Archive for the ‘Evolution of Microsoft Threat Protection’ Category

Lessons learned from the Microsoft SOC—Part 3d: Zen and the art of threat hunting

June 25th, 2020 No comments

An image of a black male developer at work in an Enterprise office workspace.

Threat hunting is a powerful way for the SOC to reduce organizational risk, but it’s commonly portrayed and seen as a complex and mysterious art form for deep experts only, which can be counterproductive. In this and the next blog we will shed light on this important function and recommend simple ways to get immediate and meaningful value out of threat hunting.

This is the seventh blog in the Lessons learned from the Microsoft SOC series designed to share our approach and experience from the front lines of our security operations center (SOC) protecting Microsoft, and our Detection and Response Team (DART) helping our customers with their incidents. For a visual depiction of our SOC philosophy, download our Minutes Matter poster.

Before we dive in, let’s clarify the definition of “threat hunting.”  There are various disciplines and processes that contribute to the successful proactive discovery of threat actor operations. For example, our Hunting Team works with threat intelligence to help shape and guide their efforts, but our threat intelligence teams are not “threat hunters.”  When we use the term “threat hunting,” we are talking about the process of experienced analysts proactively and iteratively searching through the environment to find attacker operations that have evaded other detections.

Hunting is a complement to reactive processes, alerts, and detections, and enables you to proactively get ahead of attackers. What sets hunting apart from reactive activities is the proactive nature of it, where hunters spend extended focus time thinking through issues, identifying trends and patterns, and getting a bigger picture perspective.

A successful hunting program is not purely proactive however as it requires continuously balancing attention between reactive efforts and proactive efforts. Threat hunters will still need to maintain a connection to the reactive side to keep their skills sharp and fresh and keep attuned to trends in the alert queue. They will also need to jump in to help with major incidents at a moment’s notice to help put out the fire. The amount of time available for proactive activities will depend heavily on whether or not you have a full-time or part-time hunting mission.

Our SOC approaches threat hunting by applying our analysts to different types of threat hunting tasks:

1. Proactive adversary research and threat hunting

This is what most of our threat hunters spend the majority of their time doing. The team searches through a variety of sources including alerts, external indicators of compromise and other sources. The team primarily works to build and refine structured hypotheses of what the attackers may do based on threat intelligence (TI), unusual observations in the environment, and their own experience. In practice, this type of threat hunting includes:

  • Proactive search through the data (queries or manual review).
  • Proactive development of hypotheses based on TI and other sources.

2. Red and purple teaming

Some of our threat hunters work with red teams who simulate attacks and others who conduct authorized penetration testing against our environment. This is a rotating duty for our threat hunters and typically involves purple teaming, where both red and blue teams work to do their jobs and learn from each other. Each activity is followed up by fully transparent reviews that capture lessons learned which are shared throughout the SOC, with product engineering teams, and with other security teams in the company.

3. Incidents and escalations

Proactive hunters aren’t sequestered somewhere away from the watch floor. They are co-located with reactive analysts; they frequently check in with each other, share what they are working on, share interesting findings/observations, and generally maintain situational awareness of current operations. Threat hunters aren’t necessarily assigned to this task full time; they may simply remain flexible and jump in to help when needed.

These are not isolated functions— the members of these teams work in the same facility and frequently check in with each other, share what they are working on, and share interesting findings/observations.

What makes a good threat hunter?

While any high performing analyst has good technical skills, a threat hunter must be able to see past technical data and tools to attackers’ actions, motivations, and ideas. They need to have a “fingertip feel” (sometimes referred to as Fingerspitzengefühl), which is a natural sense of what is normal and abnormal in security data and the environment. Threat hunters can recognize when an alert (or cluster of alerts/logs) seem different or out of place.

One way to think about the qualities that make up a good threat hunter is to look at the Three F’s.


This is technical knowledge and competency of investigating and remediating incidents. Security analysts (including threat hunters) should be proficient with the security tools, general flow of investigation and remediation, and the types technologies commonly deployed in enterprise environments.


This is “know thyself” and “know thy enemy” and includes familiarity with your organization’s specific environment and familiarity with attacker tactics, techniques, and procedures (TTPs). Attacker familiarity starts with understanding common adversary behaviors and then grows into a deeper sense of specific adversaries (including technologies, processes, playbooks, business priorities and mission, industry, and typical threat patterns). Familiarity also includes the relationships threat hunters develop with the people in your organization, and their roles/responsibilities. Familiarity with your organization is highly valued for analysts on investigation teams, and critical for effective threat hunting.


Flexibility is a highly valued attribute of any analyst role, but it is absolutely required for a threat hunter. Flexibility is a mindset of being adaptable in what you may do every day and how you do it. This manifests in how you understand problems, process information, and pursue solutions. This mindset comes from within each person and is reflected in almost everything they do.

