Computer security

From Ganfyd

Jump to: navigation, search

Information is valuable and the information on its computer systems if lost, corrupted or stolen can certainly break small businesses and make life distinctly uncomfortable for personal computer users. The security of bank account access is the most common high financial risk area for most computer users. However there are other risks with data and doctors in particular need to be aware of their responsibilities to secure patient identifiable data.

For your convenience articles on computer security are indexed here:

Articles on encryption, network security and secure communication are awaited. Users setting up solutions for others, as an administrator of either a home or enterprise network should have a core understanding of why all operating system default configurations can be usefully hardened (see external links below). This is a complex area but all should be aware that their default hardware storage with consumer operating systems is likely to be accessible by anyone with physical access to the hardware storage and that basic internet and email communication is by plain text, open to trivial man in the middle reading. Where the operating system makes it possible internet and email use should be not be done by a user with administration (root) rights. Ransomware is another evolving threat.

Often you will find people claiming one operating system is more secure than others. This may be fairly irrelevant if the hardware, the computer itself is stolen or destroyed. There is little doubt however that ease of function can impair security, some have better default design than others, and a balance has to be struck. It has long been known that the most secure operating system is one where the full source code is available and in particular the compiler source code has been compiled after inspection for deliberate back doors[1]. This of course has the catch that the code has to be looked at and understood. It has the other catch that you can still design in relative security holes and some hardware security holes can not be mitigated fully by software design. There are as of 2016 over 15 million lines of code in the linux kernel, over 13 million lines of code in Firefox and 7 million in Libre-office. This compares with Windows 7 that was 40 million lines and Windows 10 is easily more but of course not yet officially admitted to. Ultimately executed code is a matter of trust and convenience, given its complex dependencies in any modern computer. Also any operating system and its associated software can be insecure if not updated against faults, and is open to compromise by design or malware such as rootkits which are given for whatever reason full administration rights over the system. If you run a closed system updates are not as important as one with open interfaces such as internet connectivity or USB ports. You can sometimes be unpleasantly surprised by default boot configurations of consumer hardware. For example booting from USB may be enabled, which essentially means anyone with access to a USB port can access all the binary data in your computer. If this is not securely encrypted it can be read by many. A similar comment applies to hard disks that are not physically destroyed. Problematically to install more secure operating systems you often have to compromise other security features in a computer, not all of which may be convenient to re-enable.

In practice all consumer operating systems (eg MS Windows, Apple Mac, iOS, Android and Linux ) and their associated internet aware software need regular updating and the update load is considerable for all. There is also little doubt that out of the box, but connected to the internet, MS Windows is a relatively insecure system. However as a closed system it may well have the most functional software. The main problem is keeping it a closed system, as its friendly networking, internet and other interactions may be easy to exploit and are now part of making it functional.

The internet, email and cloud storage (as now encouraged by default now by major vendors) may seem to offer reasonably secure solutions but the protocols are imperfect and minor implementation errors can easily create issues. Yet alone the issue with even with the highest grade encryption, that unless you and only you, have physical access to the hardware kept in faraday cages at either end of a connection, the data could go astray.

Routers and switches can have insecure operating systems or configurations. As of 2017 most routers should have UPnP actively disabled. UPnP has long known issues as it allows a device on the network to configure the router. This puts all computers on networks at increased risk once a sophisticated malware load gets into the network say via an internet of things device that might be pawned because it used UPnP in first place to set up port forwarding. You now would have a trusted network that is as dangerous as any untrusted network such as the internet itself ! However with the increasing popularity of peer-to-peer applications, such as game servers, and VoIP programs to get equivalent functionality with manual configuration can be very inconvenient.


External link