Research

This week’s question in our Ask Securosis series moves past a technology question into the realm of management and statistical research. Scott asks: … It seems that the companies Jason identified in his study have become the most productive in their industry sectors by streamlining, removing wasteful process, empowering staff, etc. (scary prospects for security professionals). Obviously, there is a cross-over point, of sorts, where this starts to impact information security in a way that puts them at risk for a big set-back that was just hadn’t happened yet. I’d really like to find any references to studies regarding how “good security” positively or negatively impacts long term productivity. We security geeks have a gut feel that it has a positive effect, if done right. But we need data to prove it. In all your market research experience have you come across any such studies, and do you feel they provide solid evidence or arguments for the case of Security vs. Productivity? If you don’t know of any, then chances are that there aren’t any well documented ones. Since I don’t have the resources to do this kind of study myself, I’m thinking of approaching a university business school to see if they can do a follow-up study on the companies Jason found, and look at their records on security. Or I guess I could try to ask the author himself. I think this is a key issue for managers struggling to understand the trade-offs in security: how much productivity will they be foregoing if they commit to a real security initiative. I’d like to explore this idea more to help them understand the impacts. No, I’m not aware of any study linking security with productivity metrics. Or even correlating highly productive companies with their security postures. Since I can’t point you in the right direction to get the answer you’re looking for, I’ll focus on providing a few aspects to look at if you do decide to link up with a university and perform a proper study. My gut feeling is there is an empirical problem in attempting a study like this. While we can accurately measure the productivity impact of certain security controls, correlating that to the additional risk exposure would, by nature, involve introducing risk metrics that are neither as precise nor as accurate as those measuring productivity. Risk measurements in infosec involve the use of estimates that don’t accurately reflect the full financial exposure of insufficient controls. We can never fully measure losses or potential losses, thus the numbers will be oranges to the apples of productivity measurements. The result ensures it’s nearly impossible to use these measurements to balance security vs. productivity, and depending on how the numbers skew we will draw the wrong conclusions. For example, they may show that passwords hurt productivity by X dollars, and security risk drops by an estimated Y dollars, with said estimate being nearly impossible to calculate accurately. We might end up thinking that because we’ve never had a system compromised due to a weak password, we don’t need them at all. Okay, an extreme example, so here are a few ideas on how I’d look at the research. Ideally I would try and find two organizations with equally good productivity, but variable security. If we can normalize enough of the variables, and find a big enough sample set, that gives us a good macro view on any causality. We might also look at a very productive company vs. a very secure company that isn’t productive. Good luck finding that. But I think what you really want to do is devise a model to determine the productivity impact of potential security controls, not just security in general. You should be able to measure that for any specific security control as long as you have a corresponding measurement of productivity. You should then map in estimates of risk measurements to make a decision. Otherwise, nearly everything will reduce productivity, but the corresponding risk might exceed acceptable tolerance. Also, this should take into account any alternative controls that achieve the same goal, with a lower productivity impact. And that control impact varies over time. At this point we’ve just created enough complexity that measuring the performance impact of a security control is now greater than the performance impact of said control. My advice? We spend more time identifying the most efficient ways to be secure with the least performance impact. Share:

Websense posted a replay link for my webcast on DLP. Here ya go… Share:

Last week I gave a webinar on database security for ZDNet, sponsored by Oracle. We had an exceptionally good turnout and ran a couple of polls during the session. Oracle just posted the results on a new security blog they’ve set up. One of the questions was on data masking, something we’ve discussed here before. I asked the audience how many actively performed data masking within their organizations. We got a great response, with a sample size of 139. Not huge, but still somewhat statistically significant. Most organizations don’t data mask, and of those that do, only a combined 13% have a formalized program. No surprises, but it’s nice to see it in some real numbers. And don’t forget data masking law number 5. Here’s the obligatory pretty picture, and you can still replay the session over at ZDNet. < p style=”text-align:right;font-size:10px;”>Technorati Tags: Data, Data Masking, Database Security, Oracle, Survey Share:

I got a few emails from people asking to push SunSec up to next week due to upcoming travel, conferences, and training. Plan is for Wednesday night, the 6th, back at Furio. Email me at rmogull@securosis.com if that won’t work for you, and please spread the word… Share:

While I sometimes get annoyed with various security technologies, there are very few I consider to be complete snake oil. However, those remote “data destruction” tools or “Lojack for laptops” are complete crap when it comes to security. Absolute bullshit, and I don’t use language like that here very often. They might have some value in recovering the physical asset, but as this case shows they sure as hell won’t protect you from a data breach: Horizon Wednesday said it has notified about 300,000 of its members of the potential compromise of their personal information following the theft of a laptop containing the data on Jan 5. A security feature on the stolen laptop automatically deleted all of the confidential information on Jan. 23, a company spokesman said. But it is not clear whether the thief who stole the computer accessed the data on the system before then, he said. The data on the laptop was unencrypted but password-protected. I guarantee you’ll see some of these companies at the next security conference you go to. If you want to use them to help with physical recovery, that’s fine. But for data security? No fracking way. < p style=”text-align:right;font-size:10px;”>Technorati Tags: Data Breach, Laptop Encryption Share:

Dark Reading just posted my column for this month, entitled, “11 Truths We Hate To Admit”. Due to a miscommunication with my editor it reads as if I still live in Boulder, Colorado. I’m really down in Phoenix, but spent most of my adult life in Boulder. DR is a fun publication to write for- they want us to poke the industry with a stick and get people thinking. Nothing I wrote is any big surprise, but they aren’t the kinds of things we tend to publish. If it doesn’t piss at least a few people off I didn’t do a good job writing the article, although so far the reviews are mostly positive. Damn. Here’s a snippet: 1. Signature based desktop antivirus is an addiction, not effective security. AV is often the single biggest security expense in an organization, yet it’s one of the least effective. Gateway AV is still a reasonable investment to filter out known garbage, but desktop AV needs to seriously improve its heuristics and other non-signature techniques if it is to protect us. Independent reports indicate current AV products are full of gaping holes, and many organizations experience extensive downtime from bad signatures and poor performance. At least today’s malware doesn’t grind your computer to a halt at noon every Wednesday. 2. The bad guys beat us because they’re agnostic and we’re religious. The bad guys are always innovating for competitive advantage, but innovation isn’t something large organizations or industries do well. We get wrapped up in our own little religious battles over PKI, IDS, standards, AV, whoever we work for at the time, and what’s worked for us before. We become too personally tied to pet projects we’re experienced with — and can’t let go of. 3. Antitrust concerns force Microsoft to weaken security. Host security companies take out full-page ads in the Wall Street Journal and threaten to go to court when Microsoft adds security features that might tread on their turf. Thanks to some poor past behavior by Microsoft, these tactics work. But if it weren’t for the antitrust problems of the past, we’d all have free anti-spyware and AV in Windows, forcing those other companies to compete on merit. < p style=”text-align:right;font-size:10px;”>Technorati Tags: Dark Reading, Security Industry Share:

While I was traveling home, Martin posted the latest episode of the Network Security Podcast. Our guests this week are Marcin and Andre from http://www.tssci-security.com/. We spend most of the episode talking about web application security issues. At least, I think we do. Due to technical difficulties related to my travel I got booted half-way through the podcast just as Andre was getting into the good stuff. Sorry for the problems guys, and thanks again for being on the show. And for getting me drunk and talking me into inviting you on the show. And having me pay the tab. Um. I think I’m doing this wrong… < p style=”text-align:right;font-size:10px;”>Technorati Tags: Application security, Podcast Share:

Once upon a time, an evil virus struck the land. But the people were prepared, and they stopped the virus before too many became sick… or so they thought. The virus really learned to hide, finding a home among wayward travelers outside the gates of the city. Weeks later these travelers returned home and unknowingly infected the cities. And weeks after that the next wave of travelers came to the cities, and more were infected. And then some scientists said, “Enough! No more will we let our cities become infected by these travelers. Now is the time to protect ourselves from the threats within!” The scientists created a new defense, called NAC, which would check the health of anyone before entering the city, and all was good. But NAC was new, and the first versions didn’t work as well as everyone would have liked. Then, two famous alchemists decided that they should control NAC. Rather than providing it to the people to use, they decided to tell everyone they would provide it. Eventually. And maybe it wouldn’t work quite as expected, but it would be good because it would be big. And then other alchemists decided that the people wanted NAC, but didn’t know what NAC was, so they removed the old labels from their elixirs and put on new NAC labels. And the people were confused. And they waited. (Apologies for starting so many sentences with ‘and’, but you’ll get over it.) I was listening to Alan and Mitchell’s StillSecure podcast the other day and, as usual, the subject of NAC came up. For those of you who don’t know, NAC stands for Network Admission Control or Network Access Control, depending on who you talk with. The technology was originally developed to provide pre-connect health checks when guests or mobile employees plugged into the office network. Alan was ranting on the dilution of the term, and as much as it pains me I have to agree with him. When the SQL Slammer virus hit, most companies were well defended by blocking the port on their firewalls. Those companies then found themselves infected over the following weeks in waves, as mobile employees and contractors started coming back and plugging into the wall, behind the firewall. The concept of NAC was to prevent internal infections from systems physically connecting behind perimeter defenses. A computer would plug in and would then be scanned, or checked using an agent, before it was given an IP address or other network access. If the system wasn’t up to snuff, it could be quarantined off on a network segment outside the firewall (perhaps to download the missing security software), or simply denied access. It’s a great idea, but like all great ideas a combination of big fish and bottom feeders wanted in. “NAC” kept getting expanded and integrated with everything from 802.1x for port-based authentication (only letting a computer get a usable IP address after a user is approved- a pretty good idea) to all sorts of real-time monitoring, quarantining, VLAN weirdness, and kitchen sinks. It’s a market that Cisco and Microsoft decided they want to control, and early on they started making waves without providing much in terms of functional product. It was a way for Cisco to get their endpoint agents onto desktops and to push clients to upgrade their networking hardware, since parts of their NAC don’t work if they aren’t built into the switch. I like NAC, and if I had more than 6 computers on my network it’s the kind of thing I’d look at more closely. But I’d keep myself focused on the basics- protecting my network from malicious guest and mobile systems. I’d want a mix of agent and agentless (for managed and unmanged systems) and keep focused on pre- and post- connection health checks. I wouldn’t wait for the big vendors, knowing that in the long term they’ll own it all anyway, even if they have to buy it. Yes, Cisco has stuff now, but I hear it’s pretty complex to deploy. NAC, like much of network security, will eventually be built into the network fabric. At best, we’ll have a separate security control plane for separation of duties. This is a hell of a long way out and not something that should affect your buying decisions today. I’ll be the first to admit I have a lot more depth in data and application security than netsec, but I’ve watched for years as a great idea (NAC) has been pummeled by the market. I even did an interview on it over at SearchSecurity. It reminds me a lot of Data Loss Prevention/Content Monitoring and Protection (DLP/CMP). A good technology that provides immediate value, which quickly becomes far more confusing than needed as all sorts of people want in on the action. If you want to protect yourself from potentially malicious systems plugging into your network (including remote access) take a look at NAC. If you want all the other bells and whistles you see running around out there you can look at them too, just don’t call them NAC. < p style=”text-align:right;font-size:10px;”>Technorati Tags: NAC, Network Security Share:

If any of you are involved with the Open Group, I’ll be giving a presentation at the Forum tomorrow in San Francisco. The topic ias: The Future of Security: How Disruption and Innovation are Putting the “Information” Back In Information Security Unlike other areas of information technology, security is always battling with external forces. From external attackers to our own business initiatives, disruptions are our way of life. We are in the midst of a major shift as “information-centric” security evolves into the dominant security model, while we still need to maintain our investments of yesterday. The session will discuss these disruptive forces and assess how and where some of the established security technologies (firewalls and e-mail security), and emerging security technologies (data leak prevention, endpoint security, network access control, enterprise rights management, and on and on) will fit into our information-centric security architectures in the future. At least that’s what I think I’m talking about- the description on the conference site doesn’t seem to match. Oh well, I’ll figure it ou when I get there. Note that this is my half of an upcoming joint presentation with Chris Hoff. More details on that later… Share:

Tomorrow I’ll be giving a webcast over at ZDNet (sponsored by Oracle) on the Top 5 Database Security Resolutions for 2008. The resolutions have changed a bit since I first posted about them over here, and I decided to swap in data masking for the last one. I almost pulled it back out after I found out my sponsor (Oracle) just released a data masking product (I try to avoid being too promotional in my webinars), but it’s something I’ve been talking about for a while and it’s too important to pull just because a few people might think I was being biased. We’re up to nearly 600 people registered for the event, making it one of the largest webcasts I’ve done. But enough self-promotion; it’s time to talk about data masking. Data masking started popping up as an issue about 3 years ago. At the time I was covering database security, but client calls were bouncing around between me on the security team and someone over in application development. It’s one of these annoying security issues that crosses organizational boundaries and ends up the responsibility of those will little security experience. It’s an issue that grew organically- first popping up in some audits related to GLBA (a financial services regulation), and now something we see required for PCI and a few other regulations. Data masking is really a bad term for what we’re talking about. We can technically mask data anywhere, but when we use the term data masking we usually mean “test data generation” or “analytical data generation”. It’s the conversion of production data into either test and development data or data for a data warehouse (OLAP). For this post we’ll focus on test data generation, but the same techniques can be used for an OLAP where you want data that represents production data, but still protects the sensitive stuff. And that’s our goal- to take sensitive data from a production system and convert it into non-sensitive data suitable for testing or analysis. We can do this through substitution, transposition, obfuscation, de-coupling, scrambling, hashing, or even encryption. I’m going to quickly eliminate hashing and encryption from the discussion- those techniques are very effective at protecting data, but the result breaks the second rule of data masking- that the data is still representative of the source, without being sensitive. Organizations are increasingly finding that data masking is mandated for regulatory compliance. It’s also an extremely effective way to reduce enterprise risk. Development and test environments are rarely as secure as production, and there’s little reason developers should have access to sensitive data. Analytical systems are often accessed by a wide variety of users, most of whom shouldn’t see sensitive data, with only a fraction of the access and other security controls in transactional systems. With that, and since I get way more hits if I have the “x laws” in the title, here are the Five Laws of Data Masking: Masking must not be reversible. However you mask your data, it should never be possible to use it to retrieve the original sensitive data. The results must be representative of the source data. The reason to mask data instead of just generating random data is that masking allows you to protect sensitive information that still resembles production data for development and testing purposes. This could include geographic distributions, credit card distributions (e.g., leaving the first 4 numbers unchanged, but scrambling the rest), or maintaining human readability of (fake) names and addresses. Referential integrity must be maintained. Your masking solution should maintain referential integrity- if a credit card number is a primary key, and scrambled as part of masking, then all instances of that number linked through key pairs must be scrambled identically. Only mask non-sensitive data if it can be used to recreate sensitive data. It isn’t necessary to mask everything in your database, just those parts that you deem sensitive. But remember, some non-sensitive data can be used to either recreate or tie back to sensitive data. For example, if you scramble a medical ID but the treatment codes for a record could only map back to the original record, you also need to scramble those codes. This is called inference analysis, and your masking should protect against it. Masking must be a repeatable process. One-off masking is not only nearly impossible to maintain, but it’s fairly ineffective. Development/test data needs to represent constantly changing production data as closely as possible. Analytical data may need to be generated daily, or even hourly. If masking isn’t an automated process it’s inefficient, expensive, and ineffective. I know of some organizations that centralize masking and offer it as an internal service to the enterprise. These “laws” are just to start the discussion on masking. In future posts I’ll discuss my recommended data masking process and what features to look for in tools. And if you absolutely can’t wait until I get around to a follow-on post, join me for the webinar on Friday where I’ll dig in a little deeper. Share: