By The Open Group
With hacks and security breaches becoming more prevalent everyday, it’s incumbent on organizations to determine the areas where their systems may be vulnerable and take actions to better handle those vulnerabilities. Jack Daniel, a strategist with Tenable Network Security who has been active in securing networks and systems for more than 20 years, says that if companies start implementing vulnerability management on an incremental basis and use automation to help them, they can hopefully reach a point where they’re not constantly handling vulnerability crises.
Daniel will be speaking at The Open Group Baltimore event on July 20, presenting on “The Evolution of Vulnerability Management.” In advance of that event, we recently spoke to Daniel to get his perspective on hacker motivations, the state of vulnerability management in organizations today, the human problems that underlie security issues and why automation is key to better handling vulnerabilities.
How do you define vulnerability management?
Vulnerability detection is where this started. News would break years ago of some vulnerability, some weakness in a system—a fault in the configuration or software bug that allows bad things to happen. We used to really to do a hit-or-miss job of it, it didn’t have to be rushed at all. Depending on where you were or what you were doing, you might not be targeted—it would take months after something was released before bad people would start doing things with it. As criminals discovered there was money to be made in exploiting vulnerabilities, the attackers became more and more motivated by more than just notoriety. The early hacker scene that was disruptive or did criminal things was largely motivated by notoriety. As people realized they could make money, it became a problem, and that’s when we turned to management.
You have to manage finding vulnerabilities, detecting vulnerabilities and resolving them, which usually means patching but not always. There are a lot of ways to resolve or mitigate without actually patching, but the management aspect is discovering all the weaknesses in your environment—and that’s a really broad brush, depending on what you’re worried about. That could be you’re not compliant with PCI if you’re taking credit cards or it could be that bad guys can steal your database full of credit card numbers or intellectual property.
It’s finding all the weaknesses in your environment, the vulnerabilities, tracking them, resolving them and then continuing to track as new ones appear to make sure old ones don’t reappear. Or if they do reappear, what in your corporate process is allowing bad things to happen over and over again? It’s continuously doing this.
The pace of bad things has accelerated, the motivations of the actors have forked in a couple of directions, and to do a good job of vulnerability management really requires gathering data of different qualities and being able to make assessments about it and then applying what you know to what’s the most effective use of your resources—whether it’s time or money or employees to fix what you can.
What are the primary motivations you’re seeing with hacks today?
They fall into a couple big buckets, and there are a whole bunch of them. One common one is financial—these are the people that are stealing credit cards, stealing credentials so they can do bank wire fraud, or some other way to get at money. There are a variety of financial motivators.
There are also some others, depending on who you are. There’s the so-called ‘Hacktivist,’ which used to be a thing in the early days of hacking but has now become more widespread. These are folks like the Syrian Electronic Army or there’s various Turkish groups that through the years have done website defacements. These people are not trying to steal money, they’re trying to embarrass you, they’re trying to promote a message. It may be, as with the Syrian Electronic Army, they’re trying to support the ruler of whatever’s left of Syria. So there are political motivations. Anonymous did a lot of destructive things—or people calling themselves ‘Anonymous’—that’s a whole other conversation, but people do things under the banner of Anonymous as hacktivism that struck out at corporations they thought were unjust or unfair or they did political things.
Intellectual property theft would be the third big one, I think. Generally the finger is pointed at China, but it’s unfair to say they’re the only ones stealing trade secrets. People within your own country or your own market or region are stealing trade secrets continuously, too.
Those are the three big ones—money, hacktivism and intellectual property theft. It trickles down. One of the things that has come up more often over the past few years is people get attacked because of who they’re connected to. It’s a smaller portion of it and one that’s overlooked but is a message that people need to hear. For example, in the Target breach, it is claimed that the initial entry point was through the heating and air conditioning vendors’ computer systems and their access to the HVAC systems inside a Target facility, and, from there, they were able to get through. There are other stories about the companies where organizations have been targeted because of who they do business with. That’s usually a case of trying to attack somebody that’s well-secured and there’s not an easy way in, so you find out who does their heating and air-conditioning or who manages their remote data centers or something and you attack those people and then come in.