Where any threat analyst (or threat hunter) can take a particular alert or event and run it into the ground, a good threat hunter will take a step back and look at a collection of data, alerts or events. Threat hunters must be inquisitive and unrelentingly curious about things—to the point that it bugs them if they don’t have a clear understanding of something. Instead of just answering a question, threat hunters are constantly trying to ask better questions of the data, coming up with creative new angles to answer them, and seeing what new questions they raise. Threat hunting also requires humility, to be able to quickly admit your mistakes so you can rapidly re-enter learning mode.

Threat hunting tooling

Threat hunting naturally pulls in a wide variety of tools, but our team has grown to prefer a few of the Microsoft tools whose design they have influenced.

  • Advanced hunting in Microsoft Threat Protection (MTP) tends to be the go-to tool for anything related to endpoints, identities, email, Azure resources, and SaaS applications.
  • Our teams also use Azure Sentinel, Jupyter notebooks, and custom analytics to hunt across broad datasets like application and network data, as well as diving deeper into identity, endpoint, Office 365, and other log data.

Our threat hunting teams across Microsoft contribute queries, playbooks, workbooks, and notebooks to the Azure Sentinel Community, including specific hunting queries that your teams can adapt and use.


We have discussed the art of threat hunting, different approaches to it, and what makes a good threat hunter. In the next entry, we dive deeper into how to build and refine a threat hunting program. If you are looking for more information on the SOC and other cybersecurity topics, check out previous entries in the series (Part 1 | Part 2a | Part 2b | Part 3a | Part 3b| Part 3c), Mark’s List, and our new security documentation site. Be sure to bookmark the Security blog to keep up with our expert coverage on security matters. Also, follow us at @MSFTSecurity for the latest news and updates on cybersecurity.

The post Lessons learned from the Microsoft SOC—Part 3d: Zen and the art of threat hunting appeared first on Microsoft Security.

Defending the power grid against supply chain attacks: Part 3 – Risk management strategies for the utilities industry

April 22nd, 2020 No comments

Over the last fifteen years, attacks against critical infrastructure (figure1) have steadily increased in both volume and sophistication. Because of the strategic importance of this industry to national security and economic stability, these organizations are targeted by sophisticated, patient, and well-funded adversaries.  Adversaries often target the utility supply chain to insert malware into devices destined for the power grid. As modern infrastructure becomes more reliant on connected devices, the power industry must continue to come together to improve security at every step of the process.

Aerial view of port and freeways leading to downtown Singapore.

Figure 1: Increased attacks on critical infrastructure

This is the third and final post in the “Defending the power grid against supply chain attacks” series. In the first blog I described the nature of the risk. Last month I outlined how utility suppliers can better secure the devices they manufacture. Today’s advice is directed at the utilities. There are actions you can take as individual companies and as an industry to reduce risk.

Implement operational technology security best practices

According to Verizon’s 2019 Data Breach Investigations Report, 80 percent of hacking-related breaches are the result of weak or compromised passwords. If you haven’t implemented multi-factor authentication (MFA) for all your user accounts, make it a priority. MFA can significantly reduce the likelihood that a user with a stolen password can access your company assets. I also recommend you take these additional steps to protect administrator accounts:

  • Separate administrative accounts from the accounts that IT professionals use to conduct routine business. While administrators are answering emails or conducting other productivity tasks, they may be targeted by a phishing campaign. You don’t want them signed into a privileged account when this happens.
  • Apply just-in-time privileges to your administrator accounts. Just-in-time privileges require that administrators only sign into a privileged account when they need to perform a specific administrative task. These sign-ins go through an approval process and have a time limit. This will reduce the possibility that someone is unnecessarily signed into an administrative account.


Image 2

Figure 2: A “blue” path depicts how a standard user account is used for non-privileged access to resources like email and web browsing and day-to-day work. A “red” path shows how privileged access occurs on a hardened device to reduce the risk of phishing and other web and email attacks. 

  • You also don’t want the occasional security mistake like clicking on a link when administrators are tired or distracted to compromise the workstation that has direct access to these critical systems.  Set up privileged access workstations for administrative work. A privileged access workstation provides a dedicated operating system with the strongest security controls for sensitive tasks. This protects these activities and accounts from the internet. To encourage administrators to follow security practices, make sure they have easy access to a standard workstation for other more routine tasks.

