction based on observed blocked traffic | Validated policies without risk to production |
| Phase 4: Production Deployment (staged) | 3–6 months | Enabling blocking application by application (starting from least critical), 1–2 weeks of monitoring after each stage, escalating to critical applications | Microsegmentation active across the entire production environment |
| Phase 5: Operational Maturity | Ongoing | Policy-as-Code pipeline, quarterly policy reviews, IAM and CMDB integration, automatic tagging of new workloads, threat hunting on denied traffic | Mature microsegmentation program integrated with security and IT processes |
The most important element of the roadmap is Phase 5 — operational maturity. Deploying the technology is only half the success. Microsegmentation that is not actively managed gradually degrades: policies expire without renewal, temporary exceptions become permanent, and new workloads are not automatically covered. Without maintenance processes, the investment in microsegmentation begins to lose value 12–18 months after deployment.
Key success metrics include: percentage of workloads covered by microsegmentation policies (target: 100%), number of “unused” policies (target: minimal, indicating the effectiveness of reviews), time to deploy a policy for a new workload (target: automatic, under 1 hour), and number of incidents in which microsegmentation limited lateral movement (target: every incident reported and analyzed).
How does nFlo design and deploy microsegmentation for clients?
At nFlo, we approach microsegmentation as an architectural project, not merely a technical one. Over the course of 500+ projects delivered for 200+ clients, we have repeatedly observed the same pattern: organizations that purchased a microsegmentation tool without a proper discovery and policy design process ended up with an elaborate infrastructure and zero security benefit — because the policies were too loose to block anything.
Our engagement begins with the environment assessment phase and identification of “crown jewels” — systems whose compromise would be catastrophic for the organization. Active Directory domain controllers, ERP systems, customer databases, and backup systems are typical priorities. We deploy microsegmentation from the center outward: we first protect what matters most, and then gradually expand the scope.
We have our own application dependency mapping methodology, developed through experience with hundreds of environments. We combine network telemetry (NetFlow, firewall logs, EDR telemetry) with expert knowledge of typical communication patterns for popular technologies: Microsoft AD, Exchange, SharePoint, SAP, Oracle DB. This shortens the discovery phase and allows us to identify anomalies faster — flows that should not exist.
The response time set out in the contract for clients with active support means that production issues after blocking is switched on do not wait for the next business day. We know how painful it is when “after microsegmentation was deployed, the integration with system XYZ stopped working” — which is why for the first 30 days after each deployment stage, we provide an elevated level of monitoring and rapid response.
We also carry out verification of the effectiveness of the deployed segmentation. Our Red Team conducts controlled lateral movement simulations based on MITRE ATT&CK techniques, checking whether policies actually stop the attacker, or merely theoretically should do so. We report the reduction in breach risk per client rather than as one figure for everyone: it is the result of comparing penetration test results before and after deployment, measured by the scope that a simulated attacker was able to achieve in the environment.
For clients who are not yet ready for full microsegmentation deployment, we offer a phased approach: we start with “crown jewels” and minimum viable segmentation — protecting the most important systems with simple deny-all-with-exceptions policies. This delivers an immediate security benefit with limited effort and becomes the starting point for expanding scope in subsequent stages.
Related concepts
Explore key terms related to this article in our cybersecurity glossary:
- Network security — Network security is the practice of protecting the integrity, confidentiality, and availability of network resources…
- Zero Trust — Zero Trust is an IT security model that assumes no person, device, or application is trusted by default…
- Firewall — A firewall is a security system that monitors and controls incoming and outgoing network traffic…
- Cybersecurity — Cybersecurity is a set of techniques, processes, and practices for protecting IT systems, networks, and data…
- Penetration testing — Penetration testing is an authorized simulation of hacker attacks aimed at identifying vulnerabilities…
Learn more
Explore related articles in our knowledge base:
- Network microsegmentation: How to stop an attacker who has already gotten inside?
- What is ZTNA (Zero Trust Network Access)?
- Zero Trust Architecture — implementation in an organization
- Advanced Persistent Threats (APT): Is your organization in the crosshairs of cyber spies?
