I was interviewed a few weeks back by Chris Maher on Linkedin. The topic was concerning “Trusted Computing” and ramblings on security.
CM: Howard, as you know, I quoted you at the 2011 NSA Trusted Computing Conference & Exposition: “Well.. I believe in Americans. I believe that when we see various challenges that we individually step up and out to deal with them. We have put your faith and trust in leadership and leadership has been pounded with more work than they can handle (yes, I am being nice). That being said, it is up to us individually to lead where we are. We must individually work to change our own behavior and look to influence others by leading from where we are. If I am a Janitor, then I look for ways to be efficient in cleaning and thrifty in spending for supplies, or find ways to reuse supplies. If you are an Executive Assistant, find ways to make a difference in the office. If you are a Technical Strategist, teach everyone everything you know about Service Orientation and Trusted Computing and technical reuse models. It doesn’t matter who you are, it matters what you do. Our jobs do not define us holistically. In recent days I have seen civilian leaders (you know who you are) step up to the plate and take risks in order to share their ideas on how to create a more effective and efficient acquisition solutions. It isn’t only up to them. We will find more success together by working to change these behaviors and tackling the challenges we can see one person and one problem at a time…” (SOURCE):https://cohenovate.wordpress.com/category/howard-cohen/
It’s a great quote for a variety of reasons. That said, I want to focus on your awareness of and experiences with Trusted Computing. How you were first introduced to Trusted Computing?
Chris, Thank you very for clearly understanding and articulating the message of “leading from where we are.” I have been working for the Department of Defense for close to a decade now, before that I worked at a school division and the commercial industry. I have worked for Joint Forces Command, Joint Staff and now DISA. I started hearing about Trusted Computing while working at the school division, if anyone is going to break your system it will be the kids. I learned a great deal about system hardening as I entered the world of military architectures at J8. I started at US Joint Forces Command by using security technical implementation guides (STIGs) as we call them. Prior to that I was using non-military oriented technologies like hard drive sheriff, deep freeze, bootable cd os (barts PE), stuff like that.
And, in your estimation, why does Trusted Computing matter? Why is it important?
In enterprise computing you want to be able to leverage standards. We need the ability to look at metrics and we need to understand what “expected behavior” is. In other words, we need to be able to know when something is not working right. So you need standards so that experts can be on the same page and understand what they are looking for as “normal” as opposed to seeing something that “interesting” , if everyone is doing their own thing at the enterprise it makes it very complicated to know what the heck is going on. You have “shadow IT” that will compromise the integrity of the network simply because it exists. When working in an enterprise users and operators need to trust that mechanisms are in place to protect them. I can go on about this but the bottom line is that to know if something is wrong you need to establish that something is right. I believe that is why Trusted Computing is important.
CM: As you may know, Richard Stallman once rebranded Trusted Computing (TC) as “Treacherous Computing” which made a neutral set of technologies out to be a threat to open computing and/or our civil liberties. Stallman conflated Microsoft’s Palladium effort with the word of then TCPA. Ever since, TC has been dogged by the adjective “controversial.” For me, TC (including self-encrypting drives) actually protects my civil liberties by arming me, the digital citizen, with technologies that can defend my information from any intruder… including an intrusive government. But that’s just my opinion. How do you assess the intersection of Trusted Computing and civil liberties.
As long as there are people involved in computing, there are going to be hackers. As long as we are at war with others, there will be people who will look to harm us in the real world or through technology. Sure you are sharing the standards but I would say process and method are two different things. In other words, you may have common technological frameworks and standards but how enterprise strategists think about and employ these technologies are different. For example, I know of an organization that uses two layers of username and password and additionally requires a common access card, all of which are standardized. The practice is abnormal but if a technologist was brought in to help solve a problem once he or she understood the architecture and because they are using standardized technologies and platforms they can help solve the problem. I equate it to having a human in the loop. People are your greatest protection mechanism as well as your greatest threat. In terms of civil liberties, I think we have some problems with the law more than technology. We don’t have a right to privacy, it isn’t guaranteed by the constitution and that means corporations and people are free to snoop around our business. When that gets into information gathering and data aggregation it poses a much bigger problem than just technical mechanisms to protect our data. It is more about what information did your city just put out about you and your home value, stuff like that. So, in other words I am not sure that Trusted Computing makes a difference here unless we are just talking about me protecting my local hard drive.
CM: Much noise is made by IT professionals about the difficulties of using TC, specifically going into the BIOS and having to turn on TPMs. And it must be said that there has not been the development of many applications that leverage TPMs. In your experience, is Trusted Computing too hard to implement?