How is vulnerability management different from risk management?
It’s a subset of risk management. Risk management, when done well, gives a scope of a very large picture and helps you drill down into the details, but it has to factor in things above and beyond the more technical details of what we more typically think of as vulnerability management. Certainly they work together—you have to find what’s vulnerable and then you have to make assessments as to how you’re going to address your vulnerabilities, and that ideally should be done in a risk-based manner. Because as much as all of the reports from Verizon Data Breach Report and others say you have to fix everything, the reality is that not only can we not fix everything, we can’t fix a lot immediately so you really have to prioritize things. You have to have information to prioritize things, and that’s a challenge for many organizations.
Your session at The Open Group Baltimore event is on the evolution of vulnerability management—where does vulnerability management stand today and where does it need to go?
One of my opening slides sums it up—it used to be easy, and it’s not anymore. It’s like a lot of other things in security, it’s sort of a buzz phrase that’s never really taken off like it needs to at the enterprise level, which is as part of the operationalization of security. Security needs to be a component of running your organization and needs to be factored into a number of things.
The information security industry has a challenge and history of being a department in the middle and being obstructionist, which is I think is well deserved. But the real challenge is to cooperate more. We have to get a lot more information, which means working well with the rest of the organization, particularly networking and systems administrators and having conversations with them as far as the data and the environment and sharing and what we discover as problems without being the judgmental know-it-all security people. That is our stereotype. The adversaries are often far more cooperative than we are. In a lot of criminal forums, people will be fairly supportive of other people in their community—they’ll go up to where they reach the trade-secret level and stop—but if somebody’s not cutting into their profits, rumor is these people are cooperating and collaborating.
Within an organization, you need to work cross-organizationally. Information sharing is a very real piece of it. That’s not necessarily vulnerability management, but when you step into risk analysis and how you manage your environment, knowing what vulnerabilities you have is one thing, but knowing what vulnerabilities people are actually going to do bad things to requires information sharing, and that’s an industry wide challenge. It’s a challenge within our organizations, and outside it’s a real challenge across the enterprise, across industry, across government.
Why has that happened in the Security industry?
One is the stereotype—a lot of teams are very siloed, a lot of teams have their fiefdoms—that’s just human nature.
Another problem that everyone in security and technology faces is that we talk to all sorts of people and have all sorts of great conversations, learn amazing things, see amazing things and a lot of it is under NDA, formal or informal NDAs. And if it weren’t for friend-of-a-friend contacts a lot of information sharing would be dramatically less. A lot of the sanitized information that comes out is too sanitized to be useful. The Verizon Data Breach Report pointed out that there are similarities in attacks but they don’t line up with industry verticals as you might expect them to, so we have that challenge.
Another serious challenge we have in security, especially in the research community, is that there’s total distrust of the government. The Snowden revelations have really severely damaged the technology and security community’s faith in the government and willingness to cooperate with them. Further damaging that are the discussions about criminalizing many security tools—because the people in Congress don’t understand these things. We have a president who claims to be technologically savvy, and he is more than any before him, but he still doesn’t get it and he’s got advisors that don’t get it. So we have a great distrust of the government, which has been earned, despite the fact that any one of us in the industry knows folks at various agencies—whether the FBI or intelligence agencies or military —who are fantastic people—brilliant, hardworking patriotic—but the entities themselves are political entities, and that causes a lot of distrust in information sharing.
And there are just a lot of people that have the idea that they want proprietary information. This is not unique to security. There are a couple of different types of managers—there are people in organizations who strive to make themselves irreplaceable. As a manager, you’ve got to get those people out of your environment because they’re just poisonous. There are other people who strive to make it so that they can walk away at any time and it will be a minor inconvenience for someone to pick up the notes and run. Those are the type of people you should hang onto for dear life because they share information, they build knowledge, they build relationships. That’s just human nature. In security I don’t think there are enough people who are about building those bridges, building those communications paths, sharing what they’ve learned and trying to advance the cause. I think there’s still too many who horde information as a tool or a weapon.
Security is fundamentally a human problem amplified by technology. If you don’t address the human factors in it, you can have technological controls, but it still has to be managed by people. Human nature is a big part of what we do.
You advocate for automation to help with vulnerability management. Can automation catch the threats when hackers are becoming increasingly sophisticated and use bots themselves? Will this become a war of bot vs. bot?
A couple of points about automation. Our adversaries are using automation against us. We need to use automation to fight them, and we need to use as much automation as we can rely on to improve our situation. But at some point, we need smart people working on hard problems, and that’s not unique to security at all. The more you automate, at some point in time you have to look at whether your automation processes are improving things or not. If you’ve ever seen a big retailer or grocery store that has a person working full-time to manage the self-checkout line, that’s failed automation. That’s just one example of failed automation. Or if there’s a power or network outage at a hospital where everything is regulated and medications are regulated and then nobody can get their medications because the network’s down. Then you have patients suffering until somebody does something. They have manual systems that they have to fall back on and eventually some poor nurse has to spend an entire shift doing data entry because the systems failed so badly.
Automation doesn’t solve the problems—you have to automate the right things in the right ways, and the goal is to do the menial tasks in an automated fashion so you have to spend less human cycles. As a system or network administrator, you run into the same repetitive tasks over and over and you write scripts to do it or buy a tool to automate it. They same applies here –you want to filter through as much of the data as you can because one of the things that modern vulnerability management requires is a lot of data. It requires a ton of data, and it’s very easy to fall into an information overload situation. Where the tools can help is by filtering it down and reducing the amount of stuff that gets put in front of people to make decisions about, and that’s challenging. It’s a balance that requires continuous tuning—you don’t want it to miss anything so you want it to tell you everything that’s questionable but it can’t throw too many things at you that aren’t actually problems or people give up and ignore the problems. That was allegedly part of a couple of the major breaches last year. Alerts were triggered but nobody paid attention because they get tens of thousands of alerts a day as opposed to one big alert. One alert is hard to ignore—40,000 alerts and you just turn it off.
What’s the state of automated solutions today?
It’s pretty good if you tune it, but it takes maintenance. There isn’t an Easy Button, to use the Staples tagline. There’s not an Easy Button, and anyone promising an Easy Button is probably not being honest with you. But if you understand your environment and tune the vulnerability management and patch management tools (and a lot of them are administrative tools), you can automate a lot of it and you can reduce the pain dramatically. It does require a couple of very hard first steps. The first step in all of it is knowing what’s in your environment and knowing what’s crucial in your environment and understanding what you have because if you don’t know what you’ve got, you won’t be able to defend it well. It is pretty good but it does take a fair amount of effort to get to where you can make the best of it. Some organizations are certainly there, and some are not.
What do organizations need to consider when putting together a vulnerability management system?
One word: visibility. They need to understand that they need to be able to see and know what’s in the environment—everything that’s in their environment—and get good information on those systems. There needs to be visibility into a lot of systems that you don’t always have good visibility into. That means your mobile workforce with their laptops, that means mobile devices that are on the network, which are probably somewhere whether they belong there or not, that means understanding what’s on your network that’s not being managed actively, like Windows systems that might not be in active directory or RedHat systems that aren’t being managed by satellite or whatever systems you use to manage it.
Knowing everything that’s in the environment and its roles in the system—that’s a starting point. Then understanding what’s critical in the environment and how to prioritize that. The first step is really understanding your own environment and having visibility into the entire network—and that can extend to Cloud services if you’re using a lot of Cloud services. One of the conversations I’ve been having lately since the latest Akamai report was about IPv6. Most Americans are ignoring it even at the corporate level, and a lot of folks think you can ignore it still because we’re still routing most of our traffic over the IPv4 protocol. But IPv6 is active on just about every network out there. It’s just whether or not we actively measure and monitor it. The Akamai Report said something that a lot of folks have been saying for years and that’s that this is really a problem. Even though the adoption is pretty low, what you see if you start monitoring for it is people communicating in IPv6 whether intentionally or unintentionally. Often unintentionally because everythings’s enabled, so there’s often a whole swath of your network that people are ignoring. And you can’t have those huge blind spots in the environment, you just can’t. The vulnerability management program has to take into account that sort of overall view of the environment. Then once you’re there, you need a lot of help to solve the vulnerabilities, and that’s back to the human problem.
