2. MOVING TO AN ENDPOINT-CENTRIC ZERO TRUST SECURITY MODEL WITH SENTINELONE
SENTINELONE WHITEPAPER 2
TableofContents
Abstract
3
Introduction
3
History
4
Zero Trust Overview 5
Attack Simulation 6
Zero Trust Maturity Level 7
Building a Zero Trust security model 8
Zero Trust Best Practices 9
Zero Trust powered by SentinelOne 13
Conclusion
18
3. MOVING TO AN ENDPOINT-CENTRIC ZERO TRUST SECURITY MODEL WITH SENTINELONE
SENTINELONE WHITEPAPER 3
Introduction
The requirements for security professionals have changed dramatically in recent years with
the rapidly evolving threat landscape, redefinition of how and where employees work, and
adoption of new technologies like cloud computing, Internet of Things (IoT), and 5G. Today,
security teams embrace the concept of assuming breach, as many understand that it’s not
if you will be compromised, but when. Therefore, organizations invest in people, processes,
and technology that help them protect, detect, respond, and recover from threats as effec-
tively as possible.
Historically, most corporate applications and solutions that store corporate data were pro-
tected behind the corporate network. With the adoption of cloud applications and the mo-
bile workforce, this has changed dramatically. Today, many applications or storage solu-
tions that were unthinkable to be accessible outside the corporate network are hosted on
cloud-native solutions and are accessible from virtually anywhere. For this reason, the old
perimeter that security professionals would set and protect no longer exists, and perime-
ter-based security models are obsolete. As a result, many organizations are looking for a
new security model that helps them to protect against the modern threat landscape, sup-
ports remote work scenarios, and reduces the attack surface to a minimum aperture. These
are all critical capabilities of Zero Trust, and when successfully adopted by an organization,
they can equip organizations with a robust and flexible security model.
Today, endpoints represent the most significant attack surface, with over 70% of breaches1
originating on the endpoint. Organizations have a heterogeneous mix of them connected to
their network, whether laptops, mobile endpoints, servers, or IoT devices. Therefore, Sen-
tinelOne recommends an endpoint-centric Zero Trust security model that works regardless
of where the user is located, whether in the corporate network, at a coffee shop, or at home.
SentinelOne’s approach brings trust verification as close to the user as possible, turning
endpoints into effective zero-trust policy enforcement mechanisms.
1 IDC: 70% of Successful Breaches Originate on the Endpoint (Rapid7)
01
02
Abstract
Zero Trust Network (ZTN) concept follows the mantra of never trust, always verify. Through
this approach, organizations can reduce their open attack surface and adopt enhanced
security capabilities beyond traditional defenses. In this paper, you will learn how Sen-
tinelOne can help enterprises embrace an endpoint-centric zero trust strategy using the
SentinelOne Singularity™ platform.
4. MOVING TO AN ENDPOINT-CENTRIC ZERO TRUST SECURITY MODEL WITH SENTINELONE
SENTINELONE WHITEPAPER 4
History
There have been many security paradigms over the past decades including defense-in-
depth, least privilege, micro-segmentation, containerization, and Multi Factor Authentica-
tion (MFA). Each of these paradigms represents a mindset of how the different security
tools in your organization should be architected to work with each other.
Layered network defenses have been the traditional approach to security for decades.
Network-centric methods relied heavily on physical sensors—like firewalls, Intrusion Pre-
vention System (IPS), and Intrusion Detection System (IDS)—to control and secure north-
south traffic. Access to corporate resources was binary, based on whether the user was
inside or outside the firewall. Once inside the firewall, trust was implicit and given freely.
In recent years, rises in phishing, supply-chain-based attacks, insider threats, and cre-
dential compromise attacks have made organizations reconsider the ‘trust by default’
approach. Insider credentials can be taken advantage of for elevated access, therefore
presenting themselves as attractive targets for attackers. In contrast to attacks originating
from outside of the corporate network, adversaries can leverage the implicit trust given to
identity to move laterally within an organization’s network.
