How the Vietnam War created single sign-on — and how it's evolved since

How the Vietnam War created single sign-on — and how it's evolved since

Speakers: Connor Peshek

Conference: BSides SF

YouTube: https://www.youtube.com/watch?v=iXkK6bl42R0

Overview

In this fascinating talk, Connor Peshek, a seasoned engineer with experience at Kroger, Cloudflare, and Authentic, delves into the unexpected origins and intricate evolution of single sign-on (SSO). Titled "How the Vietnam War created single sign-on," the presentation uncovers a direct lineage from classified military communication systems of the mid-20th century to the ubiquitous authentication protocols we rely on today. Peshek argues that the fundamental principles of SSO – authenticating once at a central location to gain access across multiple trusted services – were first conceived out of a critical need for secure and efficient military communications during wartime.

The talk meticulously traces this journey, beginning with rudimentary analog scramblers of World War II, through the complex encrypted networks of the Cold War, and ultimately to the digital revolution that birthed Kerberos, SAML, and OpenID Connect. Peshek's narrative highlights the continuous interplay between military necessity, academic research, and commercial innovation that shaped these foundational technologies. For anyone interested in the often-overlooked history behind modern cybersecurity and authentication, this talk offers a compelling and detailed account of how seemingly disparate events converged to create one of the internet's most critical security mechanisms.

Background

▶ Watch: Talk introduction and the Vietnam War SSO premise (0:00)

The journey to SSO begins not in the digital realm, but in the analog world of World War II cryptography. Early voice communication security relied on devices like the A-3 scrambler, which merely performed frequency inversion – turning high frequencies to low and vice versa. As Peshek notes, Bell Labs in the 1970s famously stated that "any basic high school boy could unscramble something like the A-3 scrambler," highlighting its severe limitations. This vulnerability became acutely apparent when a critical message about the impending Pearl Harbor attack, deemed too sensitive for the A-3 scrambler due to interception fears, was sent via slower coded radio telegraph, arriving too late.

This catastrophic failure spurred the US military to demand a truly secure communication solution from Bell Labs. The result was the Green Hornet, later known as SIGSALY. This groundbreaking system introduced the vocoder, which converted analog voice signals into digital data, allowing for actual encryption. SIGSALY used one-time vinyl records for encryption and decryption, with identical copies played synchronously at each end and then destroyed. Though immensely secure (never known to have been cracked), SIGSALY was prohibitively large (over 50 tons), expensive, and impractical for widespread use, being decommissioned immediately after WWII.

In the post-WWII era, leading up to the Vietnam War, the AUTOSEVOCOM network was developed, featuring encrypted military communication devices like the KY3. These devices were still large, a "hodgepodge of a bunch of devices shoved together into a box," and relied on single key encryption in a direct device-to-device manner. This meant that every KY3 on the network needed to hold the encryption key for every other KY3 it might communicate with. This key management system scaled terribly, becoming incredibly expensive and inefficient as the network grew.

The Vietnam War brought these communication vulnerabilities into sharp focus. The US military was losing battles at an alarming rate, despite no apparent classified document or code leaks. The NSA's Operation Purple Dragon uncovered that the Vietnamese were simply paying close attention to unencrypted radio communications, even those using code words. By deciphering who was talking to whom and observing supply movements, they could predict US maneuvers. This operation is credited with establishing the principles of OpSec (Operational Security). The core takeaway from Operation Purple Dragon was the urgent need for encrypted phone communication where every phone could securely call every other phone. This pressing military requirement directly paved the way for the invention of SSO.

Key Findings

▶ Watch: Green Hornet/SIGSALY: the first digital encrypted voice (3:00)

The central revelation of Connor Peshek's talk is the direct, albeit classified, origin of single sign-on (SSO) in a Vietnam War-era military communication system called the Bell Field. This system, designed by the head of R&D at the NSA, solved the intractable key management problems of its predecessors by introducing a central server that held all device keys. When a device made a call, it authenticated with this central server, which then generated a unique, one-time session key encrypted with the keys of both communicating devices. This marked the first instance where a single authentication with a central authority granted secure access for a specific communication, perfectly aligning with the modern definition of SSO.

