Thursday, June 29, 2017

Security Data Scientists Without Borders – Thoughts from our first Colloquium

The move to the cloud is changing the security landscape. As a result, there is a surging interest in applying data-driven methods to security. In fact, there is a growing community of talented people focused on security data science. We’ve been shedding our respective “badges” and meeting informally for years, but recently decided to see how much progress we might make against some of our bigger challenges with a more structured and formal exchange of ideas in Redmond. The results far exceeded our expectations. Here’s a bit of what we learned.

The first thing to understand is that academia and industry both focus largely on security detection, but the emphasis is almost always on the algorithmic machinery powering the systems. We at Microsoft are transparent with our algorithm research and in fact are the only cloud provider to openly share the machine learning algorithms securing our cloud service. In order to build on that research and learn more about best practices for putting security data science solutions in production, we reached out to our peers in the industry.

We started by meeting with some friends at Google to swap ideas for keeping our cloud services and mutual customers secure. That one-time exercise proved so valuable that it soon turned into a recurring meeting wherein we learned that despite different approaches to data modeling, we face similar challenges. Last week, we opened the doors at Microsoft to the broader community. At first, we weren’t sure if companies would take us up the offer to discuss security data science issues in the open – nothing could have been farther from the truth. We quickly had delegates from Facebook, Salesforce, Crowdstrike, Google, LinkedIn, Endgame, Sqrrl, the Federal Reserve and researchers from the University of Washington. What was supposed to be an hour-long meetup, morphed into a full-blown conference – so much so, we had to give it a name – “Security Data Science Colloquium”.

The goal of the colloquium was simple: share learnings of how different cloud providers/services secure their systems using machine learning. No NDAs, no complicated back and forth paperwork. Our only constraint: keep it technical and be honest. This way, we could ensure that that the 300+ applied Machine Learning (ML) engineers, security analysts, and incident responders who signed up, had a collaborative environment to discuss freely!

Security Data Science > Security + Data Science

Operationalizing security and machine learning solutions is tricky, not only because security data science solutions are inherently complex from both fields, but also because their intersection poses new challenges. For instance, compliance restrictions that dictate data cannot be exported from specific geographic locations have a downstream effect on model design, deployment, evaluation, and management strategies (a data science constraint). As Adam Fuchs, CTO of Sqrrl, pointed out in his lecture, this complicated machinery requires a variety of actors to land an operational solution: threat hunters, data scientists, computer scientists and security analysts, in addition to the standard development crew of program managers, developers and service engineers.

Security Data Scientists ❤ Rules

To quote Sven Krasser (@SvenKrasser), Chief Scientist at Crowdstrike, “Rules are awesome”. This may come as a surprise to machine learning puritans who have long berated rules as futile tools. But as Sven noted in his talk, rules are very good at finding known maliciousness and we as a community must not shy away from them. During our smaller brainstorm discussions, we discussed various ways to combine rules and machine learning. For instance, at Microsoft, we have had success in using Markov Logic Networks to combine the domain knowledge of our security analysts and model them into probabilistic graphs.

Adversarial Machine Learning is Mainstream and We Don’t Know How to Solve It

Hyrum Anderson (@drhyrum) and Robert Filar’s (@filar) riveting talk on how adversaries can subvert machine learning solutions made defenders in the room uncomfortable (in a good way!). They showed different ways that attackers can successfully manipulate machine learning models, from partial to no access to the system. While instances of such attacks have been known since spammers have tried to evade detection, or when adversaries attempt to dodge antivirus systems, the biggest takeaway here is the Machine learning current system, like any system, is susceptible to attacks. For instance, attackers can use the labels alert outputs, or the decision label (such as malware or not), and work around these defenses. While this has been happening for some time, the game changer is that this feedback is instantaneous: the data that was designed as a way for defenders to act swiftly is now exploited by attackers. Research in this area is nascent, and we still don’t know how to bridge this gap.

Call for standardization and benchmarks

At our breakout sessions, we heard the need for a standardized benchmark dataset à la ImageNet – for instance, how do we know if the newest detection for anomalous process creation performs under various test cases. An interesting observation made by the “Security Platform” discussion group, was the need for something along the lines of “GitHub for feature engineering”. They reckoned that many teams waste time managing feature pipelines and sometimes re-computing the same feature, and wanted an effective management system that will make teams more efficient and code more maintainable.

The colloquium, thanks to the enthusiastic participation of our peers, ended up as a marketplace of security data science ideas – we discussed, agreed, and challenged one another with the intention of learning. My favorite quote about the conference, comes from a Salesforce participant, who remarked “we are all batting for the same team”. It particularly resonated with me, because despite our organizational boundaries, we all have a common goal: protect our customers from adversaries.

This is our commitment to share what we have learned – success and failures, so that you don’t have to waste time going down the wrong path. Given the overwhelming support from the security analytics community, my colleagues have already started planning on the next edition of the colloquium. If you are interested in participating, have ideas to make it better, or want to lend a helping hand in organizing, drop a note at ramk@microsoft.com or reach out to me on Twitter – @ram_ssk.



from Microsoft Secure Blog Staff

What are Confidence building measures (CBMs) and how can they improve cybersecurity?

Cyberspace security is too often viewed through a prism of technological terms and concepts. In my experience, even supposedly non-technical discussions of cyberspace quickly devolve into heated debates about “vulnerability coordination”, “the latest malware”, “the best analytical tools”, “threat information sharing”, and so on. While these are interesting and important topics, it is ultimately people and their personal perspectives – not technology – that largely shape governments’ political, diplomatic and military choices in cyberspace.

At the heart of government’s “human” decision-making in cyberspace are understanding and trust. The two are not the same. It’s possible for one state to understand another’s capabilities in cyberspace but not to trust their intentions. The reverse is also true, with trust existing outside of understanding another’s capabilities. But, by and large, some level of understanding about what another state can and can’t do in cyberspace should at least reduce distrust. And that can help governments make rational judgments about each other’s behaviors as well as de-escalate tensions between and among states.

One significant complication in building understanding and diffusing distrust is the fact that many systems useful in cyber-defense can also be used in cyber-offense. When a state invests in cyber to defend itself, its rivals might instead see a growth in offensive capabilities. This is not a question of technical understanding but rather of reading the intent of others. A very human response to someone seemingly gearing up for conflict is to build at the very least one’s own defenses (and to, potentially, even increase one’s offensive as well as retaliatory capabilities). Such a move is, however, equally liable to misinterpretation by others. Thus, escalation spreads, trust evaporates, and distrust balloons, leaving cyberspace, on which so much of modern life depends, akin to a powder keg, ready to explode. The potential for a cyber arms race is as real as it is dangerous.

An essential response to this critical challenge is the use of confidence building measures (CBMs) between states. Today, CBMs are still generally seen as vectors for instilling good cybersecurity practices, especially during a country’s early entry into cyberspace. Certainly, CBMs can help such countries counter the threat of cybercrime, and can also help promote international consistency in cybersecurity approaches, which is an essential part of combating cybercrime. However, CBMs are much more than this.

Coming of age under the threat of Cold War nuclear annihilation, CBMs enable states to minimize exactly the kind of misunderstandings that fuel distrust and exacerbate tensions. In many ways, they are akin to pressure valves for states to use before a situation escalates into conflict. CBMs can help states step back from thinking, “We need to get our cyber-retaliation in first”. They may not lead directly to trust but what they provide is manifestly better than its absence. They have a manifest role to play in ensuring the safety and stability of cyberspace by reducing the risk of cyberwar from breaking out. As such, they can be a necessary prerequisite to building trust.

CBMs are already being built into critical state-to-state cyberspace agreements. The UNGGE 2015 (voluntary) norms placed CBMs at the core of responsible state behavior in cyberspace. In the UNGGE’s words, they “allow the international community to assess the activities and intention of States”. That assessment of actions and intent is absolutely essential to addressing the human perspective. The UNGGE leveraged previous work done in the framework of the Organization for Security and Co-operation in Europe (OSCE), namely its 2013 CBMs. In this respect, it is significant that just last year the OSCE expanded on its CBM work precisely because, “events in cyberspace often leave room for ambiguity, speculation and misunderstanding. The worry is that miscalculations and misperceptions between states arising from activities in cyberspace could escalate, leading to serious consequences for citizens as well as for the economy and administration, and potentially fueling political tensions.”

A failure to mature and refine CBMs globally adds to distrust and militarization in cyberspace, i.e. the aforementioned cyber arms race. The consequences of the “miscalculations and misperceptions” that the OSCE warned of can easily move from the virtual world to the real one. For example, 2010’s so-called “Pakistan-India cyberwar” saw “cyber armies” from each country vandalizing official websites, exacerbating serious diplomatic and military tensions after the 2008 Mumbai terror attacks. Furthermore, recent tensions between parts of the West and Russia, North Korea or even China all feature strong elements of “cyber-distrust”. The danger, of course, is that once there is “cyber-distrust” among states it is likely spread into other spheres, if left unchecked, and vice versa.

So, if the human perspective matters at least as much as the technology when it comes to government decision-making about cyberspace, all parties should take every opportunity to promote understanding and reduced distrust between states. We should use whatever tools seem most appropriate to do so, . CBMs are essential in this regard. They are and remain a key tool in the cyber peacebuilder’s toolkit.

 



from Paul Nicholas

Tuesday, June 27, 2017

Tips for protecting your information and privacy against cybersecurity threats

This post is authored by Steven Meyers, security operations principal, Microsoft Cyber Defense Operations Center.

Introducing a new video on best practices from the Microsoft Cyber Defense Operations Center

In 2016, 4.2+ billion records were stolen by hackers. The number of cyberattacks and breaches in 2017 have risen 30 percent.

The business sector leads in the number of records compromised so far, with more than 7.5 million exposed records in 420 reported incidents. These cybercrimes are often intended for financial gain, such as opening a fraudulent credit card or accessing a company’s financial records. Today, a growing market exists in the dark web for selling credentials and sensitive information to other cybercriminals.

To help Protect your information and privacy against cyberthreats, the Microsoft Cyber Defense Operations Center has published a series of best practices videos that will help consumers, businesses and organizations enable a safer online environment. This video shares some of the policies and practices that can be used to better protect information and privacy inside and outside of your operational perimeters.

Protection starts with classifying information and then putting appropriate protections in place based on its value. Some information is meant to be public, some data is sensitive but not highly valued to outside entities, but some data is mission critical and/or could cause tremendous financial hardship if shared externally.

Cybersecurity technologies and policies such as multifactor authentication, the principles of least privilege access, just-in-time-and just-enough administrator access, and Microsoft’s cybersecurity products and services can help safeguard access to data and applications.

Some cybersecurity tips discussed include:

  • Classifying emails and data according to their level of sensitivity
  • Employing multifactor authentication for access to sensitive information
  • Only providing administrator access to individuals for the time needed to complete a task
  • Restricting access to only the information needed for the task
  • Keeping your software up-to-date

Please take a few minutes to watch the video and share it with your colleagues, friends and family. We all need to be diligent in the face of this growing and ever-more sophisticated threat.

Also, be sure to watch part one of the video series, Protecting your identity from cybersecurity threats. Check back next week for our third video, Protecting your devices from cybersecurity threats.

Additional resources:



from Microsoft Secure Blog Staff

Sunday, June 25, 2017

BSides Athens 2017 Wrap-Up

The second edition of BSides Athens was planned this Saturday. I already attended the first edition (my wrap-up is here) and I was happy to be accepted as a speaker for the second time!  This edition moved to a new location which was great. Good wireless, air conditioning and food. The day was based on three tracks: the first two for regular talks and the third one for the CTP and workshops. The “boss”, Grigorios Fragkos introduced the 2nd edition. This one gave more attention to a charity program called “the smile of the child” which helps Greek kids to remain in touch with the new technologies. A specific project is called “ODYSSEAS” and is based on a truck that travels across Greek to educate kids to technologies like mobile phones, social networks, … The BSides Athens donated to this project. A very nice initiative that was presented by Stefanos Alevizos who received a slot of a few minutes to describe the program (content in Greek only).


The keynote was assigned to Dave Lewis who presented “The Unbearable Lightness of Failure”. The main fact explained by Dave is that we fail but…we learn from our mistakes! In other words, “failure is an acceptable teaching tool“. The keynote was based on many facts like signs. We receive signs everywhere and we must understand how to interpret them or the famous Friedrich Nietzsche’s quote: “That which does not kill us makes us stronger“. We are facing failures all the time. The last good example is the Wannacry bad story which should never happen but… You know the story! Another important message is that we don’t have to be afraid t fail. We also have to share as much as possible not only good stories but also bad stories. Sharing is a key! Participate in blogs, social networks, podcasts. Break out of your silo! Dave is a renowned speaker and delivered a really good keynote!