Additionally, the COVID-19 pandemic has accelerated digital transformation efforts for
many organizations. IT teams were forced to rapidly stand up infrastructure to support an
instant remote and later hybrid workforce. New solutions were deployed to enable busi-
ness continuity, including cloud infrastructure and Software-as-a-Service (SaaS) platforms
like Zoom and Office 365. IT teams adopted solutions that could scale and deploy without
needing access to the physical data center, in some cases deploying applications that were
exposed to the open internet. In parallel, many organizations needed to provide endpoints
for new remote employees and roll out bring your own device (BYOD) programs. In reality,
securing these new operating environments was a secondary concern.
COMMERCIAL
CLOUD
Trusted Untrusted
Employees
Partners
IoT
Public
Trusted (Perimeter)
• Current state
• Implicit trust within
network
• Perimeter security
controls
• Firewall ACLs
• Static policies
Endpoints
AWS
MS Azure
Google Cloud
Platform
Office 365
G Suite
SaaS
Campus Corporate
Network IT
Public
03
Legacy
Perimeter
Internet
5. MOVING TO AN ENDPOINT-CENTRIC ZERO TRUST SECURITY MODEL WITH SENTINELONE
SENTINELONE WHITEPAPER 5
ZeroTrustOverview
Whereas legacy models are focused on neutralizing threats originating outside an orga-
nization’s network, Zero Trust acknowledges that threats may well exist both inside and
outside the network. Legacy security models trust, by default, the endpoints and identities
within their sphere of influence; In contrast, Zero Trust follows the principle of never trust,
always verify all endpoints, all identities. By successfully adopting Zero Trust, organiza-
tions can perform risk-based access control and leverage the concept of least privileged
access for every access decision.
ForresterdefinesZeroTrustas“movingsecurityfromanetwork-oriented,perimeter-based
security model to one based on continuous verification of trust.” Zero Trust is not a prod-
uct, but rather a modern security model composed of multiple cooperative trust verifica-
tion layers that are triggered and tested regardless of the device’s location. A Zero Trust
ecosystem aggregates multiple sources of trust signals from identity, endpoint, workload,
and network to provide a point-in-time access decision.
Guiding Principles of Zero Trust
• Never trust, always verify
Treat every user, endpoint, application or workload, and data flow as untrusted.
Authenticate and explicitly authorize each to the least privilege required.
04
All of these radical shifts resulted in users accessing applications and data outsides of the
traditional corporate network. While some organizations tried to scale their on-premises
infrastructure to cope, creating a new perimeter around the new compute-where-you-are
operating environment with legacy tooling requires too much effort and is prohibitively ex-
pensive. Changes in attacks and attack surfaces have necessitated a new approach that
ensures every endpoint can be trusted.
COMMERCIAL
CLOUD
Zero Trust
(Perimeter-less)
• Eliminate implicit
trust
• Incorporate trust
signals
• Conditional risk-
aware policies
• Network micro-
segmentation
• Visibility and
analytics
Campus
Employees
Partners
IoT
Endpoints
Cloud-Edge
VPN /
ZTNA
IT
Internet
Public
Trust Boundary
ZeroTrust
Perimeter
Workloads
& Data
At-rest &
In-transit
AWS
MS Azure
Google Cloud
Platform
Office 365
G Suite
SaaS
Hybrid /
Remote
Work
6. AttackSimulation
By comparing two attack simulations, we can begin to understand the strengths of Zero
Trust’ continuous security compared to the prototypical (and now legacy) perimeter-cen-
tric security model with a hard exterior and soft interior.
