Cybersecurity

Cybersecurity is the practice of keeping computers, the data they hold, and the people who depend on them safe from deliberate attack and accidental harm.

12 frameworks, from 1972 to today · Cybersecurity · Computer Science

still a working view ended · each leads to its entry below

Overview

Cybersecurity is the practice of keeping computers, the data they hold, and the people who depend on them safe from deliberate attack and accidental harm1. What makes it a distinct field rather than a corner of computer science is that its adversary thinks: a defender must anticipate what a motivated person will try next, and the defender's mistakes are often invisible until they are exploited. The history of the field is largely a history of that mismatch—between systems built on assumptions of good behavior and attackers who violate them—and of the successive ideas practitioners invented to close the gap. It runs from time-sharing mainframes in the 1960s to cloud-scale identity systems today, and its turning points are as much about institutions, law, and human behavior as about mathematics.

Time-sharing, the reference monitor, and multilevel security

The field's founding problem was access. When computers were batch machines, "security" meant physical control of a room. Time-sharing in the 1960s put many users on one machine at once, and suddenly the operating system had to decide which user could read which file2. Early systems such as CTSS at MIT and Multics, developed from 1965 by MIT, Bell Labs, and General Electric, made protection a first-class part of the operating system.

The theoretical response came from military concern about classified data. In 1973, David Bell and Leonard LaPadula at MITRE formalized a multilevel security model: each object (file, device) and subject (user, process) carries a classification, and access follows rules—no reading upward, no writing downward—that prevent a cleared analyst from leaking secrets to an uncleared one. Around the same time, Kenneth Biba proposed an integrity model with the mirror-image rules, protecting data from being corrupted by less trustworthy sources3. These access control models gave designers a vocabulary—mandatory versus discretionary control, labels, clearances—that still structures how operating systems and databases enforce policy.

Underneath the models sat a structural idea. In 1972, James Anderson's study for the U.S. Air Force introduced the reference monitor: a small, tamper-proof part of the system through which every access must pass, so that security could be verified by inspecting one component rather than the whole machine. The collection of code that must be trusted for the policy to hold became known as the trusted computing base. This was the intellectual foundation of formal assurance: if security rests on a small kernel, perhaps that kernel can be proved correct.

Orange Book, formal assurance, and the cost of proof

The U.S. Department of Defense turned these ideas into procurement rules. Its Trusted Computer System Evaluation Criteria, published in 1983 and known from its cover as the Orange Book, graded systems from D (no protection) through C1, C2, B1, B2, B3, to A1 (formally verified). The top grades demanded a formal model and a proof that the implementation matched it. The most famous attempt was the Honeywell SCOMP system, evaluated at A1 in the mid-1980s, and later the multinet gateway work and the seL4 microkernel, whose functional-correctness proof was completed in 2009—decades later, and still a research landmark rather than a mass-market product.

The lesson practitioners drew was sobering: formal verification was enormously expensive, and evaluated systems lagged the commercial market badly. The criteria movement nonetheless left durable ideas—security policy as an explicit object, evaluation as a discipline—and its successor frameworks, the Common Criteria (1996 onward, standardized internationally in 1999), are still used to certify smart cards and other high-assurance products. But the mainstream shifted toward practical hardening: least privilege, auditing, and the C2-level controls that shaped commercial operating systems.

Cryptography escapes the government

Until the 1970s, strong cryptography was effectively a state monopoly. That changed with two publications. In 1976, Whitfield Diffie and Martin Hellman published "New Directions in Cryptography," introducing public-key cryptography: each party holds a public key anyone may use and a private key kept secret, so two strangers can establish a shared secret over an open channel. The following year, Ron Rivest, Adi Shamir, and Leonard Adleman at MIT published the RSA scheme, whose security rests on the difficulty of factoring large numbers. (British signals-intelligence mathematicians had found similar ideas earlier, but kept them classified.)

Public-key cryptography made possible the secure exchange of keys at scale, and with it digital signatures and, eventually, the certificate infrastructure behind secure web traffic. It also changed who could do security: the field was no longer confined to cleared defense contractors. The U.S. National Bureau of Standards' Data Encryption Standard, adopted in 1977, standardized symmetric encryption for civilian use, and the open competition that produced the Advanced Encryption Standard in 2001 (designed by Joan Daemen and Vincent Rijmen) showed the community could build and vet ciphers in public.

Viruses, worms, and the antivirus industry

The 1980s brought the first self-replicating programs aimed at ordinary PCs. The term "computer virus" was coined by Fred Cohen in 1983–84, and the first widespread PC infections—the Brain virus of 1986, the Morris worm of 1988 that disabled a large fraction of the early Internet—made malicious code a public problem. The Morris worm also prompted the creation of the first Computer Emergency Response Team (CERT/CC) at Carnegie Mellon in 1988, an institutional answer to the question of who coordinates when many sites are hit at once.

Defense came from a new industry. Signature-based detection—scanning files for byte patterns known to belong to malware—was simple, fast, and effective against known threats, and antivirus products from companies founded in the late 1980s and early 1990s spread across desktops. Its weakness was structural: a signature can only catch what has already been seen, so attackers who wrote new code, or polymorphic code that changed its appearance each infection, could slip past. Heuristic and behavior-based detection followed, and the cat-and-mouse dynamic between detection and evasion became a permanent feature of the field.

Firewalls, the perimeter, and the rise of network defense

As organizations connected to the Internet in the 1990s, the boundary between "inside" and "outside" became the natural place to defend. The first packet-filtering routers and the firewall products that followed (the first commercial firewall, from Digital Equipment Corporation, appeared around 1988; stateful inspection arrived with Check Point's FireWall-1 in the early 1990s) enforced rules about which traffic could cross. The perimeter security model that grew around them—a hardened boundary, a "demilitarized zone" for public servers, and a trusted interior—was intuitive and matched how businesses thought about their buildings.

It also embedded an assumption that would later be attacked: that everything inside the perimeter is trustworthy. Intrusion detection systems, commercialized in the late 1990s, watched for signs of compromise, and virtual private networks let remote workers tunnel back inside. The model worked well enough while the network had a clear edge; it aged badly as laptops, phones, cloud services, and partners dissolved that edge.

Incident response, disclosure, and the offensive turn

Two practices matured in the 1990s and 2000s. Incident response became a discipline with its own phases—preparation, detection, containment, eradication, recovery, lessons learned—codified in handbooks and national CERTs1. Coordinated vulnerability disclosure emerged as a norm after researchers and vendors clashed over whether flaws should be published immediately. The tension was real: early disclosure could endanger users, but silence let vendors ignore bugs. The compromise—notify the vendor, allow a fix window, then publish—was pushed by figures such as Rain Forest Puppy, whose 1997 "full disclosure policy" essay proposed a five-day vendor notification, and later formalized by organizations such as the CERT Coordination Center and, from 2002, the Open Web Application Security Project (OWASP) for web flaws. Bug bounty programs, pioneered by Netscape in 1995 and later run at scale by Google, Microsoft, and platforms such as HackerOne, institutionalized the same bargain.

Meanwhile, offensive security became a profession and a training ground. Penetration testing—hired attackers probing a client's defenses—grew through the 1990s, and capture-the-flag (CTF) competitions, notably the DEF CON CTF beginning in 1996, turned exploitation into a sport and a teaching method. CTFs spread the practical skills of reverse engineering, web exploitation, and cryptography that formal curricula often missed, and they fed a generation of practitioners into both defense and offense.

The human factor and the software supply chain

By the 2000s it was clear that many failures were not technical at all. Usable security—the study of how people actually behave with passwords, warnings, and privacy settings—grew from work such as Alma Whitten and J. D. Tygar's 1999 paper on why security is hard to use, and the insight that users "route around" security that gets in their way. Research on phishing, password managers, and warning design followed, and the field's findings repeatedly showed that blaming users was less effective than designing systems that made the safe choice easy.

In parallel, the software industry tried to build security into how code is made. Microsoft's Security Development Lifecycle, introduced in 2004 after a series of damaging vulnerabilities, embedded threat modeling, code review, and testing into development, and the practice spread as the secure software development lifecycle. The 2010s added a supply-chain dimension: the 2013 Target breach, entered through a heating contractor's credentials, and the 2020 SolarWinds compromise, in which attackers inserted malicious code into a widely used update, showed that trust in software and vendors is itself an attack surface.