Peshek then highlights a critical, almost accidental, transmission of this classified concept into the public domain. Jerome Saltzer, an MIT professor consulting for the NSA, recognized the Bell Field's potential to revolutionize computer access at MIT, where students struggled with mainframe-based systems. Unable to disclose the classified details, Saltzer got a "lucky break" when an NSA member gave a vague description of the Bell Field at a conference. Saltzer seized this opportunity, tasking Needham and Schroeder to research it, leading to the development of the Needham-Schroeder protocol. This protocol, in turn, became the foundation for Kerberos, the first widely adopted implementation of SSO in computer networks, essentially a "clone" of the Bell Field system.

The talk further underscores the profound impact of Kerberos through Project Athena at MIT. This ambitious project transformed computing for students, moving from restrictive mainframe terminals to a network where users could log into any machine and access remote files, messaging (Zephyr, an AOL Instant Messenger clone), and collaborative documents. Project Athena was a crucible for early internet technologies, contributing to the development of X Windows for Linux and one of the first DNS servers.

The narrative then transitions to the internet era, where the need for cross-domain SSO led to the creation of SAML (Security Assertion Markup Language), the first internet-focused SSO protocol, built upon XML. The subsequent browser wars and Microsoft's introduction of the XML HTTP request (which unexpectedly birthed modern web apps) led to the abandonment of XML in favor of JSON. This shift, coupled with the growing problem of insecure authorization practices (e.g., applications asking for full email credentials), propelled the development of OpenID (initially for blog verification) and eventually OAuth and OpenID Connect, the prevalent SSO and authorization protocols of today's web.

Technical Deep Dive

▶ Watch: AUTOSEVOCOM and KY3: early network, scaling key management (5:20)

The evolution of single sign-on is a story of continuous technical innovation, each step driven by the limitations of its predecessors.

The earliest attempt at secure voice communication, the A-3 scrambler, was fundamentally flawed. It employed a technique called frequency inversion, where high-pitched sounds were converted to low, and low to high. While it garbled speech, it offered no true encryption and could be easily unscrambled, as noted by Bell Labs. It was merely "better than nothing," and its insecurity had severe consequences, as seen with the Pearl Harbor warning.

The SIGSALY (codename Green Hornet) represented a monumental leap. Developed by Bell Labs, it was the first system to use a vocoder, a device that converts analog speech into a digital signal. This digitization was crucial because it allowed for actual encryption, unlike the A-3 scrambler. SIGSALY's encryption mechanism involved two synchronized vinyl records, one at the sending end to encrypt the digital signal, and an identical one at the receiving end to decrypt it. These records were one-time use and destroyed after each session, ensuring extreme security. However, its immense size (over 50 tons), complexity, and cost limited its deployment to only 12 units.

Post-WWII, the AUTOSEVOCOM network introduced the KY3, a more compact but still substantial device. The KY3 integrated a vocoder and a KG-13 for encryption, but its core flaw lay in its single key encryption protocol. For two KY3s to communicate securely, each device had to possess the specific encryption key for the other. This meant that on a network of 'N' devices, each device needed to store 'N-1' keys, leading to an N-squared scaling problem for key management. This system was "really not time effective, very frustrating to use," and proved unsustainable for large-scale, dynamic communication needs.

The breakthrough arrived with the Bell Field system during the Vietnam War. This system introduced the concept of a central authority for key management and session establishment. Instead of direct device-to-device key exchange, a device (Phone A) would first send a message, encrypted with its own key, to a central server. This central server, holding the keys for all devices on the network, would decrypt Phone A's message, verifying its authenticity. It would then generate a unique, one-time session key. This session key was then encrypted separately with Phone A's key and Phone B's key, and sent to both phones. Phones A and B could then use their respective copies of the session key to establish a secure, encrypted communication channel. After the call, the session key was destroyed. This mechanism, where a single authentication with a central server granted temporary, secure access to another resource, is precisely what defines single sign-on.

