Networks dominate today's computing landscape and commercial technical protection is lagging behind attack technology. As a result, protection program success depends more on prudent management decisions than on the selection of technical safeguards. Managing Network Security takes a management view of protection and seeks to reconcile the need for security with the limitations of technology.
Last month, I disabused my audience of the notion that the 'academic' view of security is one to be scoffed at. So it's only fair that this month I lay my fair share of abuse on the academics of the world. And don't worry - this month I will do so without shame. But at the same time, I think it is really important to understand the vital role of the academics in the present and future of information protection.
You may reasonably ask why it is that I feel as if I can talk from the government view, the industry view, and the academic view. I seem to act as if I am from one or the other almost at will, as if I was somehow all three. That's because I am. While I rarely take on the view of an academic in this venue, in my spare time, I am indeed on the faculty of the University of New Haven where I teach 6 graduate courses and, at times, carry out funded research.
So I will don my academic hat for half of this article to defend the vital role of academic institutions in information protection - but before I do that...
OK - it turns out that unless you really know what you are talking about in this field, it's just about impossible to tell the difference between a stupid academic project proposed by someone who doesn't know their field and a brilliant academic who is pushing the world forward by leaps and bounds. And of course most of those who fund academics couldn't tell a brilliant academic from a used car salesman. That's why we have academics review each others proposals...
But wait a minute. Suppose our reviewers are not the brilliant academics, but rather the used car salesmen? Once we start to let the used car salesmen in, we will never get another legitimate researcher. Big problem. In fact, there is a major conflict of interest when you have reviewers selected from the pool of people you fund, because mostly they want to be funded again, so they will taint their evaluations - even if they do not intend to do so - by their views. But experts are in competition for funds, especially in academia where there is a shrinking pot of money and monies not going to one group goes to another. In zero sum games with memory, you will find that people who 'game the system' win and those who don't lose.
So my solution is simple. Have people from unrelated fields review work so that they don't know good from bad when it comes to the proposals. Then you will have a random chance of funding the real experts as opposed to the used car salespeople - which would be an improvement over what happens today. OK - perhaps this is not ideal, just a minor improvement. But I do have some suggestions...
Don't fund people who haven't done their homework. How can you tell? Pay some outside experts to do your homework before you fund anyone in a field. If it costs you $40,000 to find out about the state of the art in intrusion detection, it is a great deal compared to funding $10M a year for several years to find out what was already known many years before. That's a real example by the way. If the people you are thinking of funding don't know at least as much as the results of our national technical baseline study on the subject from a few years back, DO NOT FUND THEM. If they refuse to read such studies, don't invite their proposals. If that's too hard for you, check against the 50 Ways to Defeat Your Intrusion Detection System article and have them explain in detail to an expert how their technique avoids these attacks - or at least why it is that it will not do so and how much of an impact that has on the value of the work.
Don't fund people who haven't invested some time in the field. This can be detected by getting copies of some of their published papers on the subject. Get them, read them, understand them, compare them to the state of the art from the study you sponsored on the subject (or the examples above), and determine whether they know what they are talking about. If you need an expert to evaluate their proposals, get one who you are NOT funding EXCEPT for the evaluation. Pay the evaluators good money to spend the time needed to do a good and fair job of it and have them evaluate the people, their previous work, etc.
Do a bit of real science on your own. Test out what they say and do against people who know how to bypass their techniques. Not stupid crackers from the Internet - real experts. Perhaps some serious red teaming groups who are good at identifying problems and pointing them out. Preferable folks who aren't driven so much by their egos as by getting the right answers.
All right - so I am a dreamer... so sue me...
This is an easy one. Academics don't understand your problems because you haven't told them about your problems. And indeed, most of your real problems are probably pretty stupid and don't require an academic breakthrough to solve. For example, academics are terrible about understanding issues of the color of money. It sounds stupid to them when you say that you can't buy a $5,000 product, but you can spend 500 person hours at a loaded rate of $100 per hour evaluating it. They would ask why you don't just buy it and try it and save the $45,000 of wasted time. See how foolish they are?
That was a pause for dramatic effect... I am working on a budget these days trying to figure out how to turn money that I have but cannot spend into money that I can spend before I have to get rid of the people who can do the work. It's simple enough - there is money to do the work but the people who can do the work can't be paid by the money allocated to do the work - instead we have to hire someone who can't do the work and get rid of people who can do the work so that we can get the work done. Of course this will cause us to be unable to get the work done, so I am trying to turn the people who can do the work into people who can get paid to do the work, but of course the people who can do the work aren't qualified to do the work, while the people qualified to do the work can't do the work.
WARNING - if this makes sense to you, you need a vacation - as do I. The reason academics don't understand this sort of thing is not because they are stupid - which is not to say that they are all that smart - but rather because they are academics. They are people who have trained themselves and oriented themselves toward solving the deepest problems in their chosen fields using a set of mental and other sorts of hard won tools and tricks to do so. So here's the solution: (1) bring problems to academics that are suited to their ability to solve them, and (2) if they don't understand you it is either because you haven't explained yourself well enough or because it is not a problem they are likely to be able to solve.
