https://www.nap.edu/read/1581/chapter/4

It all comes down to three words:

• Authentication
• Authorisation
• Access Control

all applied at the front door, and at the back door.

Secure applications rely on and require the protection provided by a secure operating system to assure they are not corrupted/subverted by malware. Mandatory Access Controls (MAC) for integrity and secrecy policies are designed to provide precisely this sort of protection.

Secure operating systems rely on and require the protection provided by a secure boot framework to assure that the operating system has a true and trustworthy understanding of the security properties provided by the hardware. Without assurance that from the first instruction executed as a result of asserting the RST line on the CPU, the OS cannot claim to be secure. See research on the Blue Pill BIOS/Firmware attacks for real-world examples.

Secure systems rely on and require the protection provided by formal methods and Trusted Distribution of firmware, intelligent hardware devices, and operatings systems to assure that verifiably secure kernels (i.e., those that can demonstrate to third party reviewers via techniques including code correspondence maps from source to formal specifications and back) that there are no trap doors or trojan horses (e.g., malware) lurking in the security enforcing portions of the security kernel, and that there are no opportunities to introduce such into the system during the design, engineering, development, manufacturing, delivery, update or retirement phases of the system life cycle.

For more detailed guidance see the Orange Book (Trusted Computing Security Evaluation Criteria) Division A requirements. For worked examples, see the TCSEC Evaluated Product Listings for Class A1 products, such as our own, the Gemini Trusted Network Processor (GTNP), using the Gemini Multiprocessing Secure Operating System using Intel Processors (80286 and later).

Such a Class A1 MAC-enforcing Security Kernel provides a verifiably secure foundation for running even hostile applications in such a way that for all possible programs, none can subvert the integrity and secrecy protections provided by the security kernel.

That's really all it takes, along with a sensible Trusted Network Interpretation Network Security Architecture Design in which to deploy it.

Cryptography is the ultimate solution

A paper trail is critical, given how easy it is to hack electronic systems.

Eric Amankwa - here is why Cryptography, without assurance, is not the ultimate answer: https://www.bloomberg.com/news/features/2018-10-04/the-big-hack-how-china-used-a-tiny-chip-to-infiltrate-america-s-top-companies

Assurance, which indeed may include paper trails, is a hard requirement.

Assurance includes a whole suite of procedures, business practices, and technology reviews that don't deal with features, but rather the engineering and operational characteristics surrounding the system. Refer, again, to TCSEC Class A1 prescriptions for Verified (or at least verifiable) security.

Trusted boot (including bios/UEFI firmware); Trusted distribution; Known configuration (along the path towards trustworthy hardware) (means no, you can't just pick least cost hardware components from Amazon for your system), etc.

You've raised a crucial point regarding the adoption of digital electoral systems and e-Government initiatives. While these technologies offer the potential for greater convenience and accessibility in voting processes, they also introduce significant cybersecurity challenges and risks. Here are some key considerations:

It's important for governments and election authorities to approach the adoption of digital electoral systems and e-Government initiatives with a strong focus on cybersecurity. While these technologies offer potential benefits in terms of convenience and accessibility, their successful implementation depends on addressing and mitigating cybersecurity risks to maintain the integrity of democratic processes and government services.