- Threat hunting in practice — how to proactively detect threats hidden in the network
Check our services
Do you need cybersecurity support? See:
- Network infrastructure deployments — design and deployment of secure networks
- Firewall deployments — perimeter security and internal segmentation
- Penetration testing — segmentation effectiveness verification by Red Team
- SOC as a Service — 24/7 security monitoring and lateral movement detection
FAQ — Frequently asked questions
Does microsegmentation replace the perimeter firewall?
Microsegmentation does not replace the perimeter firewall — it complements it. The perimeter firewall controls traffic between the internal and external network (north-south traffic), blocking unauthorized access from the internet. Microsegmentation controls traffic inside the network (east-west traffic), limiting the lateral movement of an attacker who has already breached the perimeter. Both mechanisms are essential and mutually complementary. An organization that relies solely on a perimeter firewall has a hard outer shell and a soft interior — vulnerable to exploitation after every successful breach. Adding microsegmentation means that even a compromised system does not serve as a launchpad for further network compromise.
How long does a typical microsegmentation deployment take?
Deployment time depends on the scale and complexity of the environment. For an environment with a few dozen workloads, minimum viable segmentation (protection of key systems) can be achieved in 6–8 weeks. For enterprise environments with thousands of workloads in a hybrid setting, full deployment typically takes from 6 to 18 months. The longest stage is typically the discovery and policy design phase, not the technical tool installation itself. This phase should not be shortened — mistakes made during policy design can result in production disruption or security gaps that only surface months after deployment.
What is lateral movement and why is it so dangerous?
Lateral movement (TA0008 in MITRE ATT&CK) is a set of techniques that allow an attacker to move through a network after gaining initial access. Instead of exploiting a single system, the attacker “hops” from system to system, using stolen credentials, administrative protocols (RDP, SMB, WMI), or vulnerabilities on subsequent machines. The goal is to reach high-value systems: domain controllers, databases, backup systems. Lateral movement is dangerous because it often uses legitimate tools and protocols, making it difficult for automated detection systems to identify. This is precisely why limiting it through microsegmentation is one of the most effective forms of defense.
Does microsegmentation work in cloud and container environments?
Yes, although it requires adapting the technical approach to the specifics of the environment. In public clouds (AWS, Azure, GCP), native mechanisms are available: Security Groups, Network Security Groups, VPC Firewall Rules — which implement basic microsegmentation at the VM level. For more advanced requirements, host-based solutions (Illumio, Guardicore) install an agent on cloud virtual machines, providing consistent policy management between on-premises and cloud. For Kubernetes environments, tools such as Cilium (eBPF-based) or Calico implement granular NetworkPolicy at the pod and namespace level — with the ability to extend to identity-based policies based on Service Accounts. The key advantage of modern host-based solutions is precisely that they operate consistently across the entire hybrid environment.
What are the main success indicators of a microsegmentation project?
The success of a microsegmentation project is measured at several levels. Coverage metrics: percentage of workloads covered by an active microsegmentation policy (target: 100% for production systems), percentage of workloads operating with deny-all as the default policy (target: 100%). Policy quality metrics: number of “unused” policies (policies for which no traffic has been recorded in 90 days — an indicator of review effectiveness), time from provisioning a new workload to it being covered by a policy (target: under 1 hour, automated). Security metrics: result of a controlled lateral movement simulation (the scope an attacker achieved before being stopped by policies) measured regularly by a Red Team or penetration testers. Operational metrics: number of production incidents caused by incorrect policies, resolution time for such incidents.
Sources
- NIST Special Publication 800-207: Zero Trust Architecture — https://doi.org/10.6028/NIST.SP.800-207
- MITRE ATT&CK Framework — Lateral Movement Tactic (TA0008): https://attack.mitre.org/tactics/TA0008/
- Mandiant M-Trends 2025 Report — dwell time and incident statistics
- CrowdStrike Global Threat Report 2025 — lateral movement techniques and attack trends
- Illumio Zero Trust Segmentation — technical documentation and whitepapers
- Gartner Market Guide for Microsegmentation (2024) — microsegmentation tools market analysis
- FireEye/Mandiant SolarWinds Attack Analysis (2020–2021) — post-mortem analysis of operation UNC2452
- CSA (Cloud Security Alliance): Software Defined Perimeter and Zero Trust — implementation guidelines
Explore Our Products
Solutions mentioned in this article that can help protect your organization:
- FortiGate — Fortinet
Need expert support? nFlo team can help secure your organization:
