Juggling Risk in the Cloud Security Circus
Today's headlines point to more sophisticated and large-scale and malicious online activities. For some folks, therefore, the consensus seems to be that the cloud computing model and vision are not up to the task when it comes to security.
But at the RSA Conference earlier this year, a panel came together to talk about security and cloud computing, to examine the intersection of cloud computing, security, Internet services, and Internet-based security practices to uncover differences between perceptions and reality.
The result is a special sponsored BriefingsDirect podcast and video presentation that takes stock of cloud-focused security -- not just as a risk, but also as an amelioration of risk across all aspects of IT.
Join panelists Chris Hoff, director of cloud and virtualization Solutions at Cisco Systems; Jeremiah Grossman, the founder and chief technology officer at WhiteHat Security; and Andy Ellis, the chief security architect at Akamai Technologies. The discussion is moderated by Dana Gardner, principal analyst at Interarbor Solutions.
Listen to the podcast (45:26 minutes).
Here are some excerpts:
Jeremiah Grossman: An interesting paradigm shift is happening. When you look at website attacks, things haven't changed much. An application that exists in the enterprise is the same application that exists in the cloud. For us, when we are attacking websites and assessing their security, it doesn't really matter what infrastructure it's actually on. We break into it just as same as everything else.
Our job, in the website vulnerability management business, is to find those vulnerabilities ahead of time and help our customers fix those issues before they become larger problems. And if you look at any security report on the Web right now, as far as security goes, it's a Web security world.
What's different [with cloud] among our customer base is that they can't run to their comfort zone. They can't run to secure their enterprise with firewalls, intrusion detection systems, and encryption. They have to focus on the application. That's what's really different about cloud, when it comes to Web security. You have to focus on the apps, because you have nothing else to go on.
Andy Ellis: The first thing you have to do is to understand your own business. That's often the first mistake that security practitioners may make. They try to apply a common model of security thinking to very unique businesses. Even in one industry, everybody has a slightly different business model.
You have to understand what risks are acceptable to your business. Every business is in the practice of taking risk. That's how you make money. If you don't take any risk, you're not going to make money. So, understand that first. What are the risks that are acceptable to the business, and what are the ones that are unacceptable?
Security often lives in that gray area in between. How do we take risks that are neither fully acceptable nor fully unacceptable, and how do we manage them in a fashion to make them one or the other? If they're not acceptable, we don't take them, and if they are acceptable, we do. Hopefully we find a way to increase our revenue stream by taking those risks. ...
There's a huge gap in what people think is secure and what people are doing today in trusting in the security in the cloud. When we look at our customer base, over 90 of the top 100 retailers on the Internet are using our cloud-based solutions to accelerate their applications -- and what's more mission-critical than expecting money from your customers?
At Akamai, we see that where people are saying, "The cloud is not secure, we can't trust the cloud." At the same time, business decision makers are evaluating the risk and moving forward in the cloud.
A lot of that is working with their vendors to understand their security practices and comparing that to what they would do themselves. Sometimes, there are shifts. Cloud gives you different capabilities that you might be able to take advantage of, once you're out in the cloud.
Chris Hoff: I like to say that if your security stinks before you move to the cloud, you will be pleasantly unsurprised by change, because it's not going to get any better -- or probably not even necessarily any worse -- when you move to cloud computing.
What we're learning today is that if we secure our information and applications properly and the infrastructure is able to deal with the dynamism, you will, by default, start to see derivative impacts and benefits on security, because our models will change. At least, our thinking about security models will change.
We in the security industry in some way try to hold the cloud providers to a higher standard. I'm not sure that the consumer, who actually uses these services, sees much of a difference in terms of what they expect, other than it should be up, it should be available, and it should be just as secure as any other Internet-based service they use.
Those cloud providers -- cloud service and cloud computing providers -- are in the business of making sure that they can offer you really robust delivery. At this time, they focus there. We have a challenge to take everything we have done previously, in all these other different models, still do that, and deal with some of the implementation and operational elements that cloud computing, elasticity, dynamism, and all this fantastic set of capabilities bring.
So we get wrapped around the axle many times in discussions about cloud, where a lot of what we are talking about still needs to be taken care of from an infrastructure and application standpoint.
