Showing raw Markdown: the renderer could not load.
## Questions ### 1. An online shop already has a stateful firewall that allows only TCP 443 from the Internet to its web server. The security team still wants to stop SQL injection attempts in the login form. What should they add? *Objective: 1. Compare stateless packet filters, stateful firewalls, next-generation firewalls and web application firewalls (WAF).* - A) A second stateful firewall with the same rules, for redundancy. - B) A stateless packet filter on the router, to block suspicious source addresses. - C) A web application firewall in front of the site, which reads the HTTP requests. - D) A host firewall on the web server that blocks TCP 22. ### 2. A router access control list contains one rule: `permit tcp any host 192.0.2.10 eq 443`. It is stateless and has no other permit rules. Clients can send requests to the web server at `192.0.2.10`, yet they never receive a page. Why? *Objective: 1. Compare stateless packet filters, stateful firewalls, next-generation firewalls and web application firewalls (WAF).* - A) The clients' requests are blocked, because the rule must also list the clients' source port. - B) The server's replies are dropped, because a stateless filter judges each packet alone and no rule permits packets from port 443 back to the clients. - C) TCP 443 is encrypted, so a stateless filter cannot read the requests. - D) A stateless filter only allows UDP. ### 3. The web server in the DMZ at `10.40.2.10` needs to read from the database at `10.40.3.10`, which listens on TCP 5432. Which rule follows least privilege? *Objective: 2. Design network zones with a demilitarized zone (DMZ) and write a default-deny rule set that allows only the required flows.* - A) Allow new connections from `10.40.2.10` to `10.40.3.10` on TCP 5432 only. - B) Allow all traffic from the DMZ subnet to the internal subnet, because the DMZ is under your control. - C) Allow new connections from the Internet to `10.40.3.10` on TCP 5432, so that the web server can reach it through the Internet. - D) Move the database into the DMZ so that no rule is needed. ### 4. A rule set has a `policy drop` and these rules in this order. Users report that nobody can reach the web server on TCP 443. What is wrong? *Objective: 2. Design network zones with a demilitarized zone (DMZ) and write a default-deny rule set that allows only the required flows.* ```text 1 ct state established,related accept 2 counter drop comment "default deny" 3 iifname "net0" oifname "dmz0" ip daddr 10.40.2.10 tcp dport 443 ct state new accept ``` - A) Rule 1 blocks all new connections. - B) The `policy drop` overrides every `accept` rule. - C) Rule 3 must name the source address, or it is ignored. - D) Rule 2 drops every packet that reaches it, so rule 3, placed after it, is never evaluated. ### 5. A team wants to add an intrusion prevention system (IPS) in line with the traffic to an important customer portal. They're worried that a wrong rule would block paying customers. What is the sensible first step? *Objective: 3. Explain what intrusion detection and prevention systems (IDS/IPS) do and how they complement firewalls.* - A) Run it in detect-only (alert) mode first, review the alerts for false positives, then switch rules to block. - B) Skip it: an IPS cannot be tuned, so it always blocks real users. - C) Replace the firewall with the IPS, which does the same job. - D) Enable every rule in block mode on day one, to get the most protection. ### 6. A new exploit has just appeared. No signature exists for it yet, but it makes a web server open connections to many unknown hosts, which that server has never done. Which detection method is most likely to notice it? *Objective: 3. Explain what intrusion detection and prevention systems (IDS/IPS) do and how they complement firewalls.* - A) Signature-based detection, because it contains every known exploit. - B) A stateless packet filter, because it keeps a list of all attackers. - C) Anomaly-based detection, because it alerts on behaviour that deviates from the learned normal. - D) None: an IDS can only detect attacks that have already succeeded. ### 7. An attacker steals a laptop and finds it is still logged in to the company VPN. A traditional VPN gives it access to the whole internal network. Which zero trust measure limits the damage most? *Objective: 4. Explain the principles of zero trust architecture and how they differ from perimeter-based security.* - A) Each request to each resource is authorised on identity and device health, so the VPN connection alone grants nothing. - B) A longer password on the VPN account. - C) A second edge firewall in front of the first one. - D) Encrypting the VPN tunnel with a stronger cipher. ### 8. In the zero trust architecture of NIST SP 800-207, what is the job of the policy enforcement point (PEP)? *Objective: 4. Explain the principles of zero trust architecture and how they differ from perimeter-based security.