Though classified, the Bell Field's principles were eventually adopted into the computing world. Jerome Saltzer, an MIT professor, recognized its potential. Influenced by a vague public description, he tasked Needham and Schroeder, leading to the Needham-Schroeder protocol. This protocol became the basis for Kerberos, a widely adopted network authentication protocol. Kerberos operates similarly to the Bell Field, but introduces a Ticket Granting Server (TGS). A client authenticates with an Authentication Server (AS), which then issues a Ticket Granting Ticket (TGT) from the TGS. This TGT can then be used to obtain service tickets for various services within the network, granting seamless access without re-authentication. Project Athena at MIT was a monumental early implementation of Kerberos, demonstrating its power in enabling campus-wide, hardware-agnostic computer access. It facilitated technologies like Zephyr (an instant messenger) and remote file access, laying groundwork for modern distributed computing.

As the internet grew, the limitations of Kerberos (e.g., UDP-based, no native HTTP support) became apparent for cross-domain web authentication. This led to the development of SAML (Security Assertion Markup Language), the first SSO protocol designed explicitly for the internet. SAML leverages XML for exchanging authentication and authorization data between an Identity Provider (IDP) (which authenticates the user) and a service provider (the application the user wants to access). SAML requires pre-defined rules, including where digital signatures will be placed, whether the communication is encrypted, the URLs for exchanging SAML tickets, and the binding methods (e.g., HTTP POST or Redirect). Despite its corporate popularity, Peshek notes that many SAML integrations are "not to spec, aren't well set up," leading to potential insecurities.

The browser wars of the mid-1990s and Microsoft's "embrace, extend, extinguish" strategy profoundly impacted web technology. Microsoft's introduction of the XML HTTP request was intended to extend browser capabilities beyond static documents, paradoxically giving birth to modern web apps. However, developers quickly found XML cumbersome for dynamic web interactions, favoring JSON (JavaScript Object Notation) due to its lighter weight and native compatibility with JavaScript. This shift led to a decline in SAML's appeal for consumer-facing web applications.

This environment paved the way for OpenID and later OAuth and OpenID Connect (OIDC). The original OpenID aimed to solve blog comment verification by allowing users to prove their identity via a URL. The more significant challenge, however, was authorization, exemplified by early Facebook practices of asking for users' Gmail passwords to sync contacts. OAuth emerged to solve this by providing a secure delegation framework, allowing users to grant limited access to their resources without sharing credentials. OpenID Connect builds on OAuth 2.0, adding an identity layer that enables clients to verify the identity of the end-user and obtain basic profile information, making it the de facto standard for modern web and mobile SSO. While OIDC shares core practices with its predecessors—establishing trust, proper configuration, and token exchange—it is designed for the distributed, JSON-centric world of today's internet.

Demo / Proof of Concept

▶ Watch: Vietnam War: Operation Purple Dragon and the origin of OpSec (6:40)

This talk is a historical narrative tracing the lineage of single sign-on technologies. As such, it did not include a live demonstration or proof of concept of any of the discussed protocols or systems. The speaker focused on illustrating the historical context and technical evolution through detailed explanations and historical anecdotes.

Defensive Implications

▶ Watch: Crucial takeaway: the imperative for fully encrypted phone communications (7:50)

Understanding the historical evolution of SSO provides crucial insights for modern defensive strategies. The journey from the A-3 scrambler to OIDC highlights a continuous struggle to achieve confidentiality, integrity, and availability in authentication and communication, often balancing these with usability and cost.

1. Secure Centralized Authentication is Paramount: The Bell Field system, and its successor Kerberos, introduced the concept of a central authority for authentication. This design reduces the burden of key management and provides a single point of control for access. However, it also makes the central authentication server a critical target. Defenders must prioritize the security of their Identity Providers (IDPs), Key Distribution Centers (KDCs), or Authentication Servers with the strongest possible controls, including multi-factor authentication, robust access controls, and continuous monitoring.

1. Protocol Adherence and Proper Configuration are Non-Negotiable: Connor Peshek explicitly noted that many SAML integrations are "not to spec" or "not well set up," leading to bugs and insecurity. This observation is a critical defensive takeaway: simply using a protocol like SAML or OIDC does not guarantee security. Implementations must strictly adhere to protocol specifications, including proper handling of digital signatures, encryption, assertion validation, and token lifetimes. Misconfigurations are a common attack vector (e.g., signature bypasses, XML external entity (XXE) attacks in SAML). Regular audits of SSO configurations are essential.

1. Session Management and Key Rotation: The Bell Field's use of one-time session keys (destroyed after a call) was a highly secure practice. While modern SSO protocols use longer-lived tokens, the principle of limiting their lifespan and enabling revocation remains vital. Defenders should implement short token lifetimes, robust token revocation mechanisms, and secure storage for session tokens to minimize the impact of compromise.

1. Authorization vs. Authentication: The shift from early, insecure authorization practices (e.g., asking for plaintext passwords) to OAuth and OpenID Connect underscores the importance of proper authorization delegation. Defenders must ensure that applications only request the minimum necessary scopes/permissions from users and that these permissions are properly enforced. Misconfigured authorization can lead to privilege escalation or unauthorized data access, even if authentication is strong.

1. Understanding the Attack Surface of Modern Web Apps: The rise of XML HTTP Request and JSON-based modern web apps dramatically expanded the attack surface for authentication and authorization. Client-side logic, API endpoints, and cross-origin communication introduce new vulnerabilities. Defenders must employ secure coding practices, implement robust API security, and protect against common web vulnerabilities like Cross-Site Scripting (XSS), Cross-Site Request Forgery (CSRF), and injection attacks, which can compromise tokens or session integrity.

1. Continuous Evolution and Problem-Solving: The history of SSO shows a constant adaptation to new technologies and emerging threats. Defenders should remain vigilant about new security challenges and the evolution of authentication standards (e.g., OIDC 2.1). The question "What's going to replace OIDC?" highlights the need for ongoing innovation to address future problems that current protocols may not adequately solve.

In summary, the historical context provided by Peshek's talk emphasizes that foundational security principles, such as centralized trust, secure key exchange, and proper protocol implementation, have been critical since the earliest days of encrypted communication. Neglecting these principles, even with modern protocols, will inevitably lead to vulnerabilities.

Key Takeaways

• SSO's Military Origins: The concept of single sign-on (SSO) was first conceived during the Vietnam War with the classified Bell Field system, designed to solve complex key management problems for secure military phone communications by using a central server and one-time session keys.
• Direct Lineage to Kerberos: The Bell Field's architecture was indirectly exposed and subsequently reverse-engineered by Jerome Saltzer, leading to the Needham-Schroeder protocol, which formed the basis for Kerberos. Kerberos, with its Ticket Granting Server (TGS), essentially cloned the Bell Field's centralized authentication model for computer networks.
• Project Athena's Impact: Project Athena at MIT, a pioneering implementation of Kerberos, revolutionized campus computing by providing seamless access to shared resources from any workstation, contributing significantly to the development of early internet technologies like X Windows for Linux and DNS servers.
• Evolution Driven by Technology Shifts: SSO protocols evolved in response to changing technological landscapes: from local network solutions like Kerberos to internet-focused SAML (using XML), and then to OAuth and OpenID Connect (using JSON) to support modern web apps and address complex authorization challenges.
• Importance of Protocol Adherence: The speaker highlighted that many SAML implementations are "not to spec" or "not well set up," leading to potential insecurities. This underscores the critical importance of strictly adhering to protocol specifications and proper configuration for any SSO solution.
• Solving Authorization Challenges: The shift towards OAuth and OpenID Connect was largely driven by the need to securely delegate authorization (allowing applications limited access to user data) without requiring users to share their primary credentials, a significant improvement over earlier, insecure practices.

About the Speaker(s)

Connor Peshek is a self-taught JavaScript engineer with a diverse background in software development and infrastructure. He began his career at Kroger, where he contributed to the development of ClickList, one of the earliest grocery delivery and pickup services. Following his time at Kroger, Peshek co-founded a startup that was later acquired by Cloudflare, where he worked for approximately three and a half years. Currently, he is an engineer at Authentic, a company focused on providing open-source, self-hostable alternatives to identity management solutions like Okta. In this role, he specializes in SAML logic and regularly performs DevRel demos to showcase new features and integration capabilities. His curiosity about the history of authentication led him to research and present this talk, tracing the origins of SSO.