Hacker News Reader

Hacker News

Comment by CryoLogic | Hacker News Reader

Full thread

CryoLogic

Another poster listed the curriculum. As a security professional from a top software company I am not really impressed.

The curriculum is as follows (tech specialization, the others are worse):

1 intro to security

1 intro to policy

1 hands on lab

1 crypto

1 netsec

2 it security

choice of (2):

basic CS courses (i.e. mobile apps, operating systems)

None of that will help you get a job at any company with a serious security org. All of those courses from what I could find are the same introductory level electives you can get in your BS degree in CS.

It is not a specialized program, it is more of an introduction to security and CS at the MS level.

Things that would help:

Client XSS, Authentication / Session Management, Secure client/server architecture, defining access boundaries, phishing / social engineering, red team / blue team setup, automated vulnerability regression tests, SSDL, threat modeling, etc.

View on HN

Your list of "things that would help" are topics that (I imagine!) should be covered in intro to security, netsec, or IT security -- e.g. as 1-week or 2-week sections during the respective courses. Very few of your topics should be dedicated courses.

Agreed. Those topics are very specific instances of general principles that are usually taught in a security course.

What did OP mean by "secure client/server architecture" though? What are the basic/fundamental principles behind it, that are NOT covered in cybersec education?

I'm a grad of the OMSCS program. The "netsec" subject covered all of the "things that would help" and some in detail with practical projects.

"Client XSS, Authentication / Session Management, Secure client/server architecture, defining access boundaries, phishing / social engineering, red team / blue team setup, automated vulnerability regression tests, SSDL, threat modeling, etc. "

^ I am in no way trying to be offensive here. But if you are already a programmer, with the exception of "threat modeling" and "regression tests", it seems all those topics combined would take a week to learn?

If "learn" is "has seen and memorized a checklist", then maybe. Don't expect people that "learned" like that to be able to do much in practice with that though.

Yeah. Security is best learned after creating software for four or five years. Before that you're not going to have enough of the pieces to think about it creatively. Things turn into "rules-games" where everyone else has to tell you the rules. Also, I think studying data science is underrated for cybersec.

>Also, I think studying data science is underrated for cybersec.

Can you elaborate? I know a lot of security teams will have large amounts of machine data thrown into ELK or Splunk, but that seems like qualitative data and so there's not a lot of number-crunching to do.

Things like fraud detection come to mind. Being able to build a model based on previously detected bad actors and then use that model to detect possible additional bad actors is a specialized but useful skill in some contexts.

Statistical anomaly detection is quite valuable.

There's just so many things that are helpful from data science.

Graph Analysis => Saw some guys turn the Android codebase into a graph and use that to turn a dozen minor exploits into a chain that gave them root. Pwn2own IIRC.

Statistics => Great for detecting anomalies / understanding how to evaluate manipulation of cyber adjacent systems. For example, understanding the beta distribution lets you figure out how someone will game ratings in your app store to beat out legitimate apps with similar sounding ones. And of course all of these things cut both ways: if you're on red team it helps you masquerade more effectively.

Recommenders => Obviously useful to understand from attack detection, spam filter evasion, etc.

Linguistic analysis => De-anon attackers by the language they use. Figure out automatically which email accounts have been owned by sudden changes in speech usage.

if you have enough qualitative data it becomes quantitative data and you can start do fun stuff with that

There are a number of organizations that apply data science, ML, and HPC to support cybersec operations. The ones I am familiar with would likely consider an MS in CS to be a good entry point. A candidate that also had another substantial degree would be of even greater interest.

As a cybersec consultant with experience working with many security organizations, I wholeheartedly disagree. There is a lot more to security than things that have to do with software engineering. In my experience, the companies with the worst security are ones that primarily employ current or former software engineers as leaders of their security teams with a misguided mentality that security is just a subset of software engineering. Because of that mentality, they tend to focus on a very small portion of security principles.

For instance?

From my experience I agree. Physical, legal and network controls often get neglected.

Training and awareness, asset management, third party management, policies and procedures, legal (you should have lawyers looking over HIPAA/PCI/GDPR stuff, not a programmer), identity management, incident response, recovery & business continuity, communications plans, PR...

I could keep going. All of these things could have some software engineering component, such as with identity management: there is a component there that deals with (for example) writing code to integrate your web app with multi-factor authentication, but identity management also encompasses things like writing good access control policies, managing the legal requirements for access, training people on how to use that multi-factor auth, etc.

Engineering skills are of course of great use if you are talking about the security of a specific codebase or doing penetration testing or doing the nitty-gritty customization of a security application, but "cybersecurity" is much more than that.

Most vulnerabilities include or source from some sort of poor software engineering though, which may root from said issues, among other things. Most of these vulnerabilities (including some hardware vulnerabilities) are discovered and exploited with clever, creative software engineering / exploit development.

I know cybersec is extremely broad, but I think software engineering plays an extremely important role.

I certainly agree it's an important part of security, but that's vastly different from "you should have 4 years of engineering experience before you can start learning security".