MOVING TO AN ENDPOINT-CENTRIC ZERO TRUST SECURITY MODEL WITH SENTINELONE
SENTINELONE WHITEPAPER 6
05
Legacy
Architecture
ZeroTrust
Architecture
Legacy
Architecture
Zero Trust
Architecture
Endpoints and identities
are trusted by default
No inherent trust by
default
Malicious
Access
Malicious
Access
Compromised employee
identity or unmanaged
endpoint gives threat actor
access to environment
Compromised employee
identity or unmanaged
endpoint gives threat actor
access to environment
Malicious
Activity
Malicious
Activity
Threat actor has
employee’s access and can
move laterally
Access attempts fail due to
Zero Trust Network Access
device posture controls
• Assume breach
Operate and defend resources with the assumption that an adversary already has a
presence within the environment. Deny by default and scrutinize all users, endpoints,
data flows, and requests for access.
• Verify explicitly
Dictate access to all resources in a consistent and secure manner using multiple trust
signals for contextual access decisions.
7. MOVING TO AN ENDPOINT-CENTRIC ZERO TRUST SECURITY MODEL WITH SENTINELONE
SENTINELONE WHITEPAPER 7
ZeroTrustMaturityLevel
As organizations move from a legacy to Zero Trust security model, they look for best practic-
es and guidelines on achieving said model as quickly as possible, but changing the security
model of an organization isn’t achieved overnight. The journey is a marathon, not a sprint.
While some of the existing investments of an organization can be leveraged or integrated
into a Zero Trust security model, the transition will require additional capabilities and re-
sources to fully utilize all the benefits of a Zero Trust security model. To achieve that, Sen-
tinelOne recommends the following components:
Traditional – manual configurations and attribute assignment, static security policies,
least-function established at provisioning, proprietary and inflexible policy enforcement,
manual incident response, and mitigation capability.
Advanced – some cross-solution coordination, centralized visibility, centralized identity
control, policy enforcement based on cross-solution inputs and outputs, some incident re-
sponse to pre-defined mitigations, some least-privilege changes based on posture assess-
ments.
Optimal – fully automated assigning of attributes to assets and resources, dynamic policies
based on automated/observed triggers, assets have dynamic least-privilege access (within
thresholds), alignment with open standards for cross-pillar interoperability, centralized vis-
ibility with retention for historical review.
06
Identity Endpoint Network Workload
• Password or
multifactor
authentication (MFA)
• Limited risk
assessment
• Limited visibility into
compliance
• Basic inventory
• Large macro-
segmentation
• Minimal encryption
• Access based on local
authentication
• General purpose
protection for known
threats
• Some cloud accessibility
• MFA
• Identity federation
with cloud and on-
prem
• Compliance
enforcement
• Data access based
on device posture
• Micro-segmentation at
ingress/egress
• Basic analytics
• Access based on
centralized authentication
• Protections for known
threats with application-
specific protection
• Continuous
validation
• Real-time dynamic
analysis
• Constant monitoring
and validation
• Data access relies on
real-time risk score
• Distributed micro-
segmentation
• ML threat detection
(NDR)
• All traffic encrypted
• Access is authorized
continuously
• Analytics to provide
protections that account
for application behavior
Traditional
Advanced
Optimal
Visibility,Analytics&Automation
8. MOVING TO AN ENDPOINT-CENTRIC ZERO TRUST SECURITY MODEL WITH SENTINELONE
SENTINELONE WHITEPAPER 8
BuildingaZeroTrustSecurity
Model
Transitioning security models can be complex and time-intensive. Every organization has
its own unique requirements, use cases, and existing technologies. These factors can af-
fect the successful transition to a Zero Trust security model. SentinelOne recommends to
map based on the Zero Trust maturity level line of businesses and then define a phased
approach in transitioning the security model. The below is a suggested approach on how to
jump-start the Zero Trust project for your organization.
John Kindervag, former VP & Principal Analyst at Forrester, and creator of the Zero Trust
methodology suggests a five-step deployment guide for Zero Trust:
01. Define Your Protect Surface
Most organizations try to reduce the exposed attack surface as much as they can. But in re-
ality, regardless of the investments, there will always be an open attack surface that attack-
ers will find and exploit. Therefore instead of looking for the attack surface, the question
becomes what is the protected surface which includes critical data, application, assets, and
services (DAAS).
