Choking Lateral Movement: East-West Controls After Change Healthcare
When perimeter defenses fail, flat internal networks turn single compromised credentials into catastrophic ransomware events.

ALPHV BlackCat operators sat inside the Change Healthcare network for nine days before triggering their ransomware payload. They spent this dwell time mapping subnets. They dumped credentials. They exfiltrated terabytes of protected health information. The resulting outage halted pharmacy prescriptions across the country and disrupted cash flow for thousands of medical practices. The initial access vector was almost painfully mundane. The attackers compromised a Citrix remote access portal lacking multi-factor authentication.
The catastrophic blast radius was not a failure of the perimeter. It was a failure of internal containment. Once threat actors bypass the initial gateway, their survival depends entirely on lateral movement. If they land in a flat network, they hold the keys to the kingdom. We spend millions analyzing ingress traffic and tuning endpoint detection telemetry. Yet many enterprise networks remain structurally porous on the inside. Effective east west segmentation is the only architectural control that actively chokes an adversary who already holds valid credentials.
The Anatomy of a Perimeter Bypass
The narrative around major breaches often fixates on the initial access vector. In the Change Healthcare incident, compromised credentials gave the attackers their foothold. The gateway accepted the valid username and password and dropped the attackers onto the internal network. But initial access is just the first step in the MITRE ATT&CK framework under tactic TA0001.
The real damage happens during the lateral movement phase documented as TA0008. Attackers rarely land exactly where the most valuable data resides. They typically compromise a low-privilege workstation or a forgotten edge server. From there, they must scan the environment, harvest administrative accounts, and pivot to high-value infrastructure.
Living Off the Land in a Flat Environment
When internal boundaries do not exist, this pivoting is trivial. Operators use native administrative tools to blend in with legitimate traffic. They query Active Directory via LDAP to map trust relationships and identify domain controllers. They use the Server Message Block protocol to test credentials across hundreds of endpoints simultaneously. Because the network implicitly trusts communication between internal subnets, security monitoring tools struggle to distinguish an adversary moving laterally from an IT administrator deploying software updates.
Why Compromised Credentials Require Internal Firewalls
An attacker armed with valid remote access credentials is indistinguishable from a legitimate user to most perimeter firewalls. Once inside, they immediately execute credential dumping techniques like OS Credential Forging. They extract NTLM hashes from the Local Security Authority Subsystem Service memory using tools like Mimikatz or native Windows utilities. If the network lacks boundaries, that hash allows them to authenticate to other Windows machines without ever knowing the plaintext password.
If the network allows any endpoint to initiate a Remote Desktop Protocol session with any other endpoint, the blast radius is effectively the entire company. A compromised workstation in the marketing department can authenticate directly to a backup server in the data center. The perimeter might be a reinforced steel door, but the interior is a single massive room with no partitions.
SecurityEvent | where EventID == 4624 and LogonType == 3 | summarize count() by IpAddress, TargetUserName — KQL Hunt Query for Network Logons
That hunting query looks for network logons using high-privilege accounts. In a segmented environment, seeing a standard user workstation execute that logon against a domain controller triggers an immediate high-severity alert. In a flat network, that activity is buried under thousands of identical authentication events generated by poorly configured service accounts and legacy applications.
Phased Microsegmentation via Identity-Based Policies
Deploying internal boundaries retrospectively is notoriously difficult. Network engineers resist segmenting established environments because blocking the wrong port will break legacy applications. The traditional approach of carving up Virtual Local Area Networks and writing thousands of static IP-based firewall rules is brittle. IP addresses change constantly in modern environments, making static rules obsolete almost as soon as they are committed to the firewall appliance.
Moving Beyond Static IP Rules
The modern approach relies on identity-based microsegmentation. Policy engines abstract the underlying IP addresses away entirely. Instead of relying purely on network topology, policies are tied to logical tags, user identities, and machine certificates. A policy dictates that only endpoints tagged as database administrators can initiate Secure Shell connections to production database servers.
You do not need to segment everything simultaneously to achieve massive risk reduction. The most successful deployments start with an asset discovery phase. Security teams place agents on endpoints to map application dependencies without blocking any traffic. They analyze this NetFlow data for weeks to identify exactly which ports and protocols are strictly necessary for business operations. Only then do they switch the policies into enforcement mode.
Isolating the Tier Zero Infrastructure
Not all assets require the same level of isolation. Your primary focus must be choking access to Tier Zero infrastructure. Active Directory Domain Controllers, hypervisor management consoles like vCenter, and immutable backup servers are the absolute highest value targets for ransomware operators. If an attacker compromises vCenter, they control the entire virtualized compute environment. They can snapshot servers, modify memory, and deploy encryptors directly to the hypervisor layer.
These systems must sit in dedicated, highly restricted enclaves. Host-based firewalls on these critical servers should drop all inbound connections by default. You then explicitly allow administrative access only from designated privileged access workstations or jump hosts. Access to these jump hosts must require hardware-backed multi-factor authentication.
A standard user laptop should never be able to reach the web interface of your enterprise backup solution. By restricting access to a handful of hardened jump servers, you force the attacker through a predictable chokepoint. Even if they dump credentials from an infected workstation, they cannot route traffic to the backup server to delete your recovery points. They must first compromise the jump host, increasing their dwell time and the likelihood of detection.
Operational Runbook for Choking Lateral Movement
Theory is useless without execution. You can drastically reduce your attack surface this week by systematically restricting the three protocols most abused for lateral movement. Remote Desktop Protocol, Server Message Block, and Windows Remote Management are the highways attackers use to navigate your environment.
Step One: Restrict Remote Desktop Protocol
Remote Desktop Protocol over TCP port 3389 should never be allowed workstation-to-workstation. There is rarely a legitimate business reason for a finance laptop to remote into an engineering laptop. Push a Group Policy Object enforcing the Windows Defender Firewall to block inbound RDP originating from client subnets. Only allow inbound RDP from your dedicated IT support subnets or jump hosts.
Step Two: Segment Server Message Block Traffic
Server Message Block over TCP port 445 is essential for file sharing and Active Directory authentication. It is also the primary vector for credential harvesting and remote service execution under technique T1021.002. Block inbound SMB at the host level on all client workstations. Workstations need to talk to file servers and domain controllers, but they do not need to accept inbound SMB connections from each other. This single rule stops the majority of automated lateral movement scripts dead in their tracks.
Step Three: Control Windows Remote Management
Windows Remote Management over TCP ports 5985 and 5986 is a powerful administrative tool frequently abused by attackers using PowerShell Remoting. Similar to RDP, you must restrict which machines are allowed to initiate WinRM sessions. Configure the Windows Remote Management service to only accept connections from your management subnets. If an adversary attempts to pivot using WinRM from a compromised web server in the DMZ, the host-based firewall on the target machine will silently drop the packets.
The takeaway
Network architecture is hostile territory for an attacker only when defenders deliberately design it to be. Relying on edge firewalls to contain credentialed adversaries is a losing strategy, as the Change Healthcare incident starkly demonstrated. When you accept that initial access is inevitable, your entire defensive posture shifts toward containment and friction.
Implementing strict boundaries across your internal zones buys your incident response team the most critical resource in a breach. Time. You force the attacker to generate noise, bypass chokepoints, and burn advanced tooling just to move one subnet over. Because lateral movement often involves pivoting through internal messaging to phish secondary targets or manipulate workflows, platforms like MailSleuth.AI provide the essential visibility to correlate anomalous email behaviors with internal network flows. Stop treating the inside of your network as a trusted zone.
We dissect phishing campaigns and email infrastructure so you don't have to.


