variety of security notions
introduction#
learning network security, i had to write a post related to it
this post aims to learn or clarify hosts & networks security notions/jargon, not covering kinds of threats or attacks
i used simpler words than the ones found in my research to make it easier to read for non-native english speakers
i am not an expert by any means, please let me know if i’ve said something wrong
glossary#
defining mandatory concepts related to the notions covered
malware#
malwares are malicious piece of code or software designed to harm or hijack a device or its data by any means
payload#
payload is the part of a malware who responsible for the damages - data exfiltration, making a host unusable, etc.
vulnerability#
vulnerabilities refer to hardware, software or procedures weaknesses that could be exploited by a threat
threat#
threats are malicious or negative potential events exploiting known or yet unknown vulnerabilities
the word threat actor
comming from it refers to people behind a malicious incident
risk#
risks qualifies the probability that a threat exploits a vulnerability causing a critical damage to the host or its neighbours
risk = threat * vulnerability * damage
attack#
attacks are the usage exploitation of a vulnerability by a threat actor
classification for those are seperated, e.g: human threat, viruses…
threat model#
threat modeling is the process of identifying potential vulnerabilities or security flaws, prioritising weaknesses to address or mitigiate to minimize the risks
endpoint#
endpoints are the farrest devices on a network comming from the outside, can be hosts or servers
endpoint protection#
are covered various protections for endpoints/hosts according to many types of threats & attacks
i only wrote about relevant & still active protection solutions
hardware side#
fde#
on the hardware side, full-disk encryption
is a very good practice to preserve security & privacy for portable devices
having Luks
for all kinds of needs & BitLocker
for windows OSs
the better & common way to do fde is by using the tpm trusted platform module chip to generate the encryption keys & keeping part of it to itself
additionnaly for luks, it uses a master key asked before the boot sequence using a passphrase hash to boot into the OS
dlp#
to minimise data loss (i.e. “availability” in production use), the threat model could implement a data loss prevention
procedure
a usefull data loss model could be the 3-2-1 backup strategy
- 3 copies of the data - (or more)
- 2 backups on different storage media - really helps
- 1 backup copy offsite - can be cloud, nas…
for personnal use, backuping on two different medias (e.g: a nas & a disk or cloud) can do the job, but please do not underestimate the value of backups in production use
once an host has been infected or is showing signs to, doing a quick & tested restoration is very usefull & saves time
software side#
authorisation#
authorisation can be associated to permissions
a good practice is to always let the minimal permissions to the users, restricting them to do only what they are intended to
that can be a part of the threat model
: who can access which ressources
in other words, when an user is compromised -> what can he access, so what became at risk?
disabling the root account is also a good practice for most hosts, prefering a sudoer or proper user permissions
as always, good passwords are always a most & for the ssh protocol the usage of keys or certificates in highly recommended
authentication#
using a login & a password cannot verify the identity of the person accessing a ressource for that user
since then, human intervention has guaranteed the identity of the person accessing the resource
back then, simple questions where asked to know if the intended person using the credentials was the one intended - e.g name of its dog, where did he was born, etc.
this authentication method was highly subjected to doxing/osint
nowadays, 2fa is used, living on the intended person’s phone or an dedicated hardware device (yubikey)
2fa can take the form of push notifications (malicious ones can be injected), sms verifications (warning sim swapping attack method) or authenticators codes using the totp protocol
mfa (multifactor authentication) is also a thing
os/software side#
epp#
endpoint protection platform
define the suite of technos or solutions used to protect endpoints
ng-av/edr#
antivirus (av), next gen antivirus (ngav) & endpoint detection & response (edr)
are commonly used solutions to protect endpoints
sources i found says different things about them, so i put ng-avs & edrs together, i wonder if their names are not just a marketing thing for the same solutions
“legacy avs” are based in signature recognition to stop known malware file
an individual hash could be generated for each file. standard avs compare them to a list of malicious files hash to know if the checked file is one of them or not to flag it
it is only working against file-based attack & new or yet unknown malwares, otherwise it could not be discovered using this method
variations of a malware (malformed sinature trick) can also be done, so its bypass the hash check since it is not in the signature database
ngav use behaviour detection on top of the signature recognition, so if a software/program/service activity is suspicious -> the file or its activity can be put in quarantine or be stopped
some may introduce sandboxing & ai - machine learning although av & ngav are already well ressources hungry
edr & ng-avs are very important security solutions since only the endpoint can see the unencrypted ongoing or incomming traffics (e.g. https traffic)
be aware that more than one av could lead to more ressource usage & them trying to cancel each other, since they are accessing same files & seeing each other activity
network solutions#
network solutions are preferable so threats or attacks are stopped before reaching the endpoints
firewall#
firewalling protects networks from unwanted traffic by setting a set of pre-programmed rules
it can also provide a network segmentation, separating the lan local area network into smaller ones w/ their dedicated rules
not to compare w/ software firewalls who applies rules to an host applications only
proxy#
proxy servers could be an intermediate to access the internet in a lan (local area network)
very usefull to reduce a network attack surface since all the traffic is going through it
it can monitor traffic or gather metrics
it also provide sort of firewalling since you are restricted by what the proxy permit you to access to
it is also great for privacy since hosts are not directly exposed, the proxy is
many use of proxies can be found doing research
reverse proxy#
reverse proxies act the same as normal proxies but for incomming traffic
endpoints are behind the reverse proxy so that all incomming connexions need to pass through the reverse proxy to access the hosts
the advantages are the same
ids & ips#
intrusion detection systems & intrusion protection systems
the ids & the ips analyse real-time traffic for signature matching known attacks or suspicious behaviour
the difference between them is that ips can act as a hardware switch to cut a malicious traffic whereas the ids only raise alerts
they are oftenly shipped inside a firewall by some companies
soc#
security operations center or isoc information security operations center
is the structure (people, room, screens & devices) where logs are gathered & correlated
people are present at full-time to maintain the soc since it is a very important protection mesure (the ciso chief information security officer, analysts, devops/secdevops teams…)
the soc integrate various solutions such as a siem or a soar for example
the soc team makes decisions to act on the feedback according to the logs activity
ndr/xdr#
network detection & response and extended detection & response
the ndr monitor network layer 2-7 traffic, no agent on the endpoints
xdr tend to gather more informations by installing agents on endpoints to gather data
xdr seems to be more corporate solutions & focus on properitaty
ndr can be implemented on its own but a xdr may cause friction if it’s not the only protection system deployed
siem#
security information & event manager
is offenly used in a soc environment, it gather, centralize & organize all logs from various devices
logs gathered from the firewalls, network appliances, ids… can be filtered by the siem since all their informations aren’t always relevant
the siem is: collecting, aggregating, identifying, categorising & analysing incidents or events
the siem needs continuous learning by the security team (this report is normal because we know […], it is current that […]…) or by ai (machine learning) to keep categorising the data well but that has more to do with a soar
soar#
security orchestration, automation & response
go a step further than the siem, taking advantage of the automation
doing the same job as a siem but go a step futher by automating and orchestrating time-consuming manual tasks of the secops team, so they can speed up on real incident response time