02. Map the Transaction Flows
Most organizations as they transition from networker perimeter-based security to modern
architectures are aware of their network and how to protect it. What changes is the fact that
organizations need analytical insights of DAAS within the network? How are critical data
accessed? How can anomalies be detected?
03. Architect the Environment
There is no such thing as an architectural blueprint that fits all organizations in the world.
This statement remains true as organizations embrace the move to a Zero Trust architecture.
ZTN designs are unique per organization because they are determined by your protected
surface and DAAS. Ideally, you want to bring security controls as close as possible to your
protected surface by defining micro-perimeters and ensuring across all aspects that access
requests are always verified based on the health state of the entity requesting the access.
04. Create the Zero Trust Policy
Determine the Zero Trust policies by answering who, what, when, where, why, and how
should get access to corporate resources and services.
07
9. MOVING TO AN ENDPOINT-CENTRIC ZERO TRUST SECURITY MODEL WITH SENTINELONE
SENTINELONE WHITEPAPER 9
05. Monitor and Maintain the Environment
The final step is about gathering telemetry, leveraging autonomous solutions to perform
analytics and detect anomalous and automatically respond based on the defined zero
trust policies.
ZeroTrustBestPractices
SentinelOne can help organizations successfully adopt a Zero Trust security model for their
entire organization from the digital estate, including workplace, data center, mobile, and
cloud workloads.
Endpoints
Today endpoints, regardless if they are workstations, laptops, mobile devices, or servers,
often have different configurations, patch statuses, operating systems, leading to incon-
sistent approaches to applying security policy. This problem is compounded by the rise of
bring-your-own-endpoint (BYOD) and the loss of visibility from legacy network controls due
to the rise of remote and hybrid working practices.
Adopting Zero Trust for endpoints can assist organizations in reducing this risk by providing
the means to monitor, isolate, secure, control, and remove any endpoint from the network
at any time. When integrated into a Zero Trust ecosystem, endpoints can provide valuable
trust signals when determining whether to grant network access, including the endpoint’s
identity, health, and compliance status.
Endpoint Detection and Response (EDR) solutions provide visibility, detection, and re-
sponse and act as an organization’s primary control point for endpoint security. EDR solu-
tions collect telemetry from endpoints, correlate to detect malicious activity, and facilitate
the response and remediation of threats. When EDR is paired with Endpoint Protection
(EPP) as a preventative control, organizations have a complete understanding of the end-
point attack surface and threat posture.
Integrating endpoint trust signals into a Zero Trust ecosystem can help answer the follow-
ing questions and deny access to applications and resources based on the policy:
• Is the endpoint currently affected by malware?
• Is the endpoint demonstrating aberrant network behavior?
• Is the user accessing from a corporate managed endpoint?
• Is the endpoint accessing from a known location?
08
10. MOVING TO AN ENDPOINT-CENTRIC ZERO TRUST SECURITY MODEL WITH SENTINELONE
SENTINELONE WHITEPAPER 10
While security teams deploy EPP and EDR controls to endpoints they manage, there are
a significant number of endpoints that remain unmanaged or unable to take a manage-
ment agent.
Unmanaged endpoints are more vulnerable to compromise and introduce risk to the envi-
ronment if allowed to access corporate resources. Organizations should strive to isolate un-
managed devices, close the EDR deployment gap by leveraging technologies that can per-
form network discovery, and automatically deploy the EDR agent on unmanaged endpoints.
Workloads
Digital transformation’s innovation pace is enabled principally by nimble cloud workload
technology. However, organizations have traded reduced time to market for environmental
security. Agile development practices that emphasize iteration and speed can overwhelm
security teams who are not prepared to secure workloads as fast as they are created. This
friction between DevOps and SecOps creates bottlenecks and an incentive for develop-
ment teams to circumvent security and governance processes. As a result, there are often
blind spots for security teams tasked with keeping cloud environments secure.