That's easy too... Really good academics are really good at solving problems once and for all. That is, they are not in the business of making band aides or building a better mousetrap. They are in the business of figuring out new and better ways to limit bleeding while not exposing wounds to septic threats and finding ways to limit mouse traffic so that it doesn't do any harm or create any scares for humans. If you ask an academic to build a better mouse trap, and if they take the problem on that basis, they are either really desperate for funding or they are not really academics.
So if you want to solve problems - really solve them, and if you have the time and money required to do this task properly - then academics are probably well suited to the task. If you don't really want to solve the problem or don't have long term funding or don't have enough funding to really solve the problem, then the academics are not the right people to put to the task.
Here are some problems we might really want to solve and solve well where we are foolishly sending our money to the wrong people:
Finding a way to effectively deal with computer network attacks.
Finding an effective way to manage risks associated with computer crime on a national basis.
Developing a systematic, reliable, repeatable, and scientifically valid way to do forensics examinations involving digital systems and media which reproduces the sequence(s) of events that led to the current situation in those systems/media.
Create a new methodology for analyzing systems for vulnerabilities relative to threats and consequences and produce a systematic method and set of tools for generating analytical results comparing defensive measures.
Hopefully you get the idea.
Academia is vital for at last three things in information protection: (1) Education, (2) Research, (3) Social issues.
Education: While you might think that few would debate the role of academia in education, today, almost none of the education in information protection is done by academia. This is largely because academia has failed to take up the cause and because the government, particularly the NSA, has historically created impediments to academia in this area. The latter problem is now changing but the lack of infrastructure in the form of competent educators in this arena is creating impediments to quality undergraduate and graduate education. This will continue to be a problem until someone starts educating the existing crop of Ph.Ds in universities in this area. Such efforts have been attempted, but they have fallen short due to lack of funding. The introduction of $10M per year for a few academic institutions is a pitiful attempt by the government to change this situation.
Research: Many think that research is better done by businesses, but history has shown just how poor a choice this is. The fact of history is that almost no fruitful research has ever been done by businesses in the information protection arena, and the little useful research that has been done has focussed on optimizing specific mechanisms that are usually poor to begin with. The research is aimed at making them less poor and generally is focussed on solutions in the 6 month to 2 year time frame. Similarly, government has been funding these time frames exclusively in recent years with very few research grants running for more than a year. Historically, 5-10 year research programs have been required for real progress in complex subjects, and there is no doubt that information protection is highly complex. The total US government funding to universities in this area is probably in the range of $50M/year, which would not be so bad, except that it goes almost entirely to politically chosen institutions which are not competent to do research. Rather, they produce reports and fund others to do little projects. The other chunks of money typically go to trusted systems research, intrusion detection research, and cryptography research. This is all fine and dandy, but this represents a total lack of understanding about what is needed in research and what the future might look like.
Social issues: Universities are really good at looking at social issues, and university professors in some social sciences departments do look at these issues, but the funding and support in these issues is pitiful. Typically, the lack of cooperation between competent computer scientists and competent social scientists leads to computer science without adequate social science methodology and social science without adequate technical expertise. The result is that we are missing what is probably the most important aspect of information protection research, we do few valid experiments in this area, and we do essentially no research and development designed to address these issues.
So it looks like the areas where university research should and will ultimately play its most vital role are collapsing from benign neglect. I should point out one other really important thing. If you look at the history of information protection, you will find that almost every breakthrough that produced substantial changes came from middle-aged researchers in universities doing research funded for periods of 5 years or more. If you look at the situation today, we are practically guaranteed that these sorts of breakthroughs will not happen for the next five years and that they won't be numerous for the next 15 years. And every year we wait, the situation gets grimmer - because the total number of researchers in this area in universities is going down, the total number of Ph.D.s available to do the work is too low to sustain current levels of professors, and the best and the brightest stars who created the breakthroughs we are still depending on today, are nearing retirement.
Universities are poorly understood by industry and government and have taken a lot of abuse lately. Their ineptitude in the politics of funding has led to the movement of research dollars and quality researchers out of universities and out of this field. The side effect is that there are fewer and fewer quality researchers in information protection and they are doing less and less research and producing fewer and fewer new scholars in the field. Unless this changes, we will soon see a near-total collapse of the capability in the United states to do real research in this area. We are nearing this collapse today.
Government and industry funding have been slow, inadequate, and poorly targeted, has ignored the long term in favor of the short term, and a direct result is the increasingly staggering losses due to attacks on computer systems. We are losing scores of billions of dollars a year because of inadequate protection, and yet the total funding to stop these losses through research is less than one tenth of one percent of the losses.
The situation could not be clearer. Unless and until we start spending more money more wisely for long term research at the Ph.D. involving academic institutions with quality programs, we will continue to sink into increasingly horrendous losses. If we don't reverse the trend soon, we may even start to find that the efficiencies brought about by improvements in information technologies, which we spend billions for each year, are more than offset by the losses associated with the poor quality of protection associated with those technologies.
About The Author:
Fred Cohen is researching information protection as a Principal Member of Technical Staff at Sandia National Laboratories, helping clients meet their information protection needs as the Managing Director of Fred Cohen and Associates, and educating cyber defenders over-the-Internet as a practitioner in residence in the University of New Haven's Forensic Sciences Program. He can be reached by sending email to fred at all.net or visiting http://all.net/