What should Enterprise Architects look for in an automated solution?
It really depends on the corporate need. They need to figure out whether or not the systems they’re looking at are going to find most or all of their network and discover all of the weakness, and then help them prioritize those. For example, can your systems do vulnerability analysis on newly discovered systems with little or no input? Can you automate detection? Can you automate confirmation of findings somehow? Can you interact with other systems? There’s a piece, too—what’s the rest of your environment look like? Are there ways into it? Does your vulnerability management system work with or understand all the things you’ve got? What if you have some unique network gear that your vulnerability management systems not going to tell you what the vulnerability’s in? There are German companies that like to use operating systems other than Windows and garden variety Linux distributions. Does it work in your environment and will it give you good coverage in your environment and can it take a lot of the mundane out of it?
The challenge is a lot of people push back against high information flow because they can’t make sense of it; they can’t ingest the data, they can’t do anything with it. It’s the challenge of accepting and sharing a lot of information. It doesn’t matter whether vulnerability management or lot analysis or patch management or systems administration or back up or anything—the challenge is that networks have systems that share a lot of data but until you add context, it’s not really information. What we’re interested in in vulnerability management is different than what you’re automated backup is. The challenge is having systems that can share information outbound, share information inbound and then act rationally on only that which is relevant to them. That’s a real challenge because information overload is a problem that people have been complaining about for years, and it’s accelerating at a stunning rate.
You say Internet of Things, and I get a little frustrated when people treat that as a monolith because at one end an Internet enabled microwave or stove has one set of challenges, and they’re built on garbage commodity hardware with no maintenance ability at all. There are other things that people consider Internet of Things because they’re Internet enabled and they’re running Windows or a more mature Linux stack that has full management and somebody’s managing it. So there’s a huge gap between the managed IoT and the unmanaged, and the unmanaged is just adding low power machines in environments that will just amplify things like distributed denial of service (DoS). As it is, a lot of consumers have home routers that are being used to attack other people and do DoS attacks. A lot of the commercial stuff is being cleaned up, but a lot of the inexpensive home routers that people have are being used, and if those are used and misused or misconfigured or attacked with worms that can change the settings for things to have everything in the network participate in.
The thing with the evolution of vulnerability management is that we’re trying to drive people to a continuous monitoring situation. That’s where the federal government has gone, that’s where a lot of industries are, and it’s a challenge to go from infrequent or even frequent big scans to watching things continuously. The key is to take incremental steps, and the goal is, instead of having a big massive vulnerability project every quarter or every month, the goal is to get down to where it’s part of the routine, you’re taking small remediated measures on a daily or regular basis. There’s still going to be things when Microsoft or Oracle come out with a big patch that will require a bigger tool-up but you’re going to need to do this continuously and reach that point where you do small pieces of the task continuously rather than one big task. That’s the goal is to get to where you’re doing this continuously so you get to where you’re blowing out birthday candles rather than putting out forest fires.
Jack Daniel, a strategist at Tenable Network Security, has over 20 years experience in network and system administration and security, and has worked in a variety of practitioner and management positions. A technology community activist, he supports several information security and technology organizations. Jack is a co-founder of Security BSides, serves on the boards of three Security BSides non-profit corporations, and helps organize Security B-Sides events. Jack is a regular, featured speaker at ShmooCon, SOURCE Boston, DEF CON, RSA and other marque conferences. Jack is a CISSP, holds CCSK, and is a Microsoft MVP for Enterprise Security.
Join the conversation – @theopengroup #ogchat #ogBWI