Governance of workloads is often performed just once when the workload is deployed, or
sometimes not at all. And the specific configuration of workloads is inconsistent, with many
instances deployed without critical controls. Confusion often abounds and incorrect as-
sumptions made by DevOps regarding workload security according to each cloud provid-
er’s shared responsibility model. Regardless of the public cloud environment, it’s the orga-
nization’s responsibility to monitor their cloud attack surface, which is just as vulnerable to
compromise as user endpoints.
According to Forrester, “public cloud migrations and other disruptive IT changes have often
acted as a good vehicle for achieving a Zero Trust security model.” A Zero Trust solution for
workloads must provide a repeatable and consistent approach to securing private, public,
hybrid, and multi-cloud environments. It requires an active inventory of all cloud assets,
configuration status and health, and measures to preserve workload immutability. As such,
cloud governance is not a one-time activity but one that happens continuously.
Workload controls can help answer the following questions and adjust container operation
based upon policy:
• Is the workload deviating from the baseline?
• Is the workload affected by malicious activity?
• Is the workload vulnerable to attack?
• Who has access to the workload?
11. MOVING TO AN ENDPOINT-CENTRIC ZERO TRUST SECURITY MODEL WITH SENTINELONE
SENTINELONE WHITEPAPER 11
Using this information, security teams can create Zero Trust access policies using real-time
information about workload’s runtime security, compliance status, and security posture.
Identities
Forrester notes that IAM is one of the least mature areas and one of the top 3 vectors for
external attacks. Compromised credentials and insider threat attacks are a large and often
difficult to mitigate attack surface. With compromised identities, attackers can imperson-
ate employees as well as perform man-in-the-middle attacks to exploit trusted identities
for their advantage. Identity is a critical component of a Zero Trust ecosystem and many
organizations begin their Zero Trust journey by using identity as a lever.
Identity serves as a one-to-many enforcement point for least privilege and identity and ac-
cess management (IAM) remains an effective preventative control point. Identity manage-
ment is complex - tracking employees, customers, partners, and service accounts across
environments, each with varying levels of entitlement and privilege. Zero Trust for identity
governs entitlement and provides least-privilege access policies.
Rather than providing unfettered access, conditional access policies should provide the
least required amount of privilege required to perform a task. Continual authentication for
end-users often adds friction, so it is crucial that Zero Trust implementation automates the
experience as much as possible.
Common implementations of Zero Trust for identity are conditional access, single sign-on
(SSO), and multi-factor authentication (MFA). These technologies should be deployed in
conjunction with a formalized identity governance and entitlement access reviews to en-
sure that users are not over-provisioned privileges.
Networks
Networks have evolved due to the rise of remote work, and our perception of the network
perimeter has evolved as well. Managed networks are no longer contained to a single
location; they exist wherever devices, cloud workloads, and mobile devices access corpo-
rate resources.
Previously, it was considered good practice to mirror network security after physical securi-
ty. Once someone was through the front door, they could move around as needed, whether
that was in the building or on the network. Now, with much of the population still distributed
and working from home becoming a more permanent part of the culture, the boundary of
the workplace moves from the organization’s firewall to the endpoints. This means that the
assumption must always be made that any endpoint is connected via a hostile network and
the operator may not be who they say they are. To operate in this environment, trust must
be requested and granted on a granular scale which often means on a per application basis.
12. MOVING TO AN ENDPOINT-CENTRIC ZERO TRUST SECURITY MODEL WITH SENTINELONE
SENTINELONE WHITEPAPER 12
Zero Trust for network facilitates the proactive hardening of network-accessible resourc-
es and east-west traffic within the network using micro-segmentation. Logical micro-seg-
mentation creates isolated access zones for an application and its associated hosts, peers,
and services. Micro-segmentation furthers Zero Trust by limiting the ability for attackers to
move laterally within the environment. If a given segment of the network is compromised,
micro-segmentation will ensure that the threat actor or ransomware cannot compromise
adjacent resources or services.
In a remote environment, Zero Trust Network Access (ZTNA) helps address the security of
north-south traffic between the internal network and cloud-based internal resources. Zero
Trust Network Access solutions inspect multiple sources of trust signals, from both end-
points and identity providers to ensure that the request is valid before granting access to
SaaS applications and corporate resources. Following Zero Trust principles of verifying ex-
plicitly and assuming breach, the endpoint must prove that it is trustworthy to gain access.
Using ZTNA with endpoints provides the means for a risk-aware network access policy.
Additionally, Network Detection and Response (NDR) solutions provide visibility for detect-
ing and are an effective control point for responding to network-borne threats. NDR solu-
tions log, inspect, and continuously monitor all network traffic for suspicious activity. NDR
solutions can help answer a broader range of questions when responding to an incident or
hunting for threats, such as:
NORTH-SOUTH
NETWORK
TRAFFIC
EAST-WEST NETWORK TRAFFIC
Internal Network
External Network
Data center Data center
13. MOVING TO AN ENDPOINT-CENTRIC ZERO TRUST SECURITY MODEL WITH SENTINELONE
SENTINELONE WHITEPAPER 13
Visibility, Analytics and Automation
Hygiene - Risk - Hardening
ZeroTrustpoweredbySentinelOne
SentinelOne’s approach to Zero Trust provides the means for security teams to continuous-
ly monitor and manage the hygiene, risk, and hardening of their entire estate as part of a
Zero Trust strategy.
• Did another asset begin to behave strangely after communicating with the potentially
compromised asset?
• What service and protocol were used?
• What other assets or accounts may be implicated?
• Has any other asset contacted the same external command-and-control IP address?
• Has the user account been used in unexpected ways on other devices?
08
SentinelOne
forZeroTrust
Endpoint
Monitor, secure and
enforce endpoint
policy
Apply least-privilege
and risk-aware
policy
Inventory and
protect multi-cloud
workloads
Segment and
control network
access
Identity
Workload Network
14. MOVING TO AN ENDPOINT-CENTRIC ZERO TRUST SECURITY MODEL WITH SENTINELONE
SENTINELONE WHITEPAPER 14
Hygiene
Preventing, detecting, responding, and recovering
from cyber threats
In a Zero Trust environment, everything is assumed to be breached, and endpoints and
cloud workloads must prove otherwise. Before granting access to corporate resources, a
Zero Trust solution must first check whether the device requesting access is compromised.
• SentinelOne’s patented on-endpoint Behavioral AI predicts, stops, and corrects
the effects of known and unknown threats in real-time. SentinelOne’s agent can be
deployed across all major operating systems to monitor and continuously assess
endpoint health with or without an internet connection.
• Patented 1-click remediation automates threat resolution with fully autonomous
responses that trigger protective actions in real-time. SentinelOne provides a clear
picture of an endpoint’s health, management status, and the ability to automatically
quarantine or remediate it to bring the device into compliance.
• SentinelOne Singularity Cloud provides runtime protection and EDR for virtual
machines (VMs) and containerized workloads. Organizations can manage and secure
hybrid, private, and multi-cloud workloads from a single console with a single agent.
Workload health status is available in real-time and automatically brought back into
compliance.
• Singularity Mobile brings behavioral AI-driven protection, detection, and response to
iOS, Android, and ChromeOS devices. Part of the Singularity™ Platform, SentinelOne
delivers mobile threat defense that is local, adaptive, and real-time, to thwart mobile
malware and phishing attacks at the device, with or without a cloud connection. And
because it’s mobile, data privacy is built-in at every level. Singularity Mobile works
with or without an MDM, and integrates with all leading MDM solutions. The on-device
agent provides protection and detection of both mobile malware and phishing, known
or unknown, with minimal battery consumption for an optimal end-user experience.