Speaking practically, the vast majority of companies that experience one of the vulnerabilities you are referencing are not going to be patching the code themselves - they're going to be updating their infrastructure with code that was written by someone else. And before they even get to that point, they are going to have to have policies/procedures that alert them of that vulnerability, that assess how important patching that vulnerability is, that determine cost of patching/not patching, that determine how to receive that patch, and that determine how to apply that patch. And all of those things involve much more than programming.

One of the other largest parts of defending a company from being hacked is training and protecting employees from social engineering/phishing attacks, and that's something that doesn't involve writing any code (or even knowledge of code) at all.

Fair points, thanks!

> There is a lot more to security than things that have to do with software engineering.

I didn't say there weren't?

> In my experience, the companies with the worst security are ones that primarily employ current or former software engineers as leaders of their security teams with a misguided mentality that security is just a subset of software engineering.

Wasn't what I was advocating for.

I'm advocating for learning security after you have a base of software development. Not leading a security team with no experience learning or practicing security. Also, this is a huge and growing field, so I'm speaking generally, but I'm sure there are sub-areas where there are counter examples.

The "already a programmer" thing is quite a misconception here. Security has many topics and domains that have nearly no crossover with being a programmer. It's possible that many people entering this degree program have no past (or future) experience writing code.

Anecdotally, at my (one of the largest in the world) security consulting firms that hires from similar degree programs, a very small minority of people have any experience as a programmer, and a similarly small minority of our work involves writing/reading any code.

>"Another poster listed the curriculum"

I'm not sure why you are just reposting what someone else wrote, the course and curriculum are linked to in the second sentence of the article.

It's far more solid than your incredibly spartan summary:

https://pe.gatech.edu/degrees/cybersecurity

Your list of "things that would help" are likely found in:

"Introduction to Information Security (CS 6035) A full spectrum of information security: threats, software vulnerabilities, programming for malice, basic cryptography, operating systems protections, network security, privacy, data mining, computer crime."

I disagree, every year I go over a hundred of new-grads resumes from top schools and interview for dozens of security engineer and researcher positions. My basic impression is that no undergrad program has any worthwhile undergrad-level security emphasis. I would definitely give someone with such a masters degree a second or third look.

It's my observation that a master's is often a "second field of study" after getting a bachelor's. Whether that should be or not is debatable. But I know people in the field of infosec that have less experience than this, so if someone really gross the content of these courses, I'd hire them.

Yes. When I did my undergrad in CS, the people doing Masters degrees were taking many of the same classes. Different course catalog numbers, but they were sitting in the same room with me, hearing the same lectures, and doing the same homework.

As this is a masters-level course, arguably readying people for management, I would add a course or two on compliance. Too many security pros know how to secure a system but know nothing about security standards. If you cannot explain how what you have done complies with various standards, all your effort will be torn down.

A course on documentation would also be prudent. Knowing the difference between policies, standards, and procedures will help you in any interview for positions above entry level. Knowing qualitative from quantitative, and being able to talk about abstract security theory, would be good too.

And some law. Some basic course so you can talk about negligence and liability without sounding like a wikipedia page.

As non-degreed developer who is looking to transition to security work, I was initially excited to see this program but had a similar reaction after looking at the curriculum.

Potential gems on the list might be the OS or networking course if they are run something like MIT's 6.828 with lots of lab coding and perhaps a bit of hand-holding, but if you can just clone last semester's 6.828 repo and get osdev'ing, why drop 10k to spend most of your time on courses that look like yawn central?

I think what I really wanted to see was a lab-focused curriculum for aspiring vulnerability researchers, so perhaps I was bound to be disappointed.

a lab-focused curriculum for aspiring vulnerability researchers

A MS is an academic degree so of course it will be heavy on the theory, and you will mainly be expected to do practical work independently on top. Same as an undergraduate degree actually, at least a good one. That stuff you call “yawn central” is the principles that will last your entire career, the stuff you learn in the lab will be obsolete in a few years.

Your list of topics sounds like what should be covered in security in a software engineering context more so than a computer science context. In computer science I would expect more of a deep dive in to the crypto math stuff.

Some of that is definitely covered in the curriculum, at least in the classes I took at GT that are part of this.

Was it covered in a theoretical sense where you took quizzes on it or in a more hands on way where you find xss exploits and exploiting them?

My focus wasn't security, but the classes I took were hands on projects

Intro to Information Security, for example, has four hands-on projects:

1. Exploit a buffer overflow in a C program

2. Use Cuckoo to understand a malware attack

3. Implement CBC encryption algorithm and a brute-force algorithm to crack it

4. Demonstrate an XSRF, XSS and SQL injection attack

Remember this is an intro course and there are about 2-3 weeks per project so you won't be an expert but I'd say it makes you aware of some of the basic security attacks and how to prevent them.

Georgia Tech biases heavily towards hands on project design. It was also rated the worst school in terms of work life balance while I was there. Grad level courses will easily suck up a substantial portion of your life, even if the content seems 'simple.'

> None of that will help you get a job at any company with a serious security org

Most of the jobs are with non-serious orgs. But yeah, paying $10k to get a basic bootcamp on security probably isn't going to inspire tons of confidence by employers.

Thank you! Do you recommend a few books or courses that would have better / more relevant coverage for people interested in the space?

It would be great if you could share a list of recommended readings for each of the things you posted. Thanks in advance.

Reply on news.ycombinator.com