Lesson 5
Regulating Network Access to Improve Security
Network access control is the practice of making sure that only the right identities get into the right systems, at the right time, in the right way - and that everything else is denied. Strong access control directly protects confidentiality (who can see data), integrity (who can change it), and availability (who can affect systems or services).
Modern access control is not just "who can log in." It is an ongoing security posture that includes:
- Identity and Access Management (IAM)
- Principle of least privilege / Role-Based Access Control (RBAC)
- Multi-factor authentication (MFA)
- Network segmentation and isolation
- Continuous monitoring and response
- Enforcement at the system level (ACLs, allowlists, application control)
The rest of this lesson walks through these areas and explains how they work together.
1. Identity and Access Management (IAM)
Identity and Access Management answers two questions:
- Who are you? (authentication)
- What are you allowed to do? (authorization)
Good IAM reduces the blast radius of a compromised account and makes it easier to prove compliance. Key IAM practices include:
- Least Privilege: Every user, service account, API client, or device should have only the minimum permissions required to perform its task. Broad “admin everywhere” access is no longer acceptable.
- Role-Based Access Control (RBAC): Instead of granting permissions to each individual user, you define roles such as DBA_ReadOnly, Helpdesk_Tier1, or Prod_Deployment. Users are added to roles. Roles map to permissions. This keeps access predictable and auditable.
- Multi-Factor Authentication (MFA): MFA requires more than just a password. Typical patterns include a password (something you know) plus a one-time code, device approval, or hardware key (something you have). Enforcing MFA for VPN, remote desktop, cloud consoles, and administrative logins is considered baseline security.
- Account Lifecycle Management: Accounts should be created intentionally, reviewed regularly, and removed immediately when no longer needed. Dormant accounts are high-risk because attackers can quietly reuse them.
Under the hood, IAM depends on modern cryptography. Legacy encryption like DES is considered obsolete; secrets, passwords, and stored credentials should be protected with strong algorithms such as AES for encryption and SHA-256 or stronger for hashing.
2. Network Segmentation and Isolation
Network segmentation means you do not treat your internal network as one giant “trusted” zone. Instead, you break it into smaller zones and strictly control which systems can talk to which.
Why segmentation matters:
- If an attacker compromises a single workstation, they should not automatically be able to reach production databases, payment systems, identity providers, or backups.
- Critical workloads (for example, HR systems containing employee data or systems that process financial records) should live in hardened network segments with limited inbound and outbound paths.
- Access to high-value segments should require authenticated, logged, policy-checked connections - not just “I’m on the same VLAN.”
Typical segmentation techniques include:
- VLANs and Subnets: Group similar assets or sensitivity levels together (for example, “user laptops,” “dev servers,” “production database servers”) and route traffic between them deliberately instead of by default.
- Firewall Policies / ACLs Between Segments: Firewalls and routers enforce rules like “Web front end can talk to App tier on TCP/8443, nothing else” or “Only backup server can connect to database on TCP/1521.” If traffic is not explicitly allowed, it is denied.
- Network Zones: You can define security zones such as Public-facing DMZ, Internal Business Apps, and Restricted Data Vault. Users and services get access to zones based on job function and policy, not convenience.
Segmentation limits the damage an attacker can do after the first foothold. It also makes alerting cleaner: if a workstation in “Accounting-Laptops” suddenly talks to “Prod-DB-Network,” you know something is wrong.
3. Continuous Monitoring and Incident Response
Controlling access is not enough - you must verify that your controls are actually working and react when they fail.
- Traffic Monitoring / IDS: Intrusion Detection Systems (IDS) and network analytics tools watch for suspicious behavior, such as abnormal data exfiltration or login attempts from unexpected geographies.
- Log Aggregation and Review: Firewalls, VPN gateways, authentication services, cloud IAM, and directory services (like Active Directory) all generate logs. Collect and retain them in a central place. You cannot investigate what you did not record.
- Incident Response Plan: You need a documented plan for what happens when you detect unauthorized access - who is paged, who is allowed to isolate systems, how you communicate internally, and how evidence is preserved. This is what turns “we noticed something weird” into “we controlled the situation.”
The faster you detect and contain unauthorized access, the less time an attacker has to laterally move, elevate privileges, or tamper with logs.
4. Enforcement at the System Level
Even with IAM and segmentation in place, the operating system and applications still need to enforce rules locally. This is where concepts like Access Control Lists (ACLs) and execution control come in.
Access Control Lists (ACLs)
An Access Control List defines who (which user, service account, group, or process) is allowed to interact with a resource and how they are allowed to interact with it.
An ACL entry typically looks like: “Group X may read this file but may not write or delete it,” or “Service account Y may open TCP port 8443 to host Z.”
If a user or process attempts any action that is not explicitly allowed, the OS or service denies it and logs that decision. This protects databases, application config files, secrets vaults, and other high-value assets.
Example: A firewall or router ACL that only allows an application server to reach a backend service:
# Only allow the app server 10.20.5.17 to reach the DB on TCP/5432
permit tcp host 10.20.5.17 host 10.20.9.42 eq 5432
# Block everything else by default
deny ip any host 10.20.9.42
In plain English: the database should not be reachable by “whoever can guess the IP.” It should be reachable only by the specific application node(s) that are supposed to talk to it.
Application Execution Control / Allowlisting
Another layer of access control is restricting what programs are even allowed to run - especially on servers that handle sensitive data or on user endpoints where malware is a risk.
Historically this was described as an Execution Control List (ECL): a policy that determines which actions a program may take once it’s running. Modern operating systems and endpoint protection tools generalize this concept into “application allowlisting” or “application control.”
Key ideas:
- Only approved software (signed binaries, trusted hashes, known packages) can execute. Everything else is blocked by default.
- The OS or security agent can prevent unapproved code from sending data off-host, tampering with protected registry keys / configuration files, injecting into other processes, or opening sensitive network connections.
- If a malicious program - for example, a trojan designed to quietly email internal files outside the company - tries to perform an unauthorized action, that action is blocked and an alert is generated.
This turns “malware ran and we hope antivirus catches it” into “malware cannot perform high-risk actions even if it runs.”
5. Configuration Discipline
Even the best access control model can be destroyed by leaving defaults in place. A misconfigured service that ships with “admin / admin” credentials or exposes an unrestricted management port to the internal network can completely bypass IAM, ACLs, and segmentation.
To close that gap:
- Harden new systems during provisioning. Do not keep vendor defaults for admin accounts, SNMP communities, management URLs, etc.
- Disable services and daemons you do not actually use. Every unnecessary service is another door somebody can try.
- Regularly review firewall rules, ACLs, security groups, and privileged roles. Remove stale entries. You are accountable for what is still allowed, not for what “nobody uses anymore.”
Summary
Controlling network access is not a single product. It is a set of reinforcing practices:
- Strong identity proof (authentication) backed by MFA and modern cryptography
- Clear authorization through least privilege and RBAC
- Network segmentation that limits lateral movement
- Continuous monitoring and a real incident response plan
- Local enforcement through ACLs and application allowlisting
- Secure configuration and ongoing review
When those pieces work together, unauthorized users are blocked, compromised accounts are contained, and suspicious activity is noticed quickly instead of weeks later. That is the goal of access control.