Risk and Governance
Visualizing, managing, and mitigating risk
Making data-driven decisions is critical for security teams. Organizations need to fully un-
derstand possible risks, blind spots and the attack surface before security policies can be
effectively applied.
• With the exponential increase of connected endpoints and the often complex and
varying configurations of cloud workloads, it has become difficult for organizations to
understand who is inside the network and how workloads are configured compared to
industry standards like CIS.
15. MOVING TO AN ENDPOINT-CENTRIC ZERO TRUST SECURITY MODEL WITH SENTINELONE
SENTINELONE WHITEPAPER 15
• To gain visibility into the network, SentinelOne Singularity Ranger turns endpoints
into distributed network sensors that provide monitoring of the enterprise attack
surface in real-time. SentinelOne agents actively fingerprint and inventory all IP-
enabled endpoints on the network to identify abnormal communications and open
vulnerabilities.
• With Ranger, risk from devices that are not secured with SentinelOne can be mitigated
by either automatically deploying an agent or isolating the device from the secured
endpoints. This is how Ranger can be used to effectively reduce the attack surface.
• Singularity Conditional Policy is SentinelOne’s endpoint-centric Conditional Policy
engine. SentinelOne empowers organizations to dynamically change security policies
based on the risk level of the endpoint through this capability. With that, endpoints
are no longer trusted by default but rather are continuously verified. When an incident
occurs, the security policies are dynamically hardened in real-time to reduce the attack
surface and prevent any potential damage.
Hardening
Designing and implementing preventative measures
One of the core principles of Zero Trust is to embrace the least privilege and default-deny
policies until a user can prove they require elevated privileges.
• SentinelOne inventories all locally running applications from across the endpoint
fleet and uses Storyline Active Response (STAR) rules to create a default-deny
policy. A default-deny policy would restrict access to only approved applications and
publishers, allowing the security team to manage by only by exception. This approach
would significantly reduce the risk of compromise from unapproved or potentially
malicious applications.
• SentinelOne’s Device Control suite helps organizations embrace a more hardened
posture for data loss prevention by restricting USB, Bluetooth, and Bluetooth Low
Energy communications. Admins can restrict by endpoint class - for example, USB
mass storage endpoints - which dramatically reduces the potential attack surface for
insider threats and data loss.
• Preserving the unchanging, immutable state of a workload is an essential control for
cloud workload protection. Application Control preserves the immutable nature of the
workload by employing a default-deny posture for any new code not present in the
validated initial VM or container image. Not only does this harden the image itself, but it
prevents attackers from executing arbitrary code that could be used for compromise or
lateral movement. Additionally, cloud application access control enables default-deny
policies for access to cloud workloads and services. Cloud services are denied by default,
reducing the amount of shadow IT and shadow cloud usage. Only approved endpoints
will access the cloud resources and can be managed by exception by the security team.
16. MOVING TO AN ENDPOINT-CENTRIC ZERO TRUST SECURITY MODEL WITH SENTINELONE
SENTINELONE WHITEPAPER 16
Singularity Marketplace
Connecting The Zero Trust Ecosystem
An effective Zero Trust framework integrates best-of-breed solutions and existing infrastruc-
ture to fill security gaps without a forklift upgrade of the security stack. SentinelOne has part-
nered with leading identity and network vendors to deliver validated Zero Trust solutions:
01. Network
SentinelOne’s integration with Guardicore provides centralized visibility of network activ-
ity, including network data generated from endpoints and cloud workloads. SentinelOne
agents report metadata to Guardicore that creates detailed visibility and network to-
pology in the Guardicore for decision-making, forensics, and micro-segmentation pol-
icy creation. Policies can be exported from Guardicore, where they are enforced by
SentinelOne’s native firewall controls. Guardicore can define segmentation and micro-seg-
mentation policies and then use the SentinelOne APIs to enforce them on the agent.
