🔹Layered Security

Layer 1

(Physical Inspection/Access)

Threats

  • Rogue Device

  • Bad USB

  • Pentest Dropbox

  • HID Attacks

  • Physical Damage

Mitigations

  • Cabinets

  • Closets

  • Physical Authentication & Authorization

Wireless (802.11)

Threats

  • Dual-homed devices (2 or more interfaces)

  • Protocol Attacks

  • BYOAP (bring your own access point)

Mitigations

  • WIPS (wireless Intrusion Prevention System)

  • Disable Obsolete Protocols (WEP)

  • Disable Flawed Control Mechanisms (WPS)

  • Enable PMF (Protected Management Frames) if possible - available in 802.11w to block spoofing, deauth, disass and replay attacks

  • Guest Management (no auth for guests, internet access only, WPA2-PSK)

  • Station/Client Isolation (clients cant access other clients on the same AP)

  • WPA2 Enterprise (use unique encryption key for each client and authenticate with RADIUS servers)

Wireless (RFID badges)

Threats

  • Cloned Cards/Tags

Mitigations

  • Use cards that support rolling codes (code change on each use) or challenge-response

  • Use protective sleeves if using old RFID cards

Layer 2 & 3

Switches

Threats

  • MAC flooding

  • 802.1Q and ISL Tagging

  • Double Encapsulated 802.Q / VLAN hopping

  • ARP Poisoning

  • Private VLAN attack

  • VLAN Tunneling Protocol Attack

  • Multicast Brute Force

  • Spanning-tree Attack

  • Random Frame Stress

  • Rogue DHCP

  • DHCP Starvation

  • NTP Amplification

  • NTP Spoofing

  • STP Spoofing

  • Sniffing

Mitigations

  • Disable CDP to prevent credential spoofing

  • Setup port security to prevent MAC flood or hardcode system MAC address

  • Disable unused ports, services, protocols, interfaces, etc.

  • Setup MAC limitation and sticky MAC address

  • DHCP Snooping

  • Dynamic ARP Inspection (DAI)

  • ARP IDS (arpwatch)

  • Configure trusted interface for DHCP

  • Use private VLANs

  • use SSHv2

  • Enable logging

  • Use Cisco type 8 or 9 password hashes

  • Configure banner correctly

  • Always use SNMP v3 with secure community strings

Routers

Threats

  • DoS

  • IP spoof

  • IP Source Routing

  • ICMP Flood

  • Smurf Attack

  • Routing Table Poisoning

  • IPv6 Router Advertisement

  • Unauthorized Routing Updates

  • Wormhole Attack (Unauthenticated Tunneling)

Mitigations

  • Cisco Autosecure

  • DISA STIGs

  • CIS Cisco Benchmark

  • Use audit tools (Nipper Studio, nipper-ng)

  • Set logon filters on the outermost external router

  • Block non-routed IPs

IPv6

Threats

  • IPv6 MitM

  • DHCPv6 Takeover

  • Neighbor Impersonation

  • NDP Spoofing

Mitigations

  • Dynamic ARP Inspection

  • Secure Neighbor Discovery

  • Private VLANs

  • Port Security

  • 802.1x

  • Use IPsec if possible

  • Block protocol 41 to prevent IPv6 tunneling (Cisco IOS ACL)

  • Use rogue advertisement (RA) guard

Layer 7

Application Proxy

Forward Proxy : remote access through proxy, hides the identity of client.

  • Web Proxy

    • SSL Inspection

    • Block sites by categories

    • Website whitelisting

    • Authentication

  • SMTP Proxy (spam application)

    • per-email encryption

    • modify/auto-route mail (add custom header/footer or remove attachments)

    • anti-spam/anti-spoof (SPF/DKIM/DMARK)

    • sender authentication

    • rate limiting

    • sender blocking

Reverse Proxy : for security analysis, hides the identity of server (WAF, three-tier model, etc.)

Placement

  • DMZ proxy (between DMZ and internal network)

  • Internal proxy (betwwen clients and internet)

  • VPN access (for remote connection)

  • Cloud proxy (systems directly connect to cloud infrastructure)

WAF

Web Application Firewall (proxy) : strong focus on mitigating OWASP Top 10.

Capabilities

  • SSL offloading (client connects to reverse proxy with SSL/TLS and reverse proxy connects to web server with http)

  • content decoding

  • HTTP attack vector mitigations

  • virtual patching (mitigation without recoding the web app)

  • CAPTCHA & rate limiting

  • add HSTS header

  • error page control

  • dynamic action based on risk level

  • dynamic traffic routing for load balancing

Deployment:

  • Automatic Learning (traffic goes through WAF for a period of time so it can learn the patterns, after learning can flip to deny)

  • manual (all settings are manually tuned, many templates are available based on OS, programming language and back-end database )

Firewall

Internal Network Design

Break up internal network into separate segments based on :

  • Business & regulation requirements

  • Criticality of assets

  • Threats

  • Risk appetite

Segmentation Principles

classify systems and data in different levels (tiers):

  • tier 1 - critical components (DCs, exchange server, etc.)

  • tier 2 - internal systems

  • tier 3 - external facing data-providing systems

gateways are the control points with egress/ingress inspection and ACL.

Best Practices:

  • Define port-based rules first

  • Block geolocations

  • Authenticate all outbound traffics

  • Use both host-based and network-based

DMZ

DMZ is usually implemented in one of these models:

  • Three Legged DMZ Model (Single Firewall) : uses a single firewall with at least three network interfaces to make the architecture that holds a DMZ.

  • Dual Firewall DMZ Model : the DMZ is placed between 2 firewalls and traffix is filtered in both sides.

In case of having multiple systems/servers in the DMZ, its highly recommended to separate them or split them into multiple DMZs (or trust zones).

In DMZ, segmentation is done using VLANs, VRFs or physical separation.

DMZ + Active Directory

In case of using Active Directory connected servers in the DMZ (such as a public website or domain), the DMZ should not be directly connected to primary/root domain controllers. in this scenario, a read-only domain controller (RODC) is placed in DMZ for authentication and authorization.

Services Like LDAP and SMB network shares shouldn't be visible from the DMZ point of view. for better security and easier management, use deferent domains for public-facing and internal use.

IDS vs IPS Design

IPS should be implemented in in-line mode so that all the traffic can pass through and be analyzed.

IDS should be implemented out-of-band with a network tap or port mirroring to prevent speed drop.

Last updated