Zero trust and the present

The perimeter model's collapse produced its successor. In 2010, John Kindervag at Forrester Research argued for zero trust: no user or device is trusted by virtue of location, and every request is authenticated and authorized. Google's BeyondCorp papers (2014 onward) described an implementation in which employees reach internal applications through an identity-aware proxy rather than a VPN. Zero trust is less a product than a design stance—verify explicitly, grant least privilege, assume breach—and it now shapes how organizations architect access to cloud services and remote work.

Today the field's main approaches descend directly from these turning points: access control models and the reference monitor in operating systems and cloud identity; public-key cryptography in the certificate system that secures the web; detection and response in endpoint and network monitoring; disclosure and bounty programs in the vulnerability economy; secure development and supply-chain scrutiny in software engineering; and human-factors research in authentication design. Open questions remain sharp: how to verify the security of machine-learning systems and the software supply chain at scale, how to make strong authentication usable by everyone, how to govern offensive tools and vulnerability markets, and whether formal assurance can ever be cheap enough to matter outside a few high-stakes niches. The field's oldest tension—systems built for cooperation, attacked by adversaries—has not gone away; it has only grown more complex.

References

  1. Computer security (introduction, Computer security incident management), Wikipedia
  2. Penetration test (History), Wikipedia
  3. Biba Model, Wikipedia

The frameworks of Cybersecurity

Each framework is a way of working the field has used: when it held, and the question it answers.

  1. 1972-Present

    Reference monitor and trusted computing base

    How can every access be checked by one small trusted component?

    Fly to its tower in Nousfarer

  2. 1973-Present

    Multilevel security and access control models

    Who may read or write which data under what rules?

    Fly to its tower in Nousfarer

  3. 1976-Present

    Public-key cryptography

    How can two strangers share secrets over an open channel?

    Fly to its tower in Nousfarer

  4. 1983-Present

    Formal verification and evaluation criteria

    How can we prove a system actually enforces its policy?

    Fly to its tower in Nousfarer

  5. 1987-Present

    Malware detection and antivirus

    How can we spot malicious code before it does harm?

    Fly to its tower in Nousfarer

  6. 1988-Present

    Incident response

    What should be done once an attack has succeeded?

    Fly to its tower in Nousfarer

  7. 1988-2015

    Perimeter security and firewalls

    Which traffic should be allowed to cross the network boundary?

    Fly to its tower in Nousfarer

  8. 1996-Present
  9. 1997-Present

    Coordinated vulnerability disclosure

    When and how should a discovered flaw be made public?

    Fly to its tower in Nousfarer

  10. 1999-Present

    Usable security

    Why do people ignore or misuse security controls?

    Fly to its tower in Nousfarer

  11. 2004-Present

    Secure software development lifecycle

    How can security be built into software from the start?

    Fly to its tower in Nousfarer

  12. 2010-Present

    Zero trust architecture

    Should any user or device be trusted by default?

    Fly to its tower in Nousfarer

How the ideas connect

Each arrow runs from an earlier idea to the later one it shaped. The list below says what passed between them.

Reference monitor and trusted computi…Multilevel security and access contro…Public-key cryptographyFormal verification and evaluation cr…Malware detection and antivirusIncident responsePerimeter security and firewallsOffensive security and penetration te…Coordinated vulnerability disclosureUsable securitySecure software development lifecycleZero trust architecture

Beyond the field

Where Cybersecurity's ideas came from in other fields, and what they led to there: the strongest 12 of 47 links.

A city in Nousfarer: a red-carpeted stair climbing terraces to a crowned tower on an acropolis, more towers round it, forest and snowy mountains beyond.
A city in Nousfarer, the game played in the browser.

Cybersecurity is a planet in Nousfarer

Its city has a tower for each of these 12 frameworks. Land, walk in, and climb one concept at a time.

Land on Cybersecurity

Related fields

Fields its ideas are linked to

More in Cybersecurity

See where Cybersecurity sits on the Map of Human Knowledge