The following security best practices will also reduce your risk:

  • Whitelist approved applications. Define the list of software applications and executables that are approved to be on your networks. Block everything else. Your organization should especially target systems that are internet facing as well as Human-Machine Interface (HMI) systems that play the critical role of managing generation, transmission, or distribution of electricity
  • Regularly patch software and operating systems. Implement a monthly practice to apply security patches to software on all your systems. This includes applications and Operating Systems on servers, desktop computers, mobile devices, network devices (routers, switches, firewalls, etc.), as well as Internet of Thing (IoT) and Industrial Internet of Thing (IIoT) devices. Attackers frequently target known security vulnerabilities.
  • Protect legacy systems. Segment legacy systems that can no longer be patched by using firewalls to filter out unnecessary traffic. Limit access to only those who need it by using Just In Time and Just Enough Access principles and requiring MFA. Once you set up these subnets, firewalls, and firewall rules to protect the isolated systems, you must continually audit and test these controls for inadvertent changes, and validate with penetration testing and red teaming to identify rogue bridging endpoint and design/implementation weaknesses.
  • Segment your networks. If you are attacked, it’s important to limit the damage. By segmenting your network, you make it harder for an attacker to compromise more than one critical site. Maintain your corporate network on its own network with limited to no connection to critical sites like generation and transmission networks. Run each generating site on its own network with no connection to other generating sites. This will ensure that should a generating site become compromised, attackers can’t easily traverse to other sites and have a greater impact.
  • Turn off all unnecessary services. Confirm that none of your software has automatically enabled a service you don’t need. You may also discover that there are services running that you no longer use. If the business doesn’t need a service, turn it off.
  • Deploy threat protection solutions. Services like Microsoft Threat Protection help you automatically detect, respond to, and correlate incidents across domains.
  • Implement an incident response plan: When an attack happens, you need to respond quickly to reduce the damage and get your organization back up and running. Refer to Microsoft’s Incident Response Reference Guide for more details.

Speak with one voice

Power grids are interconnected systems of generating plants, wires, transformers, and substations. Regional electrical companies work together to efficiently balance the supply and demand for electricity across the nation. These same organizations have also come together to protect the grid from attack. As an industry, working through organizations like the Edison Electric Institute (EEI), utilities can define security standards and hold manufacturers accountable to those requirements.

It may also be useful to work with The Federal Energy Regulatory Committee (FERC), The North American Electric Reliability Corporation (NERC), or The United States Nuclear Regulatory Commission (U.S. NRC) to better regulate the security requirements of products manufactured for the electrical grid.

Apply extra scrutiny to IoT devices

As you purchase and deploy IoT devices, prioritize security. Be careful about purchasing products from countries that are motivated to infiltrate critical infrastructure. Conduct penetration tests against all new IoT and IIoT devices before you connect them to the network. When you place sensors on the grid, you’ll need to protect them from both cyberattacks and physical attacks. Make them hard to reach and tamper-proof.

Collaborate on solutions

Reducing the risk of a destabilizing power grid attack will require everyone in the utility industry to play a role. By working with manufacturers, trade organizations, and governments, electricity organizations can lead the effort to improve security across the industry. For utilities in the United States, several public-private programs are in place to enhance the utility industry capabilities to defend its infrastructure and respond to threats:

Read Part 1 in the series: “Defending the power grid against cyberattacks

Read “Defending the power grid against supply chain attacks: Part 2 – Securing hardware and software

Read how Microsoft Threat Protection can help you better secure your endpoints.

Learn how MSRC developed an incident response plan

Bookmark the Security blog to keep up with our expert coverage on security matters. For more information about our security solutions visit our website. Also, follow us at @MSFTSecurity for the latest news and updates on cybersecurity.

The post Defending the power grid against supply chain attacks: Part 3 – Risk management strategies for the utilities industry appeared first on Microsoft Security.

MITRE ATT&CK APT 29 evaluation proves Microsoft Threat Protection provides deeper end to end view of advanced threats

April 21st, 2020 No comments

As attackers use more advanced techniques, it’s even more important that defenders have visibility not just into each of the domains in their environment, but also across them to piece together coordinated, targeted, and advanced attacks. This level of visibility will allow us to get ahead of attackers and close the gaps through which they enter. To illustrate that imperative, the 2019 MITRE ATT&CK evaluation centered on an advanced nation-state threat actor known to the industry as Advanced Persistent Threat (APT) 29 (also known as Cozy Bear) which largely overlaps with the activity group that Microsoft calls YTTRIUM. . The test involved a simulation of 58 attacker techniques in 10 kill chain categories.

Microsoft participated in the second MITRE ATT&CK endpoint detection product evaluation published today. The evaluation is designed to test security products based on the ATT&CK (Adversarial Tactics, Techniques & Common Knowledge) framework, which is highly regarded in the security industry as one of the most comprehensive catalog of attacker techniques and tactics. Threat hunters use this framework to look for specific techniques that attackers often use to penetrate defenses. Testing that incorporates a comprehensive view of an environment’s ability to monitor and detect malicious activity with the existing tools that defenders have deployed across an organization is critical.

Although this test was focused on endpoint detection and response, MITRE ran the simulated APT29 attack from end to end and across multiple attack domains, meaning defenders benefited from visibility beyond just endpoint protection. This gave Microsoft the unique opportunity to bring Microsoft Threat Protection (MTP) to the test.

Microsoft Threat Protection expands Microsoft Defender ATP from endpoint detection and response (EDR) to an extended detection and response (XDR) solution, and is designed to provide extended detection and response by combining protection for endpoints (Microsoft Defender ATP), email and productivity tools (Office 365 ATP), identity (Azure ATP), and cloud applications (Microsoft Cloud App Security/MCAS). As customers face attacks across endpoints, cloud, applications and identities, MTP looks across these domains to understand the entire chain of events, identifies affected assets, like users, endpoints, mailboxes, and applications, and auto-heals them back to a safe state.

Microsoft Threat Protection delivers coverage across the entire kill chain, not just the endpoint

To fully execute the end to end attack simulation of APT29, MITRE required participants to turn off all proactive protection and blocking capabilities. For Microsoft Threat Protection, this meant that all the capabilities that would normally block this kind of attack such as automatic remediation flows, application isolation, attack surface reduction, network protection, exploit protection, controlled folder access, and next-gen antivirus prevention were turned off. However, Microsoft Threat Protection audit capabilities for these features enabled recording of a variety of points during the attack when MTP (had it been fully enabled) would have prevented or blocked execution, likely stopping the attack in its tracks.

During this evaluation Microsoft Threat Protection delivered on providing the deep and broad optics, near real time detection through automation, and a complete, end-to-end view of the attack story. Here is how Microsoft Threat Protection stood out:

  • Depth and breadth of optics: Our uniquely integrated operating system, directory, and cloud sensors contributed deep and broad telemetry coverage. AI-driven, cloud-powered models collaborating across domains identified malicious activities and raised alerts on attacker techniques across the entire attack kill chain:
    • Microsoft Defender ATP recorded and alerted on endpoint activities including advanced file-less techniques, privilege escalation, and credential theft and persistence – leveraging deep sensors like AMSI, WMI, and LDAP.
    • Azure ATP watched and detected account compromise at the domain level, and lateral movement, such as pass-the-hash and the more sophisticated pass-the-ticket (Golden Ticket attack).
    • Microsoft Cloud App Security identified exfiltration of data to the cloud (OneDrive).
  • Detection and containment in near real time:Nation state attacks of this magnitude can take place over the course of as little as a few hours, which means that Security Operations Centers (SOCs) often have little to no time to respond. Near-real-time automated detection of advanced techniques is critical to address this challenge. Where possible, active blocking, prevention and automatic containment will make the difference between an attempted versus a successful compromise. MTP’s prevention capabilities along with fast detection and behavioral blocking are exactly designed for this purpose.
  • A complete attack story: Throughout this evaluation, Microsoft Defender ATP, Azure ATP, and Microsoft Cloud App Security, combined with the expertise of Microsoft Threat Experts generated nearly 80 alerts – for SOC teams, manually following up on each one of these alerts is overwhelming. MTP consolidated the alerts into just two incidents, dramatically simplifying the volume of triage and investigation work needed. This gives the SOC the ability to prioritize and address the incident as a whole and enables streamlined triage, investigation, and automated response process against the complete attack. With MTP we have built in automation that identifies the complex links between attacker activities and builds correlations across domains that piece together the attack story with all of its related alerts, telemetry, evidence and affected assets into coherent incidents. These comprehensive incidents are then prioritized and escalated to the SOC.


Microsoft Threat Experts, our managed threat hunting service, also participated in the evaluation this year. Our security experts watched over the signals collected in real time and generated comprehensive, complementary alerts, which enriched the automated detections with additional details, insights and recommendations for the SOC.

Real world testing is critical

Attackers are using advanced, persistent, and intelligent techniques to penetrate today’s defenses. This method of testing leans heavily into real-world exploitations rather than those found solely in a lab or simulated testing environment. Having been part of the inaugural round of the MITRE ATT&CK evaluation in 2018, Microsoft enthusiastically took on the challenge again, as we believe this to be a great opportunity, alongside listening to customers and investing in research, to continuously drive our security products to excellence and protect our customers.

This year, for the first time, we were happy to answer the community call from MITRE, alongside other security vendors, to contribute unique threat intelligence and research content about APT29, as well as in evolving the evaluation based on the experience and feedback from last year, yielding a very collaborative and productive process.

Thank you to MITRE and our customers and partners for your partnership in helping us deliver more visibility and automated protection, detection, response, and prevention of threats for our customers.

– Moti Gindi, CVP, Microsoft Threat Protection

The post MITRE ATT&CK APT 29 evaluation proves Microsoft Threat Protection provides deeper end to end view of advanced threats appeared first on Microsoft Security.