Paths Not Taken: a Secure Computing Tutorial
William Earl Boebert · 2025
Read Paths Not Taken: a Secure Computing Tutorial by William Earl Boebert (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-16088.txt.
Introduction
William Earl Boebert, a solo author, offers a tutorial on secure computing ideas from tagged and descriptor architectures in 2025. These older designs are proven but little used today, he argues, and they can block or remove many common flaws.
The tutorial shows the core rules through a model of a special guard system that hosts defensive software. It places that system in its working setting and uses systems engineering habits to frame the design. Boebert hopes teachers will add this past work to their courses. He also hopes self-taught students will add it to their own study of secure systems.
Catalog record
| CREATOR | William Earl Boebert |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 38 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2504-16088
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Paths Not Taken: a Secure Computing Tutorial William Earl Boebert Abstract This paper is a tutorial on the proven but currently under-appreciated security mechanisms associated with ”tagged” or ”descriptor” architectures. The tutorial shows how the principles behind such architectures can be applied to mitigate or eliminate vulnerabilities. The tutorial incorporates systems engineering practices by presenting the mechanisms in an informal model of an integrated artifact in its operational environment. The artifact is a special-purpose hardware/software system called a Guard which robustly hosts defensive software. It is hoped that this tutorial may encourage teachers to include significant past work in their curricula and students who are self-teaching to add that work to their exploration of secure computing. Organization of the Tutorial This tutorial is divided by topic area to facilitate incorporation in curricula and to assist those who are using it to self-teach. The topic areas are arranged in order of detail, from the most general to the most specific. The first area, Concept of Operations, discusses the issues of systems engineering and the importance of understanding and exploiting the environment in which systems operate. The second section, Models, discusses mental models and the use of abstraction in understanding computer systems. An overview of the Guard model is then given. The Mechanisms topic area is the most specific, where the known technology is described both in detail and in the context of the systems design given in the earlier sections. The final section offers suggestionss to those who may be interested in carrying this work further into an operational system. Concept of Operations No system can solve every aspect of a given problem, and experience has shown that explicit prior consideration of focus, limits, and approaches will often save time, money, and blood. A major goal of the tutorial is to show how an understanding of the context in which a system operates influences the design of mechanisms. To do so requires presenting a hypothetical operating environment for the hypothetical Guards. A proven systems engineering [ 1 ] technique is to define an operating environment and a system’s place in it by producing a Concept of Operations, or Conops [ 2 ] . Three Conops-level topics are included to help the student understand the rationale behind the mechanisms: a statement of problem the Guard is intended to solve, the general approach taken for the solution, and the goals of the assurance process. Statement of the Problem A generic network of interest to attackers is one whose purpose is to provide economic or social value. It does this by running applications on general-purpose platforms which directly or indirectly access the internet. The platforms are typically feature-rich operating systems such as Windows, Linux, or Mac OS. The applications may be a mix of locally developed, purchased, and open source software. Some networks, such as those which control processes such as water supplies and electrical distribution, will be comprised of specialized platforms and applications. With rare exception, such networks are designed and administered with the applications as a primary concern and defense against attack as secondary. This setting of priorities encourages attackers who already enjoy inherent advantages: attackers can choose the time, place, and nature of an attack and they only need to find one exploitable fault, while defenders must cope with all known and lurking ones, and attacker’s motivation is enhanced by the often overlooked fact that for many individuals attack is an adventure while defense is just a job. In today’s world attackers may also be directly or indirectly protected from retaliation by sovereign states. They will, in general, be adequately funded either by sponsors or by the unregulated transnational money flows made possible by cryptocurrency [ 3 ] . Being funded enables them to be persistent, to perform repetitive attacks and learn from each one. And attackers can exploit features of the internet which permit anonymous action at a distance. [ 4 ] . General Approach The sole purpose of a Guard is to host Security Services code. Development of that code is hosted elsewhere. There is also no requirement to support a browser, or data bases, or any other form of productive application. The isolation and concentration of Security Services in the Guard nodes means that Guards will be the primary targets of competent attacks, for their defeat would leave the network largely undefended. The design, implementation, and administration of Guard nodes recognizes this fact. Design and implementation places resistance to attack as a requirement above all others, and administration recognizes that social engineering and supply chain attacks will accompany direct attacks on the network. Separation of Security Services from applications is motivated by the radically different characteristics of the two classes when viewed from a life-cycle perspective. Application platforms are large, complex bodies of software which are subject to structural decay [ 5 ] unless updated carefully, which means at wide intervals. When new attacks appear, however, Services code must be updated with all deliberate speed so that attackers have the smallest practical window of exploitation. Putting Services code on the same platform leads to one of two undesirable options: if the Services+applications platform is updated at the rate required by Services, the application is at significant risk of decay. If the Services+application platform is updated at the careful pace necessary to prevent decay, updated Services code will be delayed and its effectiveness diluted. Figure 1 depicts potential applications of Guard instances to network defense. Figure 1: A Network With Guards The most basic application is a traditional firewall that filters packets and raises alarms at the juncture with the internet. Similar tasks could be carried out within the application network to detect or prevent adverse interactions between application nodes. Guards could interact with each other by using out-of-band communication or virtual private network (VPN) technology. A Guard could protect a critical but vulnerable internal node such as a machine learning system for intrusion detection, a network attached storage for backups, or a “honeypot” to trap attackers. In all cases the relevant Security Services are determined by the combination of applications and known attacks and therefore system-specific. The isolation of the Guards from vulnerable network elements is an adaptation of a design principle called Zero Trust : The Zero Trust Model is simple: cybersecurity professionals must stop trusting packets as if they were people. Instead, they must eliminate the idea of a trusted network (usually the internal network) and an untrusted network (external networks). In Zero Trust, all network traffic is untrusted. [ 7 ] Assurance Assurance determines how human administrators view the cyber systems they administer. Assurance is essential to effective response. Administrators who are assured of known system behavior are more likely to respond decisively and correctly to alarms and other indicators of potential attacks. There are two assurance goals associated with a Guard. The first is that the internal mechanisms of the Guard operate as advertised even when attacked. The second goal is to export assurance to the Services code it hosts and from that to the administrators. The second, external assurance goal can be somewhat flippantly expressed as ‘‘WYSIWYG 2 2 endnote: 2 “What You See Is What You Get,” originally a lighthearted way to