* - A) It decides whether a user is trustworthy, based on identity and device health. - B) It stores the company's access policies and rules. - C) It encrypts all the traffic between users and servers. - D) It sits in front of a resource and opens or closes the path as the policy decision point directs. ### 9. You apply a default-deny policy and scan the database server `10.40.3.10` from the outside. You expect the firewall to hide it. The scan reports `Not shown: 1024 closed tcp ports (conn-refused)`. What does that result tell you? *Objective: 5. Verify a firewall policy by testing allowed and blocked flows and reading the firewall's logs.* - A) The firewall is dropping the probes silently, which is what you wanted. - B) The scan failed and must be repeated with more retries. - C) The probes reached the host, which answered with refusals, so the firewall is not blocking this path. - D) The database is down. ### 10. After loading the policy, a test from the web server to the database on TCP 5432 times out. The `default deny` counter rose by 2 packets, the `web to db: 5432` counter stayed at 0 and the drop log shows `10.40.2.99.41822 > 10.40.3.10.5432`. What is the most likely cause? *Objective: 5. Verify a firewall policy by testing allowed and blocked flows and reading the firewall's logs.* ```text iifname "dmz0" oifname "int0" ip saddr 10.40.2.10 ip daddr 10.40.3.10 tcp dport 5432 ct state new counter accept comment "web to db: 5432" ``` - A) The database is not listening on port 5432. - B) The connection state is `invalid`, so the first rule dropped it. - C) The log rule consumed the packets before the allow rule. - D) The web server's packets come from `10.40.2.99`, but the rule allows only the source `10.40.2.10`. --- ## Answer key <details> <summary>Show answers</summary> | # | Answer | Explanation | |---|---|---| | 1 | C | SQL injection travels inside valid HTTP requests to an allowed port, so only a layer 7 control that reads those requests can recognise it. A second stateful firewall or a stateless filter sees only addresses and ports, and a rule on TCP 22 is unrelated. | | 2 | B | A stateless filter has no connection table, so the replies are new packets to judge: they need their own rule (source port 443 to the client's ports). The requests match the rule and arrive, and the rule doesn't need the clients' source port (it is `any` by default). Encryption is irrelevant to a filter that reads headers only, and a stateless filter handles TCP. | | 3 | A | Least privilege names the exact source, destination and port. Allowing the whole DMZ subnet gives an attacker who takes over any DMZ host access to every internal host. Routing database traffic through the Internet exposes the database, and a database in the DMZ loses the protection of the internal zone. | | 4 | D | The first matching rule wins and rule 2 matches everything that rule 1 did not, so rule 3 is never reached. The `policy drop` is only the default for packets that match no rule, it doesn't override an accept. Rule 1 accepts only replies, and naming the source is not required. | | 5 | A | An inline IPS can block real traffic on a false positive, so new rules run in detect-only mode first, are tuned against real traffic, and are switched to block when they are trusted. IPS rules can be tuned, and an IPS complements a firewall, it doesn't replace it. | | 6 | C | Anomaly-based detection compares behaviour with a learned baseline, so a server that suddenly calls many unknown hosts stands out even without a signature. Signature detection only knows patterns that already exist, and a packet filter doesn't keep a list of attackers. | | 7 | A | Zero trust grants access per request, for one resource, after checking identity and device posture, so a VPN session on a stolen laptop opens little. A stronger cipher or password doesn't help once the device is already authenticated, and another edge firewall is still a perimeter. | | 8 | D | The PEP is the gate in front of each resource: it enables or ends the connection according to the decisions of the policy decision point (policy engine and policy administrator). Deciding is the PDP's job, not the PEP's. | | 9 | C | `closed (conn-refused)` means the probe reached a host that answered with a refusal. A firewall that drops shows `filtered (no-response)`. So this path is open to the Internet: a rule is too broad, or the packets don't pass through the firewall. | | 10 | D | The packets reached the default-deny rule, so no allow rule matched them. The log shows their source is `10.40.2.99`, while the rule matches only `10.40.2.10`, which is why that rule's counter stayed at 0. The log rule doesn't consume packets (it lets them continue), and an `invalid` packet wouldn't have matched a new-connection rule anyway. | </details>