Then talks were split across the two main rooms. For the first one, I decided to attend the Thanissis Diogos’s presentation about “Operation Grand Mars“. In January 20167, Trustwave published an article which described this attack. Thanassis came back on this story with more details. After a quick recap about what is incident management, he reviewed all the fact related to the operation and gave some tips to improve abnormal activities on your network. It started with an alert generated by a workstation and, three days later, the same message came from a domain controller. Definitively not good! The entry point was infected via a malicious Word document / Javascript. Then a payload was download from Google docs which is, for most of our organization, a trustworthy service. Then he explained how persistence was achieved (via autorun, scheduled tasks) and also lateral movements. The pass-the-hash attack was used. Another tip from Thanissis: if you see local admin accounts used for network logon, this is definitively suspicious! Good review of the attack with some good tips for blue teams.

My next choice was to move to the second track to follow Konstantinos Kosmidis‘s talk about machine learning (a hot topic today in many conferences!). I’m not a big fan of these technologies but I was interested in the abstract. The talk was a classic one: after an introduction to machine learning (that we already use every day with technologies like the Google face recognition, self-driving card or voice-recognition), why not apply this technique to malware detection. The goal is to: detect, classify but, more important, to improve the algorithm! After reviewing some pro & con, Konstantinos explained the technique he used in his research to convert malware samples into images. But, more interesting, he explained a technique based on steganography to attack this algorithm. The speaker was a little bit stressed but the idea looks interesting. If you’re interested, have a look at his Github repository.

Back to the first track to follow Professor Andrew Blyth with “The Role of Professionalism and Standards in Penetration Testing“. The penetration testing landscape changed considerably in the last years. We switched to script kiddies search for juicy vulnerabilities to professional services. The problem is that today some pentest projects are required not to detect security issues and improve but just for … compliance requirements. You know the “checked-case” syndrome. Also, the business evolves and is requesting more insurance. The coming GDP European regulation will increase the demand in penetration tests.  But, a real pentest is not a Nessus scan with a new logo as explained Andrew! We need professionalism. In the second part of the talk, Andrew reviewed some standards that involve pentests: iCAST, CBEST, PCI, OWASP, OSSTMM.

After a nice lunch with Greek food, back to talks with the one of Andreas Ntakas and Emmanouil Gavriil about “Detecting and Deceiving the Unknown with Illicium”. They are working for one of the sponsors and presented the tool developed by their company: Illicium. After the introduction, my feeling was that it’s a new honeypot with extended features.  Not only, they are interesting stuff but, IMHO, it was a commercial presentation. I’d expect a demo. Also, the tool looks nice but is dedicated to organization that already reached a mature security level. Indeed, before defeating the attacker, the first step is to properly implement basic controls like… patching! What some organizations still don’t do today!

The next presentation was “I Thought I Saw a |-|4><0.-” by Thomas V. Fisher.  Many interesting tips were provided by Thomas like:

  • Understand and define “normal” activities on your network to better detect what is “abnormal”.
  • Log everything!
  • Know your business
  • Keep in mind that the classic cyber kill-chain is not always followed by attackers (they don’t follow rules)
  • The danger is to try to detect malicious stuff based on… assumptions!

The model presented by Thomas was based on 4 A’s: Assess, Analyze, Articulate and Adapt! A very nice talk with plenty of tips!

The next slot was assigned to Ioannis Stais who presented his framework called LightBulb. The idea is to build a framework to help in bypassing common WAF’s (web application firewalls). Ioannis explained first how common WAF’s are working and why they could be bypassed. Instead of testing all possible combinations (brute-force), LightBuld relies on the following process:

  • Formalize the knowledge in code injection attacks variations.
  • Expand the knowledge
  • Cross check for vulnerabilities

Note that LightBulb is available also as a BurpSuipe extension! The code is available here.

Then, Anna Stylianou presented “Car hacking – a real security threat or a media hype?“. The last events that I attended also had a talk about cars but they focused more on abusing the remote control to open doors. Today, it focuses on ECU (“Engine Control Unit”) that are present in modern cars. Today a car might have >100 ECU’s and >100 millions lines of code which means a great attack surface! They are many tools available to attack a car via its CAN bus, even the Metasploit framework can be used to pentest cars today! The talk was not dedicated to a specific attack or tools but was more a recap of the risks that cars manufacturers are facing today. Indeed, threats changed:

  • theft from the car (breaking a window)
  • theft of the cat
  • but today: theft the use of the car (ransomware)

Some infosec gurus also predict that autonomous cars will be used as lethal weapons! As cars can be seen as computers on wheels, the potential attacks are the same: spoofing, tampering, repudiation, disclosure, DoS or privilege escalation issues.