Ellis: That's the challenge for people who are moving out to the cloud. That area may be in the purview of the provider. While they may trust the provider, and the provider has done the best they can do in that arena, when they still see risks, they can no longer say, "I'll just put in a firewall. I'll just do this." Now, they have to tackle a really sticky wicket. Do you have a safe application wherever it lives?
That's where people run into a challenge: "It's cloud. Let me make the provider responsible." But at the end of day, the overall risk structure is still the responsibility of the business. Ultimately, the data owner, the business who is actually using whatever the compute cycles are.
Grossman: To piggyback on what Andy said, something has been lost. When you host an application internally, you can build it, you can deploy it, and you can test it. Now, all of a sudden, you've brought in a cloud provider, on somebody else's infrastructure, and you have to get permission to test it. It's not yours anymore.
Actually, one of the big things [to attend to] out there is a right to test. You have no right to test these infrastructure systems. If you do so without permission, it's illegal. So you have lost visibility. You've lost technical visibility and security of the application.
When the cloud provider changes the app, it changes the risk profile of the application, too, but you don't know when that happens and you don't know what the end result is. There's a disconnect between the consumer, the business, and the cloud computing provider or whatever the system is.
Hoff: Cloud computing has become a fantastic forcing function, because what its done to the business and to IT. We talked about paradigm shifts and how important this is in the overall advancement of computing.
The reality is that cloud causes people to say, "If the thing that's most important to me is information and protecting that information, and applications are conduits to it, and the infrastructure allows it to flow, then maybe what I ought to do is take a big picture view of this. I ought to focus on protecting my information, content, and data, which is now even more interestingly a mixture of traditional data, but also voice and video and mixed media applications, social networks, and mashups."
The complexity comes about because with collaboration, we have enabled all sorts of fantastic interconnectivity between what was previously disparate, little mini-islands, with mini-perimeters that we could secure relatively well.
The application security and the information security, tied in and tightly coupled with an awareness of the infrastructure that powers it, even though it's supposed to be abstracted in cloud computing, is really where people have a difficult time grasping the concepts between where we are today and what cloud computing offers them or doesn't, and what that means for the security models.
Ellis: There's a great initiative going on right now called "CloudAudit," which is aimed at helping people think through this security of a process and how you share controls between two disparate entities, so we can make those decisions at a higher level.
If I am trusting my cloud provider to provider some level of security, I should get some insight into what they're doing, so that I can make my decisions as a business unit. I can see changes there, the changes I am taking advantage of, and how that fits my entire software development life cycle.
It's still nascent. People are still changing their mindset to think through that whole architecture, but we're starting to see that more and more -- certainly within our customer base -- as people think, "I'm out in the cloud. How is that different? What can I take advantage of that's there that wasn't there in my enterprise? What are the things that aren't there that I am used to that now I have to shift and adapt to that change?"
Hoff: What's interesting about cloud computing as a derivative set of activities that you might have focused on from a governance perspective, with outsourcing, or any sort of thing where you have essentially given over control of the operation and administration of your assets and applications, is that you can outsource responsibility, but not necessarily accountability. That's something we need to remember.
Think about the notion of risk and risk management. I was on a panel the other day and somebody said, "You can't say risk management, because everyone says risk management." But, that's actually the answer. If I understand what's different and what is the same about cloud computing or the cloud computing implementation I am looking at, then I can make decisions on whether or not that information, that application, that data, ought to be put in the hands of somebody else.
In some cases, it can't be, for lots of real, valid reasons. There's no one-size-fits-all for cloud. Those issues force people to think about what is the same and what is different in cloud computing.
Previously, you introduced the discussion about the CSA. The thing we really worked on initially were 15 areas of concerns, and they're now consolidated to 13 areas of concern. What's different? What's the same? How do I need to focus on this? How can I map my compliance efforts? How can I assess, even if there are technical elements that are different in cloud computing? How can I assess the operational and cultural impacts?
Grossman: What I've seen in the last couple of years is that what drives security awareness is break-ins. Whether the bad guys are nation- or state-sponsored actors or whether they are organized criminals after credit card numbers, breaches happen. They're happening in record numbers, and they're stealing everything they can get their hands on.
