Imagine writing a piece of networking software.

• It needs to enable two neighbouring (i.e., directly connected) devices to communicate.
• It also needs to enable two devices separated by multiple hops to communicate.

The Open Systems Interconnection (sometimes Open Systems Interconnect or simply OSI) model 1️⃣ enables networking software to be built in a structured manner; 2️⃣ provides an interoperability framework for networking protocols.

History: The OSI model was introduced in 1983 by several major computer and telecom companies; and was adopted by ISO as an international standard in 1984 [Imp22].

• The second and latest edition of the international standard is ISO/IEC 7498-1:1994 [ISO94].

Watch the following video for a quick overview:

Learning the seven layers from Networking Foundations: Networking Basics by Kevin Wallace

More details follows.

The OSI model is a logical (as opposed to physical) model that consists of seven nonoverlapping layers (going bottom-up, opposite to Fig. 1):

5. Layer 5 (L5), session layer [TW11, pp. 44-45]: This layer enables users on different machines to establish sessions between them.

   Sessions offer various services, including dialog control (keeping track of whose turn it is to transmit), token management (preventing two parties from attempting the same critical operation simultaneously), and synchronisation (checkpointing long transmissions to allow them to pick up from where they left off in the event of a crash and subsequent recovery).

   For example, SOCKS5 (see RFC 1928) is a session-layer protocol.

6. Layer 6 (L6), presentation layer [TW11, p. 45]: This layer manages the syntax and semantics of the information to be transmitted.

   For most protocols however, the distinction between the presentation layer and application layer is blur. For example, the HyperText Transfer Protocol (HTTP) is commonly classified as an application-layer protocol although it has clear presentation-layer functions such as encoding, decoding, and managing different content types [FNR22].

7. Layer 7 (L7), application layer [TW11, p. 45]: This layer implements a suite of protocols for supporting end-user applications.

   For example, HTTP is a stateless application-layer protocol for distributed, collaborative, hypertext information systems [FNR22]. HTTP supports eight “methods”, including the GET method for requesting transfer of a target resource in a specified representation [FNR22, Sec. 9.3.1].

Physical unclonable functions (PUFs, see Definition 1) serve as a physical and unclonable alternative to digital cryptographic keys.

Think of a PUF as a keyed hash function, where the key is built-in and unique due to manufacturing variations [GASA20].

• Given an input, which we shall call a challenge, a PUF outputs a response. The challenge-response pair (CRP) is unique to the PUF.
• Every CRP is used only once.

Types of PUFs include 1️⃣ optical PUFs, 2️⃣ arbiter PUFs, 3️⃣ memory-based intrinsic PUFs [GASA20].

• An intrinsic PUF is a PUF that is already embedded within a device at the time of manufacturing.
• The first intrinsic PUF was introduced in 2007 in the form of an SRAM PUF.
• Flash memory PUFs and DRAM PUFs were subsequently introduced.
• A memory-based PUF usually offers desired independence among response bits, so its primary application is on-demand derivation of volatile cryptographic keys.

Watch a high-level introduction to SRAM PUF:

Proximity-1 covers the data link layer [CCS20d] and physical layer [CCS13b].

Proximity-1 enables communications among probes, landers, rovers, orbiting constellations, and orbiting relays in a proximate environment, up to about 100,000 km [CCS13c].

These scenarios are devoid of manual intervention from ground operators, and furthermore, resources such as computational power and storage are typically limited at both ends of the link.

In fact, Proximity-1 has been field-tested in the 2004-2005 Mars missions; see Figs. 1-2 for illustration.

In contrast, the AOS/TC/TM Space Data Link Protocols are meant for Earth-deep space links, over extremely long distances.

Proximity-1 supports symbol rates of up to 4,096,000 coded symbols per second.

Designed for the Mars environment, the physical Layer of Proximity-1 only uses UHF frequencies [CCS13b, Sec. 1.2].

The frequency range consists of 60 MHz between 390 MHz to 450 MHz with a 30 MHz guard-band between forward and return frequency bands, specifically 435-450 MHz for the forward channel and 390-405 MHz for the return channel [CCS13b, Sec. 3.3.2].

Also known as asymmetric-key cryptography, public-key cryptography (PKC) uses a pair of keys called a public key and a private key for 1️⃣ encryption and decryption, as well as 2️⃣ signing and verification.

Digital certificates and public-key infrastructure (PKI)

Suppose 👩 Alice is somebody everybody trusts.

• When 👩 Alice signs 🧔 Bob’s public key, anybody can verify Bob’s public key using Alice’s public key.
• Successful verification means we can trust that the public key is Bob’s because we trust Alice.
• Essentially, 🧔 Bob’s public key with 👩 Alice’s signature on it serves a 📜 digital certificate (see Definition 2) certifying Bob’s identity.

Watch a quick introduction to digital certificates on LinkedIn Learning:

Digital certificates and signing from Ethical Hacking: Cryptography by Stephanie Domas

Digital certificates are only useful if we can trust their signatories.

To ensure signatories and hence certificates can be trusted, PKC relies on a public-key infrastructure (PKI, see Definition 3) to work.

Watch a quick introduction to PKI from an operational viewpoint on LinkedIn Learning:

Cryptography: Public key infrastructure and certificates from CISA Cert Prep: 5 Information Asset Protection for IS Auditors by Human Element LLC and Michael Lester

A PKI, as specified in the ITU-T X.509 [ITU19] standard, consist of certification authorities (CAs).

• One or more CAs are trusted to create and digitally sign public-key certificates in response to certificate signing requests (CSRs).

• A CA may optionally create the subjects’ keys.