I have seen full disk encryption at the corporate level and while working with the government. I have not seen BIOS based modules employed and I don’t have personal experience with BIOS based secure computing. As I mentioned earlier, while working at the school division we used a device call hdd sheriff and some technology out of Israel to perform persistent drive management and encryption. This was over 10 years ago too but the concepts have been around for a long time. There aren’t a lot of commercial options that I have seen at the application level that use TPM’s but I think there is value there depending upon the requirement. This is all about balance. Risk is the key. How much is this going to cost you? What are the implications? If I am working in the financial sector, I want as much technology as I can to protect my information. The same could be said for the medical industry, I haven’t figured that one out yet but I am sure there is a good reason.
CM: It’s been my contention that government MUST take the lead in adopting and recommending Trusted Computing. In this regard, I’ve been heartened by the NSA’s (more or less) full-throated endorsement of TC and by the CESG’s recommendation in favor its use. Further, as you may know, NIST 800-155 (in draft form) has recommended (or will recommend) the use of a hardware root of trust as a foundation for BIOS Integrity metrics. Still, it seems like .gov and .mil domains have been quite slow to fully adopt these open standards and technologies. In your view, what’s the state of play re: TC adoption within our government?
This is about cost of implementation and ability to implement. In other words, as long as there are programs that are “Programs of Record” with Title 10 authority, essentially meaning that they can control their own technical destiny there won’t be adoption unless it becomes part of the culture. For example, while working for Joint Forces Command I stood up one of if not the first accredited virtual infrastructure. Most people were getting rejected at the time because hardening didn’t exist aside from the vendor best practices. Information Assurance folks were afraid to take the risk, although it could mean millions in savings. It boiled down to courage and tenacity. The government leadership I worked with and for championed the idea and helped me bring people together by supporting our teams ideas. It took many briefs and I think I have stock in some chocolate company now as well to get people to believe that there was value in virtualizing the infrastructure. I know that sounds funny now because so many have adopted virtual technologies. Here is the kicker though, today even though virtualization has proven to be of great value there are many government programs that haven’t virtualized and / or won’t go because of requirements and title 10 authorities. CM: A great deal of academic and industry research has focused on the value of TC when it comes to authenticating users in a cloud-computing context…as well as using TC to protect user’s data in the cloud from the “insider threat.” Speaking specifically about the cloud-computing context, how important do you think TC technologies (TPMs) and protocols are as enablers?
As I started working on enterprise computing concepts and strategies, I started to see a trend. Thomas Erl talks about this in his service oriented architecture books but it has to do with understanding dependency. Cloud computing may increase risk. Notice I say “may” instead of will, the reason is that every enterprise situation and IT ecosystem is different, remember earlier when I was referring to process and method being two different things. Regardless of the situation organizations will have dependencies, for example you need communication services to connect to the Internet. As you increase services and connectivity requirements it is likely that you introduce more dependency.
The cloud really refers to “off premise” services. These services are interconnected enterprise services that go beyond an organizations local physical infrastructure. This is very important to realize because it means that hardware and IT resources are still potentially under trusted controls of an organization which of course then leads to leveraging organizational standards etc.
The difference is that when you have a dependency on a “cloud provider” that is outside of your organization you build dependencies in which you may lose control over the IT resources. As you give up autonomy or operational governance, you become more reliant on legal remedies. In other words, SLA’s or Service Level Agreements become critical to the organization. This relates to Trusted Computing in a lot of ways, for example a service provider may need to employ certain (TPM’s) prior to an agreement of use. This increases the cost to service providers and also may limit choices as to what service providers’ organizations can use. An example is that Amazon offers Federal services with enhanced security. I am not advocating for any service provider, I am simply saying that as cloud services increase, the costs of these services will increase and the demands of security and stability increase. In the grand scheme of things it wasn’t that long ago that most folks were on dial-up, it was $9.95 to $19.95, today most people pay $40.00 for Internet services not including the extra services they pay for while on the Internet. As these costs increase, it pushes the price of everything up, simple economics. Trusted Computing in the cloud is costly, but organizations when moving to the cloud will need to absorb these costs.
My key point is that we can’t rely on technology alone. Technology as it is today can be overcome by the human brain. That being said, we still must put barriers in place to slow down attackers enough so that we can identify in some manner that our information is being attacked. It is the difference between having a lock on the door and adding a security system. Some people would say that adding a security system adds no value or is a waste of time. I think as we continue to build technological solutions to thwart attackers or secure the enterprise, we strongly need to consider how we can keep “a human in the loop” and have people involved in watching the various stores. As we move forward with these kinds of discussions we truly need to consider people, process, methods and finally tools which in my mind is where a lot of the Trusted Computing area currently addresses.