Archive

Author Archive

A decade inside Microsoft Security

November 9th, 2017 No comments

Ten years ago, I walked onto Microsofts Redmond campus to take a role on a team that partnered with governments and CERTs on cybersecurity. Id just left a meaningful career in US federal government service because I thought it would be fascinating to experience first-hand the security challenges and innovation from the perspective of the IT industry, especially within Microsoft, given its presence around the US federal government. I fully expected to spend a year or two in Microsoft and then resume my federal career with useful IT industry perspectives on security. Two days after I started, Popular Sciences annual Ten worst jobs in science survey came out, and I was surprised to see Microsoft Security Grunt in sixth place. Though the article was tongue-in-cheek, saluting those who take on tough challenges, the fact that we made this ignominious list certainly made me wonder if Id made a huge mistake.

I spent much of my first few years hearing from government and enterprise executives that Microsoft was part of the security problem. Working with so many hard-working engineers, researchers, security architects, threat hunters, and developers trying to tackle these increasingly complex challenges, I disagreed. But, we all recognized that we needed to do more to defend the ecosystem, and to better articulate our efforts. Wed been investing in security well before 2007, notably with the Trustworthy Computing Initiative and Security Development Lifecycle, and we continue to invest heavily in technologies and people – we now employ over 3,500 people in security across the company. I rarely hear anymore that we are perceived as a security liability, but our work isnt done. Ten years later, Im still here, busier than ever, delaying my long-expected return to federal service, helping enterprise CISOs secure their environments, their users, and their data.

Complexity vs. security

Is it possible, however, that our industrys investments in security have created another problem – that of complexity? Have we innovated our way into a more challenging situation? My fellow security advisors at Microsoft have shared customer frustrations over the growing security vendor presence in their environments. While these different technologies may solve specific requirements, in doing so, they create a management headache. Twice this week in Redmond, CISOs from large manufacturers challenged me to help them better understand security capabilities they already owned from Microsoft, but werent aware of. They sought to use this discovery process to identify opportunities to rationalize their security vendor presence. As one CISO said, Just help me simplify all of this.

There is a large ecosystem of very capable and innovative professionals delivering solutions into a vibrant and crowded security marketplace. With all of this IP, how can we best help CISOs use important innovation while reducing complexity in their environments? And, can we help them maximize value from their investments without sacrificing security and performance?

Best-of-suite capabilities

Large enterprises may employ up to 100 vendors technologies to handle different security functions. Different vendors may handle identity and access management, data loss prevention, key management, service management, cloud application security, and so on. Many companies are now turning to machine learning and user behavior technologies. Many claim best of breed or best in class, capabilities and there is impressive innovation in the marketplace. Recognizing this, we have made acquisition a part of Microsofts security strategy – since 2013 weve acquired companies like Aorato, Secure Islands, Adallom, and most recently Hexadite.

Microsofts experience as a large global enterprise is similar to our enterprise customers. Weve been working to rationalize the 100+ different security providers in our infrastructure to help us better manage our external dependencies and more efficiently manage budgets. Weve been moving toward a default policy of Microsoft first security technology where possible in our environment. Doing so helps us standardize on newer and familiar technologies that complement each other.

That said, whether we build or buy, our focus is to deliver an overall best in suite approach to help customers deploy, maintain, monitor, and protect our enterprise products and services as securely as possible. We are investing heavily in the Intelligent Security Graph. It leverages our vast security intelligence, connects and correlates information, and uses advanced analytics to help detect and respond to threats faster. If you are already working with Microsoft to advance your productivity and collaboration needs by deploying Windows 10, Office 365, Azure, or other core enterprise services, you should make better use of these investments and reduce dependency on third-party solutions by taking advantage of built-in monitoring and detection capabilities in these solutions. A best-of-suite approach also lowers the costs and complexity of administering a security program, e.g. making vendor assessments and procurement easier, reducing training and learning curves, and standardizing on common dashboards.

Reducing complexity also requires that we make our security technologies easy to acquire and use. Here are some interesting examples of how our various offerings connect to each other and have built-in capabilities:

  • The Windows Defender Advanced Threat Protection(ATP) offer seamlessly integrates with O365 ATP to provide more visibility into adversary activity against devices and mailboxes, and to give your security teams more control over these resources. Watch this great video to learn more about the services integration. Windows Defender ATP monitors behaviors on a device and sends alerts on suspicious activities. The console provides your security team with the ability to perform one-click actions such as isolating a machine, collecting a forensics package, and stopping and quarantining files. You can then track the kill chain into your O365 environment if a suspicious file on the device arrived via email. Once in O365 ATP, you can quarantine the email, detonate a potentially malicious payload, block the traffic from your environment, and identify other users who may have been targeted.
  • Azure Information Protection provides built-in capabilities to classify and label data, apply rights-management protections (that follows the data object) and gives data owners and admins visibility into, and control over, where that data goes and whether recipients attempt to violate policy.