• A CA certificate is a public-key certificate for one CA [ITU19, Sec. 7.4]

  • issued by another CA, in which case the CA certificate is a cross-certificate;

  • issued by the same CA, in which case the CA certificate is a self-issued certificate.

    If the signing key is the private key associated with the public key signed, the self-issued certificate is a self-signed certificate.

• Thus, CAs can clearly exist in a hierarchy, e.g., the two-tier hierarchy in Fig. 1, or the three-tier hierarchy in Fig. 2.
• In a hierarchy, the root CA serves as the trust anchor [ITU19, Sec. 7.5].
• Examples of CAs: IdenTrust, DigiCert Group, others.
• An example of a software solution that implements CA functionality is Cloudfare’s CFSSL.

A PKI also has registration authorities (RAs).

• One or more RAs are responsible for those aspects of a CA’s responsibilities that are related to identification and authentication of the subject of a public-key certificate to be issued by that CA.
• An RA may either be a separate entity or be an integrated part of the CA.
• CAs typically play the role of RA as well.
• An example of a software solution that implements RA functionality is PrimeKey’s EJBCA Registration Authority.

Public-key cryptosystems

Algorithmically speaking, there is more than one way of constructing a public-key cryptosystem.

Standard public-key cryptosystems: 1️⃣ Rivest-Shamir-Adleman (RSA) cryptosystem, 2️⃣ elliptic-curve cryptosystems.

These cryptosystems rely on the hardness of certain computational problems for their security.

The hardness of these computational problems has come under threat of quantum computers and quantum algorithms like Shor’s algorithm.

As a countermeasure, NIST has been searching for post-quantum cryptography (PQC, also called quantum-resistant cryptography).

As of writing, there are three PQC candidates.

See 👇 attachment (coming soon) or the latest source on Overleaf.

Not ready for 2023 but see reference below.

A safe programming language is one that is memory-safe (see Definition 1), type-safe (see Definition 2) and thread-safe (see Definition 3).

A significant percentage of software vulnerabilities have been attributed to memory safety issues [NSA22], hence memory safety is of critical importance.

Examples of violations of memory safety can be found in the discussion of common weakness CWE-787 and CWE-125.

Type safety ⇒ memory safety, but the converse is not true [SM07, Sec. 6.5.2], hence type safety is commonly considered to be a central theme in language-based security [Fru07].

Type-safe programming languages, e.g., Java , Ruby, C#, Go, Kotlin, Swift and Rust, have been around for a while. However, memory-unsafe languages are still being used because:

• Type-safety features also come at the expense of resource requirements. Most memory-safe languages use garbage collection for memory management [NSA22], and this translates to higher memory usage.

• Although most type-safe languages are supported on the mainstream computing platforms (e.g., Wintel), the same cannot be said of embedded platforms.

  It can be challenging to program a resource-constrained platform using a type-safe language.

• There is already a vast amount of legacy code in C/C++ and other memory-unsafe languages.

  The cost to port legacy code, including the cost of training programmers, is often prohibitive.

  Depending on the language, interfacing memory-safe code with unsafe legacy code can be cumbersome.

• Besides invoking unsafe code, it is all too easy to do unsafe things with a type-safe language, e.g., not checking user input, not implementing access control.

  Programmers often use this as an excuse to not use a different language than what they are familiar with.

Nevertheless, adoption of type-safe languages, especially Rust, has been on the rise [Cla23].

Thread safety rounds up the desirable properties of type-safe languages.

Watch the following LinkedIn Learning video about thread safety:

Thread safety from IoT Foundations: Operating Systems Fundamentals by Ryan Hu

Rust is an example of a type-safe language that is also thread-safe.

Fig. 1 depicts the usage of different frequency bands.

IEEE Standard 512 [IEE20] defines the letters designating the frequency bands.

According to ESA, the main satellite frequency bands range from L to Ka; see also the same link for discussion of different usage of the frequency bands.

The increasing usage of cutting-edge technologies in safety-critical applications leads to strict requirements on the detection of defects both at the end of manufacturing and in the field [VDSDN+19].

This gives birth to the design-for-testability (DFT) paradigm, which necessitates additional test circuits to be implemented on a system to 1️⃣ provide access to internal circuit elements, and thereby 2️⃣ provide enhanced controllability and observability of these elements [BT19, Sec. 4.7.1].

Scan is the most popular DFT technique [BT19, Sec. 3.7.2].

• The technique of scan design (see Definition 1) offers simple read/write access to all or a subset of the storage elements in a design.

  Definition 1: Scan design [IEEE13, p. 10]

  A design technique that introduces shift-register paths into digital electronic circuitry, providing controllability and observability in deeply embedded regions of circuity and thereby improving testability.

• Scan design is realised by replacing flip-flops by with scan flip-flops (SFFs, see Fig. 1) and connecting them to form one or more shift registers in test mode.

  Watch YouTube video for a revision of D-type flip-flip:

• SFFs are generally used for clock-edge-triggered scan design, whereas level-sensitive scan design (LSSD) cells are used for level-sensitive, latch-based designs.

A scan chain is a chain of SFFs [BT19, Sec. 3.7.2.2]; see Fig. 2.

When Test Enable (TE) is high, the scan chain works in test/shift mode [BT19, Sec. 3.7.2.2; VDSDN+19, p. 95].

• Inputs from the Scan In (SI) pin are shifted through the scan chain to the Scan Out (SO) pin.
• The circuit is then set to normal mode and run for a specified number of clock cycles.
• When the circuit reaches a target state, the circuit is switched back into test mode and the content of the scan chain is shifted out from the SO port.
• Finally, a test program compares the SO values with the expected values for validation.

Multiple scan chains are often used to reduce the time to load and observe.

The integrity of a scan chain should be tested prior to application of a scan test sequence.

For more information, see the attachment.