The next slot was assigned to me. I presented “Unity Makes Strength” and explained how to improve interconnections between our security tools/applications. The last talk was performed by Theo Papadopoulos: A “Shortcut” to Red Teaming. He explained how .LNK files can be a nice way to compromize your victim’s computer. I like the “love equation”: Word + Powershell = Love. Step by step, Theo explained how to build a malicious document with a link file, how to avoid mistakes and how to increase chances to get the victim infected. I like the persistence method based on assigning a popular hot-key (like CTRL-V) to shortcut on the desktop. Windows will trigger the malicious script attached to the shortcut and them… execute it (in this case, paste the clipboard content). Evil!

The day ended with the CTF winners announce and many information about the next edition of BSides Athens. They already have plenty of ideas! It’s now time for some off-days across Greece with the family…

[The post BSides Athens 2017 Wrap-Up has been first published on /dev/random]



from Xavier

Thursday, June 22, 2017

"Hacker Village - At the #SecAwareSummit"

Editor's Note: Taylor Lobbis a security community manager for developers within Adobe. Heis one of the speakers for the upcoming Security Awareness Summit 2/3 Aug in Nashville, TN. Below hegives an overview on hisupcoming talk onbuilding a Hacker Village. I am a manager of security and privacy engineering for Adobe. One of our core goals &hellip; Continue reading Hacker Village - At the #SecAwareSummit

from lspitzner

[SANS ISC] Obfuscating without XOR

I published the following diary on isc.sans.org: “Obfuscating without XOR“.

Malicious files are generated and spread over the wild Internet daily (read: “hourly”). The goal of the attackers is to use files that are:

  • not know by signature-based solutions
  • not easy to read for the human eye

That’s why many obfuscation techniques exist to lure automated tools and security analysts… [Read more]

[The post [SANS ISC] Obfuscating without XOR has been first published on /dev/random]



from Xavier

Wednesday, June 21, 2017

Tips for securing your identity against cybersecurity threats

This post is authored by Simon Pope, Principal Security Group Manager, Microsoft Security Response Center.

Introducing new video on best practices from the Microsoft Cyber Defense Operations Center

Ask any CISO or cybersecurity professional about their greatest security challenge, and it’s a good chance the answer will be “the actions of our people.”

While virtually all employees, contractors, and partners have the best of intentions, the fact is that protecting their online credentials, identifying and avoiding phishing scams, and evading cybercriminals is getting more difficult each day. More of our time each day is spent online, and as more financial transactions and social activities are conducted online, adversaries are becoming ever-more sophisticated in their cyberattacks.

Microsoft faces these same threats, and we have made deep investments in training our people to be more aware and diligent in the face of such dangers. Our cybersecurity success depends on our customers’ trust in our products and services, and their confidence that they can be safe on the internet. To help keep our customers and the global online community safe, we want to share some of our Cyber Defense Operations Center’s best practices for Securing your identity against cybersecurity threats in this video.

In this video, we discuss some best practices around securing your identity, such as avoiding social engineering scams that trick people into giving up their most sensitive secrets, recognizing phishing emails that falsely represent legitimate communications, and how to spot false impersonations of your trusted colleagues or friends. We also discuss some of the types of information you don’t want to share broadly (i.e. credentials, financial information and passwords), and tips for protecting your sensitive data.

Some cybersecurity tips that we discuss include:

  • Be vigilant against phishing emails
  • Be cautious when sharing sensitive information
  • Don’t automatically trust emails from people you know, it may not be from them
  • Keep your software up-to-date

Please take a few minutes to watch the video and share it with your colleagues, friends and family. We all need to be diligent in the face of this growing and ever-more sophisticated threat. And check back next week for our second video on Protecting your devices from cybersecurity threats, and in two weeks, we will share more on Protecting your information and data from cybersecurity threats on the Microsoft Secure blog.

Additional resources:



from Microsoft Secure Blog Staff