Model for Network Security

1. The Stallings Message-Security Model

Stallings (Fig 1.2) models any secure transfer between Alice and Bob as four elements plus a trusted third party:

Alice                                    Bob
 +------+  Transform   X  Secure    Y  Transform  +------+
 | Msg M| -----------> | Channel | -----------> | Msg M|
 +------+  Encrypt/MAC +-----------------------> +------+
      \\                                                /
       \\________ Trusted Third Party _________________/
        (arbiter / key distribution centre / CA)
         Secrets (keys), opponent info, control input

Flow: Alice applies a security transform (encryption, MAC) using secret info (key) to get $X$ ; $X$ crosses a possibly hostile channel as $Y$ (opponent may intercept/modify); Bob inverts the transform to recover $M$ . Both draw secrets and policy from the trusted third party.

2. Four Design Tasks (Memorize)

X.800/Stallings lists exactly four tasks to design the transform:

Design an algorithm for the security transform (e.g. AES, HMAC) — public, published, peer-reviewed.
Generate the secret information (keys) — random, long, distributed securely.
Develop methods to distribute and share secrets (key distribution, Module 3 Topic 7).
Specify a protocol using the algorithm + secrets to achieve the service (e.g. TLS handshake, Kerberos exchange).

Fail any one and security collapses: strong AES + weak password-derived key = breakable; good key + ECB mode protocol error = pattern leak.

3. The Two Network-Access Security Models

Beyond message transfer, Stallings models protecting a system (e.g. a server) from attackers — the gatekeeper model:

                [ Hacker ]          [ Insider ]
                     \\              /
                      v            v
               +-----------------------+
               |  Gatekeeper function  |
               | (firewall / IDS /     |
               |  access control)      |
               +-----------+-----------+
                           |  audited internal users
                    +------+------+
                    | Information |
                    |   System    |
                    +-------------+

Two sub-models: (a) gatekeeper — password login, firewall rules, IDS filtering, malware scanning keep outsiders out; (b) internal audit — logging, anomaly detection catch malicious insiders the gatekeeper let through. Access control + authentication at the gate; accountability + audit inside.

4. Opponent Capabilities and Kerckhoffs Link

Model assumes opponent knows everything except the key (Kerckhoffs's principle — next topic): algorithm public, channel fully observed, ciphertext known. Security rests only on secret info. This is why tasks 2–3 (key generation/distribution) dominate real failures, and why "secret algorithm" designs (security through obscurity) violate the model.

Exam Points

Draw message-security diagram with trusted third party and $M \to X \to Y \to M$ labels — 5-mark diagram question.
List the four design tasks verbatim — 2–4 mark core.
Gatekeeper + insider-audit diagram — second 5-marker from this topic.
Opponent-knows-all-but-key assumption → Kerckhoffs bridge.

Model for Network Security