Fortunately or unfortunately, from a cloud computing standpoint, all the attacks are largely the same, whether one application is here or in the cloud. You attack it directly, and all the methodologies to attack a website are the same. You have things like cross-site scripting, SQL injection, cross-site request forgery. They are all the same. That's one way to access the data that you are after.
The other way is to get on the other half of Web security. That's the browser. You infect a website, the user runs into it, and they get infected. You email them a link. They click something. You infect them that way. Once you get on to the host machine, the client side of the connection, then you can leverage those credentials and then get into the cloud, the back-end way, the right way, and no one sees you.
Breaches make headlines. Headlines make people nervous, whether it's businesses or consumers. When a business outsources things to the cloud or a SaaS provider, they still have this nervous reaction about security, because their customers have this nervous reaction about security. So they start asking about security. "What are you doing to protect my data?"
All of a sudden, if that cloud provider, that vendor, takes security seriously and can prove it, demonstrate it, and get the market to accept it, security becomes a differentiating factor. It becomes an enabler of the top line, rather than a cost on the bottom line.
Ellis: I like to look at security as being a business-enabler in three areas. The obvious one, we all think, is risk reduction. How can I reduce my risk with cloud-based security services? Are there ways which I can get out there and do things safer? I'm not necessarily going to change anything else about my business. That's great and that's our normal model.
Security can also be a revenue-enabler and it can also be a protection of revenue. Web application firewalls is a great example of fraud mitigation services. There are a lot of services available through the cloud that can be used to protect your brand and your revenue against loss, but also help you grow revenue. As you just said, it's all about trust. People go back to brands that they trust, and security can be a key component of that.
It doesn't always have to be visible to the end user, but as you noted with the car industry, people build the perception around incidents. If you can be incident-free compared to your competition, that's a huge differentiator, as you go down into more and deeper activities that require deep trust with your end users.
A lot of what we try to do is build a wrapper in a sandbox around each customer to give them the same, consistent level of security. A big challenge in the enterprise model is that for every application that you stand up, you have to build that security stack from the ground up.
One advantage cloud does give you is that, if you are working with somebody who has thought about this is, you can take advantages of practices that they have already instituted. So, you get some level of commonality. Then, if a customer sees something and says, "You should improve this," that improvement can affect an entire customer base. Cloud has a benefit there to match some of the weaknesses it may have elsewhere.
Historically, in the enterprise model, we think about data in terms of being tied to a given application. That's not really accurate. The data still moves around inside an enterprise. As Jeremiah noted, the weak point is often the browser. Compromise the client, and you get access to the data. As people move to cloud, they start to change their risk thinking. Now, they think about the data and everywhere it lives and that gives them an opportunity to change their own risk model and think about how they're protecting the data and not just a specific application it used to live in.
As we noted earlier, a large fraction of the Internet retailers are using cloud for their most mission-critical things, their financial data, coming through every time somebody buys something.
If you are willing to trust that level of data to the cloud, you are making some knee-jerk reaction about an internal web conference between 12 people and a presentation about something that frankly most people aren't going to care about, and you are saying, "That's too sensitive to be in the cloud." But your revenue stream could be in the cloud. Sometimes it shows that we think parochially about security in some places.
Grossman: What's interesting about security spending versus infrastructure spending or just general IT spending is that it seems security is diametrically opposed to the business. We spend the most money on applications and our data, but the least amount of security risk spend. We spend the least on infrastructure relative to applications, but that's where we spend the most of our security dollars. So you seem to be diametrically opposed.
What cloud computing does, and the reason for this talk, is that it flattens the world. It abstracts the cloud below and forces us to realign with the business. That's what cloud will bring in a good way. It's just that you have to do it commensurate with the business.
Dana Gardner is president and principal analyst at Interarbor Solutions, which tracks trends, delivers forecasts and interprets the competitive landscape of enterprise applications and software infrastructure markets for clients. He also produces BriefingsDirect sponsored podcasts. Follow Dana Gardner on Twitter. Disclosure: Akamai Technologies sponsored this podcast.