SentinelOne’s integrations with Zscaler and Cloudflare use device signals from Sen-
tinelOne to inform Zero Trust Network Access decisions. Information about the end-
point, including whether it is managed and has a SentinelOne agent installed, is provided
to Zscaler and Cloudflare. This information is combined with contextually relevant infor-
mation from an identity provider to determine a point-in-time network access decision.
SentinelOne integrates with a number of NDR solutions including Vectra AI, Awake Security
(Arista Networks), and Fidelis. The combination of SentinelOne’s EDR with partner NDR
Step 1:
User requests access
to an application.
Identity Management
Contextual Factors
Device Security Posture
Step 2:
A ZTNA provider directs user to an authentication
page where SentinelOne evaluates device posture
for policy adherence. Additionally, user’s identity
and context such as geolocation and certificate are
also assessed.
User Apps
Step 3:
If user’s request
adheres to policy, access
is granted.
Self-Hosted
SaaS
Zero Trust Network
Access Provider
17. MOVING TO AN ENDPOINT-CENTRIC ZERO TRUST SECURITY MODEL WITH SENTINELONE
SENTINELONE WHITEPAPER 17
EnforceAccessControl
Requirements
Changerequirements uppon
ofdetection of risky eventsor
activities
capabilities provides visibility, detection, and response for both managed and unmanaged
endpoints. While NDR connects related network activity into a broader attack map, Sen-
tinelOne provides contextual awareness by enriching information coming from managed
endpoints such as device name, last logged-in user, operating system details, and other
endpoint characteristics. This provides comprehensive threat detection, rapid and effective
response, endpoint containment, and forensic analysis capabilities.
02. Identity
SentinelOne integrates with Azure Active Directory to provide identity-focused Zero
Trust solutions. Conditional access is a key part of Zero Trust because it helps to en-
sure the right user has the right access to the right resources. Enabling Conditional Ac-
cess allows Azure Active Directory to make access decisions based on computed risk
and preconfigured policies. When an endpoint is compromised SentinelOne pushes
this information in real-time to Azure Active Directory ensuring that the organization
can leverage their conditional policy to block a user, limit user’s access, or trigger MFA.
Additionally, SentinelOne can share identity risk information with Azure AD that is factored
into conditional policy. For example, if SentinelOne detects an attack on an endpoint, it will
provide the last logged-in user to Azure AD with a high user risk level. Using this informa-
tion, Azure AD will enforce identity policy, such as resetting the password or blocking a user.
Finally, the risk score that Azure AD generates for a given identity can be used in Sen-
tinelOne to inform triage and investigation. Identity risk information is surfaced within the
SentinelOne console and provides analysts with at-a-glance context about a given identity.
• Phishing Attack
• Compromised Credential
• Lateral Movement
• Abnormal Activity
• Threat Detection
Type of
Devices
Modify user risk
Conditional Access Policy
Enforcement
Azure AD
User Related
Activity
Location Application
Device
Trust Level
Limit
Access
Require
MFA
Block
Access
18. MOVING TO AN ENDPOINT-CENTRIC ZERO TRUST SECURITY MODEL WITH SENTINELONE
SENTINELONE WHITEPAPER 18
SentinelOne natively integrates with Okta to bring identity context and response actions
directly within the Singularity XDR platform. SentinelOne consumes logs and contextual
events from Okta and displays them alongside relevant endpoint detections. With identity
visibility, analysts can see additional identity activity that may be relevant for an investiga-
tion. When corrective mitigation needs to take place, response actions within SentinelOne
can revoke an identity or trigger MFA.
Conclusion
SentinelOne’s approach helps organizations advance Zero Trust maturity by leveraging ex-
isting endpoint, cloud workload, identity, and network security investments. With native ca-
pabilities and integrations, organizations can begin the journey to more effectively applying
Zero Trust principles.
10
To learn more about
Singularity Marketplace
ecosystem partners, visit
To get started on your
Zero Trust journey today,
request a demo from a
SentinelOne expert
REQUEST A DEMO MARKETPLACE