From a pure cyber security and technical point of view the most important aspects are: (1) The detection rate and (2) The width of coverage (how much attack surface is protected).
For the first one, it is unfortunately very difficult to assess the detection rate of a solution unless you are an expert with a large dataset of threats (known and unknown) at your disposal to benchmark the solution against. In any case, you should make sure the solution is capable of detecting unknown and novel threats - this is, the solution must go beyond heuristics and possess a profound understanding of cyber threats.
Second, the width of coverage means that the solution covers a large number of threat verticals but more importantly is deployed at anywhere where a threat may appear. In several cases, customers do not cover all the areas of their network.
Search for a product comparison in Intrusion Detection and Prevention Software (IDPS)
So.. the technical aspects of your ID should be based on your environment.
So not trying to oversimplify or downplay the need for the IDS to meet your technical thresholds, but those get discussed a lot. Operational impacts/criteria are often overlooked, in my opinion. I think the key factors are usability (how easy it is to train my folks to use it? how easy is it to integrate it into my operational processes, etc.) and interoperability with my current security ecosystem.
If I have to do a lot of tweaking to get it to work or I need a workaround to get feed accuracy then I'm probably not inclined to pursue a product.
Tuning is one thing. I expect that, but I don't expect the coding equivalent to duct tape and bubblegum to get it to talk and interact with my SIEM, SOAR, etc.
- Capabilities, if we don't understand what these are it is unlikely we will have a success story.
- The expertise to operate
- Product documentation
- Training provided by a supplier
- Best practices
- Successful use case scenario (ideally from the same industry),
- Pricing (matters for local gov), etc.
Documentation. Algorithmic transparency. Ability to get someone smart on the phone FAST at the vendor, without going through gatekeepers. Confidence levels (statistical validity).
Find out what your peers are saying about Darktrace, Vectra AI, Check Point Software Technologies and others in Intrusion Detection and Prevention Software (IDPS). Updated: March 2024.
What Is an Intrusion Detection System? Intrusion detection systems (IDSes) analyze network traffic for signatures of known attackers. The systems can be hardware devices or software solutions. An IDS can mitigate existing malware, such as backdoors, rootkits, and trojans.The goal of an intrusion detection system is to detect an attack as it occurs. The system starts by analyzing inbound and outbound network traffic for signs of known attackers.
Some activities an IDS performs...
From a pure cyber security and technical point of view the most important aspects are: (1) The detection rate and (2) The width of coverage (how much attack surface is protected).
For the first one, it is unfortunately very difficult to assess the detection rate of a solution unless you are an expert with a large dataset of threats (known and unknown) at your disposal to benchmark the solution against. In any case, you should make sure the solution is capable of detecting unknown and novel threats - this is, the solution must go beyond heuristics and possess a profound understanding of cyber threats.
Second, the width of coverage means that the solution covers a large number of threat verticals but more importantly is deployed at anywhere where a threat may appear. In several cases, customers do not cover all the areas of their network.
So.. the technical aspects of your ID should be based on your environment.
So not trying to oversimplify or downplay the need for the IDS to meet your technical thresholds, but those get discussed a lot. Operational impacts/criteria are often overlooked, in my opinion. I think the key factors are usability (how easy it is to train my folks to use it? how easy is it to integrate it into my operational processes, etc.) and interoperability with my current security ecosystem.
If I have to do a lot of tweaking to get it to work or I need a workaround to get feed accuracy then I'm probably not inclined to pursue a product.
Tuning is one thing. I expect that, but I don't expect the coding equivalent to duct tape and bubblegum to get it to talk and interact with my SIEM, SOAR, etc.
- Capabilities, if we don't understand what these are it is unlikely we will have a success story.
- The expertise to operate
- Product documentation
- Training provided by a supplier
- Best practices
- Successful use case scenario (ideally from the same industry),
- Pricing (matters for local gov), etc.
Education, documentation, use cases and best practices.
Documentation. Algorithmic transparency. Ability to get someone smart on the phone FAST at the vendor, without going through gatekeepers. Confidence levels (statistical validity).