Thousands of companies around the world are innovating, competing, and partnering to defeat adversaries and to secure the computing ecosystem. No single company can do it all. But by making it as convenient as possible for you to acquire and deploy technologies that integrate, communicate and complement each other, we believe we can offer a best-of-suite benefit to help secure users, devices, apps, data, and infrastructure. Visit https://www.microsoft.com/secure to learn about our solutions and reach out to your local Microsoft representative to learn more about compelling security technologies that you may already own. For additional information, and to stay on top of our investments in security, bookmark this Microsoft Secure blog.


Mark McIntyre, CISSP, is an Executive Security Advisor (ESA) in the Microsoft Enterprise and Cybersecurity Group. Mark works with global public sector and commercial enterprises, helping them transform their businesses while protecting data and assets by moving securely to the Cloud. As an ESA, Mark supports CISOs and their teams with cybersecurity reviews and planning. He also helps them understand Microsofts perspectives on the evolving cyber threat landscape and how Microsoft defends its enterprise, employees and users around the world.

Categories: Uncategorized Tags:

SSN for authentication is all wrong

October 23rd, 2017 No comments

Unless you were stranded on a deserted island or participating in a zen digital fast chances are youve heard plenty about the massive Equifax breach and the head-rolling fallout. In the flurry of headlines and advice about credit freezes an important part of the conversation was lost: if we didnt misuse our social security numbers, losing them wouldnt be a big deal. Let me explain: most people, and that mainly includes some pretty high-up identity experts that Ive met in my travels, dont understand the difference between identification and verification. In the real world, conflating those two points doesnt often have dire consequences. In the digital world, its a huge mistake that can lead to severe impacts.

Isnt it all just authentication you may ask? Well, yes, identification and verification are both parts of the authentication whole, but failure to understand the differences is where the mess comes in. However, one reason its so hard for many of us to separate identification and verification is that historically we havent had to. Think back to how humans authenticated to each other before the ability to travel long distances came into the picture. Our circle of acquaintances was pretty small and we knew each other by sight and sound. Just by looking at your neighbor, Bob, you could authenticate him. If you met a stranger, chances are someone else in the village knew the stranger and could vouch for her.

The ability to travel long distances changed the equation a bit. We developed documents that provided verification during the initiation phase, for example when you have to bring a birth certificate to the DMV to get your initial drivers license. And ongoing identification like a unique ID and a photo. These documents served as a single identification and verification mechanism. And that was great! Worked fine for years, until the digital age.

The digital age changed the model because rather than one person holding a single license with their photo on it, we had billions of people trying to authenticate to billions of systems with simple credentials like user name and password. And no friendly local villager to vouch for us.

Who are you? Prove it!

This is where the difference between the two really starts to matter. Identification answers the question: Who are you? Your name is an identifier. It could also be an alias, such as your unique employee ID number.

Do you want your name to be private? Imagine meeting another parent at your kids soccer game and refusing to tell them your name for security reasons. How about: Oh your new puppy is so adorable, whats her name? And you respond, If I told you, Id have to kill you. Or you try to find an address in a town with no street signs because the town is super security conscious. Ridiculous, right? Identifiers are public specifically so we can share them to help identify things.

We also want consistency in our identifiers. Imagine if that town had street signs, but changed the names of the streets every 24 hours for security reasons. And uniqueness, if every street had the same name, youd still have a heck of a time finding the right address wouldnt you?

Now that were clear on what the identifier is, we can enumerate a few aspects that make up a really good one:

  • Public
  • Unchanging
  • Unique

In a town or public road, we have a level of trust that the street sign is correct because the local authorities have governance over road signs. Back in our village, we trust Bob is Bob because we can verify him ourselves. But in the digital world, things get pretty tricky how do you verify someone or something youve never met before? Ask them to- Prove It!

We use these two aspects of authentication almost daily when we log into systems with a user ID (identification) and password (verification). How we verify in the real world can be public, unchanging, and unique because its very hard to forge a whole person. Or to switch all the street signs in a town. But verification online is trickier. We need to be able toprovide verification of who we are to a number of entities, many of whom arent great at protecting data. And if the same verification is re-used across entities, and one loses it, attackers could gain access to every site where it was used. This is why experts strongly recommend using unique passwords for every website/app. This goes for those challenge questions too. Which can lead to some fun calls with customer service, Oh, the town where I was born? Its: xja*21njaJK)`jjAQ^. At this point in time our fathers middle name, first pets name, town where we were born, school we went to and address history should be assumed public, using them as secrets for verification doesnt make sense anymore.

If one site loses your digital verification info, no worries. You only used it for that site and can create new info for the next one. What if you couldnt change your password ever? It was permanent and also got lost during the Yahoo! breach? And it was the one you use at your bank, and for your college and car loans, and your health insurance? How would you feel?

So, with that in mind, youd probably agree that the best digital verifiers are:

  • Private
  • Easily changed
  • Unique

Your turn

OK, now that you know the difference between identification and verification and the challenges of verification in a digital world, what do you think – Is your SSN a better identifier or verifier?

Categories: cybersecurity, Data Privacy, Tips & Talk Tags: