Cloud Assessment and Authorisation (January 2024)
Cloud Assessment and Authorisation (January 2024)
Introduction
Cloud computing offers a range of potential cybersecurity benefits for cloud consumers to leverage, providing access
to advanced security technologies, shared responsibilities, fine-grained access management, comprehensive
monitoring and highly redundant geographically dispersed cloud services. For many organisations, cloud computing
can provide significant improvements to their cybersecurity, mitigating the risk of many current cyberthreats.
While cloud computing can significantly enhance an organisation’s cybersecurity, it also presents other risks that need
to be considered, such as multi-tenancy architectures, reduction in visibility of the physical and virtualisation layers,
and possible foreign interference.
At its core, cloud computing involves outsourcing a part, or all, of a consumer’s information technology (IT) capability
to a Cloud Service Provider (CSP). This outsourcing brings a reduction in control and oversight of the technology stack,
as the CSP dictates both the technology and operational procedures available to the cloud consumers using its cloud
services.
Cloud computing, by default, does not provide improved cybersecurity without effort on behalf of the cloud consumer
to perform their security responsibilities in securing the cloud. If not properly managed, maintained and configured, it
can increase the risk of a cybersecurity incident occurring. Cloud consumers need to consider the benefits and risks of
cloud computing, including their own responsibilities for securing the cloud and determining whether cloud
computing meets their security needs and risk tolerance.
One of the biggest barriers to cloud consumers adopting cloud computing is the difficulty identifying and
understanding the risks of using a CSP and its cloud services. Cloud computing presents a uniquely complex and
layered technology stack that is rapidly evolving and resists traditional point-in-time assessments. This publication
guides CSPs, cloud consumers and IRAP assessors on how to perform a comprehensive assessment of a CSP and its
cloud services so that a risk-informed decision can be made about its suitability to store, process and communicate
data.
The assessment and authorisation process detailed in this publication uses the security requirements and cloud
guidance detailed in the Department of Home Affairs’ Protective Security Policy Framework (PSPF) and the Information
security manual (ISM). These documents provide the requirements and controls for cloud consumers to use in the
assessment of the CSP, its cloud services and a cloud consumer’s own systems.
The terminology and definitions used in this publication for cloud computing are consistent with the National Institute
of Standards and Technology’s (NIST) Special Publication (SP) 800-145, The NIST Definition of Cloud Computing.
1
Audience
This publication is intended for CSPs, Infosec Registered Assessors Program (IRAP) assessors and Non-Corporate
Commonwealth Entities (NCCEs, referred to as cloud consumers in this publication) who are subject to the Public
Governance, Performance and Accountability Act 2013 to the extent consistent with legislation.
This publication assists and guides IRAP assessors, cloud consumer’s cybersecurity practitioners, cloud architects and
business representatives on how to perform an assessment of a CSP and its cloud services, and the cloud consumer’s
own self-developed systems hosted in the cloud.
While this publication is primarily intended for cloud consumers, this guidance can be used by any organisation
considering cloud computing.
An organisation who owns and manages its own IT infrastructure is responsible for securing all aspects of it, including
achieving the desired security baseline, maintaining it and updating it as malicious actor tradecraft evolves; depending
on the size of the environment, this may necessitate significant effort and resources on behalf of the organisation to
achieve this.
As part of an organisation’s examination of cloud computing, it needs to consider its own capabilities to secure their
systems and protect their data from current and future cyberthreats. If an organisation’s basic security practices such
as patching, upgrading and system hardening are ineffective or inconsistent, cloud computing may provide significant
security improvements. By transferring some security responsibilities to the CSP, an organisation can prioritise other,
more specific security mitigations such as access control, authentication and monitoring. In addition to transferring
some security responsibilities, leveraging the advanced security technologies available from many CSPs can provide
substantial cybersecurity improvements beyond what is feasible when an organisation owns and manages its own IT
infrastructure.
While cloud computing can improve an organisation’s cybersecurity, this can only be achieved by understanding the
shared responsibility between the cloud consumer and the CSP, and which party is responsible for securing which
parts of the cloud. The cloud consumer also needs to plan, design and configure the cloud services it is using to
achieve its desired security baseline. The cloud consumer then needs to monitor the CSP, its cloud services and its
own systems and respond to any changes to the security baseline that are outside the cloud consumer’s risk
tolerances.
Principles
This publication, in conjunction with the ISM, Cloud security assessment report template and Cloud controls matrix
(CCM), is designed to assist cloud consumers to identify the risks associated with a CSP and its cloud services, and
make a risk-informed decision about using cloud computing.
To support this self-authorisation and risk-based decision by cloud consumers is the independent assessment of a CSP
and its in-scope cloud services by an IRAP assessor. This assessment forms the basis of the review by cloud consumers
of a CSP and its cloud services. This assessment is performed against the controls in the ISM and other related
guidance from the Australian Signals Directorate (ASD), such as this publication. The IRAP assessor will document their
One of the key principles of the assessment is the separation of the CSP and its cloud services. This emphasises the
importance of assessing and reviewing the security fundamentals of the CSP itself. Cloud consumers need to ensure
the CSP itself is operating securely and meets the cloud consumer’s security and risk requirements.
Adopting cloud computing requires a degree of trust given to the CSP to handle a cloud consumer’s data. The cloud
consumer needs to gain enough assurance of the CSP so that this trust is not unwarranted. The other part of the
assessment focuses on the specific cloud services of the CSP that are in scope of the assessment.
Each cloud service will be documented individually in the cloud security assessment report. This is so cloud consumers
can review a cloud service without needing to read about other, unrelated cloud services to obtain the required
information to make a risk-based decision about using a particular cloud service.
To help cloud consumers maintain awareness of the risks of using a CSP and its cloud services, CSPs are responsible for
maintaining the accuracy and currency of the report between independent assessments by adding addendums to the
cloud security assessment report. These addendums are to detail any changes that have occurred to the CSP or its
cloud services that result in any part of the original report becoming inaccurate. New addendums are to be
communicated to any cloud consumer who has used the cloud security assessment report as part of their assessment
of the CSP. This supports cloud consumers to maintain continual awareness of the CSP and its cloud services and
respond to any changes that impact the risks and the security baseline of the systems.
Personnel security
As part of the cloud consumer’s review of a CSP to determine if it meets its security requirements and risk tolerance,
the cloud consumer needs to consider the classification of the data it intends to use on the CSP’s cloud services, as
well as the suitability of the CSP to store, process and transmit this data. Understanding the classification of the data
to be used on a CSP’s cloud services will assist in identifying the personnel security requirements the CSP needs to
adhere to.
To understand the personnel security risks to a cloud consumer’s data, cloud consumers need to assess the CSP’s
personnel pre-employment screening and the CSP’s controls that prevent and detect its personnel accessing customer
data without proper authorisation.
To manage, support and update their cloud services, CSPs require privileged access to perform these responsibilities.
This is a risk that is common with most outsourcing arrangements, whereby an external party has some degree of
privileged access to systems that handle customer data. This risk needs to be carefully considered and accepted by
cloud consumers before any of their data is transferred to the CSP’s cloud services.
separation of duties, such as personnel with physical access to IT infrastructure not having logical access and vice
versa
secure storage and customer supplied and/or management of encryption keys for customer data
just in time and just enough access methodologies for its personnel’s access
real-time monitoring to detect and log when CSP personnel access customers’ data, and the ability to quickly
terminate any access that is unauthorised
providing the cloud consumer with the capability to provide explicit approval before the CSP’s personnel access
its data
providing cloud consumers with flexible support arrangements including the ability to choose where support is
provided from
contractual clauses with customers that require the CSP to disclose to the cloud consumer any incidents of its
personnel accessing, or encountering, the cloud consumer’s unencrypted data.
While the risk of a CSP’s personnel accessing or encountering its customers’ data cannot be eliminated, it can be
significantly reduced with the implementation of the above controls.
The absence of effective controls to prevent and detect a CSP’s personnel accessing customer data also increases the
risk that classified data will be accessed by personnel without an appropriate need-to-know, including those CSPs
whose personnel hold an Australian Government security clearance. These CSPs will also be unlikely to be able to
inform its cloud consumers of any incidents of unauthorised access to their data within a reasonable time period. The
implementation and effectiveness of these controls needs to be carefully considered as part of a cloud consumer’s
review of a CSP.
CSPs who store, process and communicate data marked up to OFFICIAL: Sensitive, are not required to have personnel
with Australian Government security clearances to handle this classification of data. Cloud consumers only using data
marked up to OFFICIAL: Sensitive need to ensure the CSP’s personnel pre-employment screening aligns to, or meets
the intent of the pre-employment screening requirements detailed in the PSPF.
CSPs who store, process or communicate data classified at PROTECTED and above are required to have personnel who
hold a current Australian Government security clearance commensurate with the classification of the data being
stored, processed and communicated on the CSP’s systems. This requirement applies to any of the CSP’s personnel
who have physical access to infrastructure that handles classified data, and to personnel with logical access (including
potential logical access) to infrastructure that handles classified data.
Personnel security clearances, like technical controls, provide a degree of assurance that the personnel are suitable to
have access to classified data. Personnel security clearances do not provide a guarantee against maleficence on behalf
of the personnel who hold a security clearance, and this needs to be compensated with additional effective controls
such as those listed above. The combination of both personnel security clearances and effective controls provide the
most assurance that a cloud consumer’s data will be protected against illegitimate access on behalf of the CSP.
The following minimum protections are required to safeguard data accessed on a temporary basis:
Australian Government entities must limit the duration of access to classified data as follows:
Australian Government entities must supervise all temporary access. Examples include:
management oversight of the work of personnel who have the temporary access
monitoring or audit logging incidents of contact with classified data (e.g. contract conditions that require
service providers to report when any of their contractors have had contact with classified data).
NCCEs are required to conduct a risk assessment to determine whether to allow temporary access to classified data.
the need for temporary access, including if the role can be performed by a person who already holds the
necessary clearance
confirmation from the authorised vetting agency that the person has no identified security concerns, or a
clearance that has been cancelled or denied
the quantum and classification level of data that could be accessed, and the potential business impact if this
data was compromised
how access to classified data will be supervised, including how access to caveat or sensitive compartmented
information will be prevented
other risk mitigating factors such as pre-engagement screening, entity specific character checks, knowledge of
personal history, or having an existing or previous security clearance.
Where an entity intends to grant temporary access to classified data from another entity or third party, the
Department of Home Affairs recommends consulting the other entity or party, where appropriate, and obtaining
agreement for temporary access to their classified data.
The Department of Home Affairs considers there is merit in obtaining an undertaking (e.g. through a confidentiality or
non-disclosure agreement) from the person to protect official data.
Personnel security is just one consideration in the overall assessment and review of a CSP and its cloud services.
Personnel security needs to be reviewed by cloud consumers in the context of all other aspects of CSP and its cloud
services detailed in this publication and the cloud security assessment report.
Physical security
Cloud consumers are responsible for ensuring the physical facilities that contain their data, or are used to access their
data, including those owned by third-parties such as CSPs, meet the PSPF physical security requirements. These third-
party facilities are most commonly the CSP’s data halls within a data centre, other points of presence, and the CSP’s
administrative and support locations from which a cloud consumer’s data can be accessed. Cloud consumers need to
ensure these facilities meet the zone requirements defined in the PSPF that are commensurate with the Business
Impact Level (BIL) of their data if it was compromised, lost or damaged.
To ascertain if a CSP’s physical facilities that handle a cloud consumer’s data meet the necessary physical security
requirements, these facilities can be evaluated by a Security Construction and Equipment Committee (SCEC) Endorsed
Security Zone Consultant, or by an NCCE’s Agency Security Adviser (ASA).
A SCEC Endorsed Zone Consultant or ASA can be engaged to evaluate the relevant CSP’s facilities and document all
compliances and non-compliances with the PSPF and the Australian Security Intelligence Organisation (ASIO) T4
Technical Notes. The SCEC Endorsed Security Zone Consultant or ASA produces a report that benchmarks the CSP’s
facilities against the PSPF and Technical Notes requirements, and details their findings and recommendations.
Cloud consumers whose ASA performs this assessment are recommended to share their report with other cloud
consumers who are considering using the CSP. This will prevent other cloud consumers or a SCEC Endorsed
Consultants conducting a duplicative assessment.
Before undertaking an assessment of a CSP’s facilities, the CSP needs to map the network and systems in its facilities
to identify where its customers’ unencrypted data is stored, processed and communicated. Vulnerable sensitive areas
can occur anywhere that data is stored, processed or transmitted in its unencrypted state within the CSP’s
infrastructure; whether in a data hall, a meet-me room or a computer operations room.
In addition, where cryptographic operations are performed and key material is stored is an important consideration
when determining which vulnerable sensitive areas need to be secured to meet the PSPF security zone requirements.
If the unencrypted data passes outside the security zone, either of the following two options should be implemented:
Expand the security zone perimeter to include all rooms in which the unencrypted data is stored or
transmitted. While this option is simpler to implement, it may pose physical security access restrictions
(including security clearance requirements) on some areas, which the CSP may not be able to practicably
manage through their business model; or
Encrypt the data each time it is stored or is transmitted outside the security zone. This option may have
significant implications for network infrastructure, including requirements for additional cryptographic ICT
hardware and secure rack protection to protect the classified ICT hardware.
A cloud consumer’s chief security officer (CSO) or delegated security adviser must, before consuming a CSP’s cloud
services, certify the CSP’s facilities (such as the data halls, administrative and support locations, and points of
presence) in accordance with the PSPF and ASIO Technical Notes.
IRAP assessors are required to document in the cloud security assessment report, the physical facilities where a CSP
stores, processes, transmits and accesses cloud consumer classified data, and if these facilities have been evaluated by
a SCEC Assessor and have an accompanying report and letter. IRAP assessors are also required to address the relevant
ISM controls for physical security in their assessment.
For further information on classified data, including BILs, refer to the PSPF.
Cloud consumers also need to consider where the CSP’s administration and support is provided from. Depending on
the locations, this can impact personnel pre-employment screening practices, as different countries have different
laws about the degree of data that employers can request from their employees.
These aspects of the CSP need to be assessed by cloud consumers considering using its cloud services, to identify any
potential risks this presents to the cloud consumers’ systems and data.
Cloud consumers, as part of their review of a CSP, need to consider the following:
the locality of the CSP’s offices, datacentres and administrative and support personnel
the potential for any extrajudicial control and interference over a CSP by a foreign entity.
ASD recommends cloud consumers use CSPs and cloud services located in Australia for handling their classified data.
CSPs that are owned, based and solely operated in Australia are more likely to align to Australian standards and legal
obligations, and this reduces the risk of any data type being transmitted outside of Australia. These CSPs are also less
susceptible to extrajudicial control and interference by a foreign entity.
Foreign-owned CSPs, including those located in Australia, present additional risks that need to be considered as part
of the overall risk posture. This includes foreign ownership, foreign interference and extrajudicial control over the
CSP’s operations and data holdings. Foreign interference may occur where their own laws, policies or powers allow
access to, or control over, the CSP’s operations and data. ASD considers that the involvement of CSPs who are likely to
be subject to extrajudicial directions from a foreign government that conflict with Australian law, may risk failure to
adequately protect Australian Government data from unauthorised access or interference.
While the locality, ownership and potential for foreign interference are crucial elements to be considered as part of a
cloud consumer’s review of a CSP, other elements of the CSP also need to be considered to ascertain a complete
understanding of its suitability to handle a cloud consumer’s data. For example, a CSP may be at low risk of foreign
interference, but may lack controls in other areas, possibly posing a greater risk than a provider who is not solely
Cloud consumers need to consider all aspects of a CSP to make an informed decision about its use and not rely on a
single factor to determine a CSP’s suitability.
IRAP assessors as part of their assessment of a CSP are to document in the cloud security assessment report, the
ownership of the CSP, the locality of its offices (including support and administrative locations), data centres and the
locations its cloud services are provided from.
For further information on securing the cyber supply chain, refer to the Cyber supply chain risk management and
Identifying cyber supply chain risks publications.
The ISM is updated on a regular basis to provide up-to-date guidance to mitigate the latest malicious actor tactics,
techniques and procedures as observed by ASD. CSPs and IRAP assessors should review the ISM changes and adjust
their assessments accordingly to include the latest guidance, or alternatively, make a reference in the assessment
report where the updated guidance would have altered the assessment.
The CCM provides indicative guidance on the scoping of cloud assessments, and inheritance for systems under a
shared responsibility model. The guidance is not definitive, and needs to be interpreted by the IRAP assessor in the
context of the assessed system.
Importantly, the CCM also captures the ability of cloud consumers to implement ISM controls for systems built on the
CSP’s services, identifying where the cloud consumer is responsible for configuring the cloud service in accordance
with the ISM. Information regarding the maintenance of the CCM is addressed within the ‘Maintaining the CSP
security fundamentals and cloud services report’ section below.
The Cloud security assessment report template can be customised as needed to best document the findings from the
assessment of a CSP and its cloud services. IRAP assessors should, however, limit the changes to the report to only
what is necessary, maintaining its structure and headings to ensure reports are consistent.
IRAP assessors may, however, use the evidence from other assessments, provided the evidence is applicable, accurate
and valid. Given control alignment with other standards is rarely perfect, IRAP assessors should not rely on compliance
statements from other standards, but should instead review the supporting evidence and determine whether a
control is effective or not.
When reusing evidence from existing certifications or previous assessments, attention must be given to the scope of
the certification or assessment. For example, an existing cloud service certification may only be applicable to instances
of that cloud service in certain data centres and within certain regions. Similarly, cloud services may consist of a range
of integrated cloud service products and a previous assessment of a cloud service may only be applicable to specific
products in a particular configuration.
As part of the cloud security assessment report, IRAP assessors are to document which party is responsible for
securing key aspects of each cloud service in scope of the assessment. This provides cloud consumers with a clear
understanding of the different responsibilities each party has for securing the cloud service, including their own.
Regardless of the shared responsibility model, cloud consumers remain accountable for their own data. This includes
ensuring the data is appropriately secured, as well as any compromises, losses or damages that occur to the data
while using cloud computing.
Third-party cloud solutions can increase the assessment difficulty and complexity for IRAP assessors, and make it
harder for cloud consumers to identify the risks of its use. This is primarily due to the addition of an extra party who is
involved in providing the complete cloud solution. IRAP assessors need to consider both the third-party and its
security, plus that of the underlying CSP. Third-parties do not automatically inherit all controls made available to them
by the CSP, as third-parties may or may not implement certain controls for their business and operational
requirements. Third-parties also rely on their own supply chain, personnel security and secure administration
practices to provide their cloud solution. These factors need to be considered by cloud consumers as part of their
overall risk assessment.
Similarly, as part of their review of a third-party cloud solution, cloud consumers need to consider the security of the
underlying CSP, its cloud services and how it is configured. Ideally, the underlying CSP has been previously assessed by
an IRAP assessor and can provide a report. This will allow cloud consumers to better identify the risks of the complete
cloud solution, including the underlying CSP and its cloud services.
If an IRAP assessment of the underlying CSP and its cloud services has not been performed, this may make it difficult
for cloud consumers to understand the risks of using a third-party cloud solution. Cloud consumers can consider other
international standards and certificates as an indication of a CSPs security practices and posture. However, these
standards are not a substitute for an IRAP assessment against the ISM.
Customer data: This is data the cloud consumer creates, generates or uploads to the CSP for storing, processing
and sharing using the CSP’s cloud services, this includes the cloud consumer’s authentication details. The cloud
Account data: This is data about the cloud consumer’s account with the CSP and can include billing, contact and
usage details.
Metadata: This includes data about the cloud consumers’ use of the CSP’s cloud services and can include cloud
consumer generated data such as resource names, service tag details and utilisation details.
Support and administrator data: This data type is provided to the CSP’s support personnel and administrators
for technical support purposes. This can include logs, monitoring alerts and error reporting details.
The above definitions are to be used as a guide only. Each CSP will likely have their own data type definitions, as well
as other data types not covered in this publication. Cloud consumers can refer to the CSP’s cloud security assessment
report to identify the data types used by the CSP.
Each CSP’s handling of data often differs per data type, for example, a globally distributed CSP may retain customer
data in Australia, but might transmit account data to another country for processing and storing.
It is also common for CSPs to handle the names of customers’ virtual machines, networks and accounts not as
customer data, but as metadata. This is often used for analytical purposes and can subsequently be stored in a
different location with different controls. In this example, it is advisable for cloud consumers to not place any
classified data into fields not treated as customer data due to the risks related to the different handling of the data.
As part of the CSP security fundamentals and cloud services assessment, IRAP assessors need to document the
different data types, their definitions, where they are stored, and how they are handled and secured by the CSP.
Cloud environments can be complex, dynamic, large and unique. These attributes can make it challenging to assess a
CSP’s security fundamentals and cloud services to determine if it secure and suitable for handling a cloud consumer’s
data. Each assessment is unique and needs to be tailored to the CSP’s particular operating environment and bespoke
cloud services.
This publication guides IRAP assessors through an assessment of a CSP and its cloud services to determine its security
and residual risks, and to document these findings in the Cloud security assessment report template so that cloud
consumers can review and determine if the CSP meets their security requirements and risk tolerances.
Assessment frameworks
While assessments consistent with this publication should always consider the CSP’s operation against the ISM
controls, IRAP assessors may also choose to draw in other control frameworks. This provides an opportunity to draw
from prior work by the CSP and to supplement any areas where additional, or more comprehensive or insightful
coverage can be provided by frameworks other than the ISM.
Level of standardisation: Many ICT environments are centrally managed. For example, if checking the validity of
configuration on servers which are configured using one technical policy, then potentially one server can be
representative of all systems. Also, CSPs often have cloud services that support other cloud services. The built-in
security (security that cannot be altered by cloud consumers) of these supportive cloud services can provide a
security baseline representative of many cloud services.
Truly representative: IRAP assessors need to ensure that any points they sample are truly representative, and
not a contrived example created only for assessment purposes.
Different management zones/arrangements: Systems may be operated and administered in different security
zones, for different purposes and under different management arrangements. IRAP assessors should consider
these differences, and determine if they need to sample data points from across these zones.
Ease of data collection: IRAP assessors should plan to use and leverage tools as part of their assessment. By
driving down the cost of each individual sample, this allows the assessor to more comprehensively gather
evidence which will lead to a more accurate assessment.
Confirmation of unexpected results: IRAP assessors may find they come across results which are inconsistent
with their professional experience, such as a CSP demonstrating significant over or underperformance against
assessment criteria relative to other similar CSPs. In these situations, IRAP assessors should determine how to
take an additional sample/s to confirm the unexpected result.
Scoping principles
Scoping of the cloud security assessment is an important step that helps identify expected authorisation boundaries
and the limit of significant dependencies and responsibilities. It also creates abstraction layers that allow systems to
be individually identified and described without necessarily having to describe all other related components
simultaneously.
Experience with cloud assessments indicates that it is difficult to correctly determine scope at the outset. While IRAP
assessors and CSPs should determine an initial scope, this scope should be regularly reviewed to ensure it remains
representative of the assessment. In particular, if significant dependencies, potential weaknesses or significant
sources of risk are identified which are beyond the initial scope, then the scope should be adjusted.
In addition to any specific agreement between the CSP and the IRAP assessor, the following general scoping principles
should be applied:
the CSP’s corporate network may be in scope depending on secure administration practices and segmentation
and segregation between the corporate network and the CSP’s cloud infrastructure.
In general terms, the evidence of the effectiveness of controls varies from weak evidence, such as a claim that a
control exists (e.g. a policy statement), through to strong evidence, such as evidence that a policy is routinely followed
or a simulated test which verifies that a technical control performs as expected. More specific guidance on evidence is
provided below.
While broad control coverage is important, in performing assessments IRAP assessors should give considerable weight
to the quality of evidence about control effectiveness presented to them, or made available to them on request.
Gathering evidence can be time consuming, and IRAP assessors may need to decide on a case by case basis at what
point they have sufficient evidence to consider a control effective. IRAP assessors should also consider what evidence
can be collected efficiently. For example, verifying that a technical configuration is in place by reviewing the
configuration is both a higher standard of evidence, and likely faster and more efficient, than reviewing
documentation to try and identify the same thing.
Depending on the size of the assessment, IRAP assessors may need to ensure their assessment team has sufficient
skills to efficiently collect, understand and interpret the evidence they will need to review as part of the assessment.
See the Expertise Principle below for further information.
Evidence quality
IRAP assessors will find that not all evidence types are covered by these examples, but can still use them as a guide for
determining the quality of evidence suitable to assess controls:
Poor evidence: A policy statement (e.g. repeating the ISM control in an internal document, irrespective of the
amount of boilerplate included).
Fair evidence: Reviewing a copy of the relevant system’s configuration to determine if it should enforce the
expected policy.
Good evidence: Reviewing the technical configuration on the system (through the systems’ interface) to
determine if it should enforce the expected policy.
Excellent evidence: Testing the control with a simulated activity designed to confirm it is in place and effective
(e.g. attempting to run an application to check for application control, or attempting to access an external
website using a privileged account).
Acknowledging the inheritance of controls can be an efficient strategy but care needs to be taken to ensure that later
assessments do not try and claim controls which do not provide coverage against the full operating context of the
later assessment.
All aspects of a CSP, its cloud platform, and any other environments that it is responsible for that interconnect with
the cloud platform, are in scope at the commencement of an assessment. Any environments that are deemed out of
scope of the assessment are documented in the report and accompanied by a justification by the IRAP assessor for
why it has been excluded from the assessment.
The IRAP assessor as part of their assessment needs to determine if the CSP has implemented or met the intent of the
Secure Administration and Implementing Network Segmentation and Segregation publications, as well as the
associated controls in the ISM. If the IRAP assessor determines that the CSP has not sufficiently mitigated the risk of
an attacker pivoting from its corporate environment to its cloud infrastructure, the corporate environment is to be
included in the scope of the assessment.
Depending on the scope of the assessment, it may require more than one individual to adequately perform the
assessment. In these cases, IRAP assessors should ensure they are supported by a sufficiently diverse and skilled team
to assist with the assessment. CSPs and cloud consumers should enquire as to the cybersecurity and technical depth
of the individual or team that will perform an assessment in line with this publication to ensure they have the relevant
expertise to perform the assessment.
confirm the intended classification of the data to be handled by the CSP and its cloud services
confirm the purpose of the assessment, i.e. to be assessed for the purposes of identifying the CSP’s
implementation or alignment to the controls defined in the ISM and other relevant Australian Government
security policies
gain an understanding of the CSP, its cloud services and any third-parties that are in scope of the assessment – if
any third-parties are involved in providing the cloud solution, such as a SaaS provider using another CSP’s IaaS,
has the third-party been assessed by an IRAP assessor, and if not do they need to be
identify the sources of data, locations, and evidence required to complete the assessment
identify assessment methods and how data and evidence will be collected and verified
identify and document the shared security responsibility model for each cloud service in scope
identify the ISM controls that are in scope of the assessment using the Cloud controls matrix
identify which party (the CSP, cloud consumer, or any third-parties) are responsible for implementing and
maintaining the effectiveness of each ISM control
obtain evidence of the implementation of ISM controls and their effectiveness – if the CSP has implemented an
alternative control, also include how it meets the intent of the control in the ISM
document any non-implemented or ineffective ISM controls and how the absence of these controls is being risk
mitigated by the CSP
make recommendations to mitigate the risk of absent ISM controls that have not been risk mitigated by the CSP
document the assessment using the Cloud security assessment report template and the Cloud controls matrix.
The objective of the CSP’s security fundamentals assessment is to assess and document the security practices and
posture of the CSP itself. This is so cloud consumers can determine if the CSP itself is operating securely and producing
secure cloud services, and is suitable for handling the cloud consumer’s data.
The other part of the Phase 1a assessment focusses on the CSP’s cloud services that are in scope of the assessment.
These cloud services are assessed against the applicable ISM controls, so cloud consumers can determine the security
risks of using the CSP’s cloud services.
The IRAP assessor will document the findings, evidence and remediation actions in the Cloud security assessment
report template. This report will then be provided to the CSP for sharing with cloud consumers that are considering
consuming its cloud services.
CSP fundamentals
governance
personnel security
data protections
vulnerability management
support model
physical security
system hardening
secure administration
cloud infrastructure
physical security
network security
decommissioning hardware
data transfers
process automation
description
data portability
security baseline
Between assessments, CSPs are to keep their customers informed of changes to their security fundamentals and cloud
services that impact their security baseline and that of its customers’ systems. This continual disclosure to the CSP’s
customers is primarily to be achieved by updating the cloud security assessment report with addendums as soon as is
reasonable to do so. Any new addendums added to the report are to be communicated with the CSP’s customers who
have used the report as part of their assessment of the CSP.
For those aspects of the CSP and its cloud services where there has been no change, or only insignificant changes that
have not impacted the CSP’s security baseline, the evidence used in previous assessments can be reused to validate
controls. However, IRAP assessors are to consider the age of the evidence being supplied and determine if the
evidence is still valid and accurate.
Although CSPs are to be assessed at least once every 24 months, assessment reports that are older than 24 months
are not automatically invalidated by this timeframe. As reports age, the likelihood of them being inaccurate and
invalid increases. They may, however, still be relevant depending on the CSP’s change cadence. Cloud consumers need
to check with the CSP on the validity and currency of the report if it has not been recently completed. Due to the rapid
nature of change in cloud computing, it is likely a CSP’s systems and cloud services have changed, even within a short
period of time. CSPs can support cloud consumers by updating their reports with addendums, document and detailing
any changes to their systems and cloud services that have occurred post the original report.
changes to security policies relating to the CSP, its cloud platform or its clouds services
detection of new or emerging cyberthreats to the CSP, its cloud platform or its cloud services
the discovery that controls for the CSP, its cloud platform or any of its cloud services are not as effective as
intended
major architectural changes to the CSP, its cloud platform or its cloud services.
the cloud service or services were not in scope of the CSP’s cloud security assessment report
the CSP has released a new cloud service or services post the completion of its cloud security assessment report
the CSP has made significant changes to a cloud service or services that impacts the security documented in the
cloud security assessment report.
Phase 1b cloud service assessments are performed by an IRAP assessor. These assessments can be performed
independently of the Phase 1a assessments and between reassessments, alleviating the need for cloud consumers to
wait for a Phase 1a reassessment to occur, and allowing them to quickly use a CSP’s cloud services. A Phase 1a
assessment is required to be completed before a Phase 1b assessment can occur, as cloud consumers need to review
a Phase 1b report (referred to as a supplementary, new and updated cloud services report) in combination with the
Phase 1a cloud security assessment report.
These assessments are documented in the cloud security assessment report, omitting the Introduction and CSP
Security Fundamentals Assessment sections from the report.
The Phase 2a assessment provides additional guidance for assessing and authorising the cloud consumer’s own cloud
systems. This assessment is performed by an IRAP assessor. Phase 2 concludes with the review by the cloud
consumer’s authorising officer or their delegate of the Cloud Authorisation Package, which includes the CSP’s cloud
security assessment report, and if applicable, any supplementary, new or updated cloud services report, and the cloud
consumer cloud systems report. The authorising officer provides the authority to operate based on the acceptance of
security risks associated with the operation of the entire cloud solution.
The guidance in this publication, the Cloud controls matrix, the ISM and other relevant guidance from ASD should be
used in the assessment of a cloud consumer’s own systems. As part of this assessment, cloud consumers need to
identify which controls from the CSP and its cloud services have, and have not been inherited. Where controls have
not been inherited, cloud consumers need to consider the risk, and determine if any compensating controls are
required. This will enable cloud consumers to select the controls that are applicable and need to be implemented for
their cloud-based systems as part of the risk management framework detailed in this publication. These controls can
The assessment of cloud consumer systems needs to be conducted during all phases of the system development
lifecycle, ensuring any security issues are identified and remedied early on. It is common for cloud systems to be
continuously developed throughout their lifecycle, rather than remaining static. As such, point-in-time assessments
are of limited value, and an iterative approach to assessing and validating these systems is necessary. Agile and
DevSecOps practices are encouraged for the development and security of cloud consumer systems to perform this
iterative approach to development and security.
The shared responsibility model for each cloud service should also be reviewed and updated as necessary. Depending
on the configuration of a cloud consumer’s systems, this could alter the original shared responsibility model
documented in the cloud security assessment report, resulting in responsibilities transferring between the CSP and
the cloud consumer.
The Phase 2a findings are documented in the Cloud Authorisation Package and reviewed by the cloud consumer’s
authorising officer.
The authorising officer or their delegate reviews the Cloud Authorisation Package and determines if the cloud
environment, comprising the CSP, its cloud services and the cloud consumer’s systems meet their security
requirements and does not exceed their risk tolerances. In some cases, the security risks associated with a system’s
operation will be acceptable and it will be granted authorisation to operate; however, in other cases the security risks
associated with operation of a system may be unacceptable. In such cases, the authorising officer may request further
work, and potentially another security assessment to be performed. In the intervening time, the authorising officer
may choose to grant authorisation to operate but with constraints placed on the system’s use.
Finally, if the authorising officer deems the security risks to be unacceptable regardless of any potential constraints on
the system’s use, they may deny authorisation to operate until such time that sufficient remediation actions, if
possible, have been completed to an acceptable standard.
Activities to support continuous monitoring of the threat environment, security risks and controls associated with a
system, will differ from system to system, but it is important to consider the dependencies of the system, including
software and other cloud services.
From time to time, ASD will release publications and advisories to assist CSPs and cloud consumers with identifying
and mitigating security risks. While there is no requirement for CSPs and cloud consumers to be immediately
compliant with every update to the ISM, timely processing of ISM changes is recommended to assist with ongoing
identification of risks. ASD security advisories are released in response to current and active threats it is aware of.
These advisories are usually more time sensitive and affected entities are recommended to review and implement the
advice in a timely manner.
Measures to proactively monitor and manage vulnerabilities in systems can provide CSPs and cloud consumers with a
wealth of valuable information about their exposure to cyberthreats, as well as assist them to determine security risks
associated with the operation of their systems. Intentional changes between different releases or ongoing
improvements should also be considered for their impacts on the security risks of a system.
CSP systems
The CSP is responsible for monitoring and assuring the security of its systems and related hardware, software and
facilities. Security-related information determined through continuous monitoring, where relevant to cloud
consumers, can be included in the addendums to the CSP’s cloud security assessment report, and used to issue CSP
security advisories as appropriate.
Additionally, CSPs are in a unique position to proactively undertake automated activities to identify and report on
possible cloud consumer misconfigurations that do not align with security best practices. CSPs have some visibility
across all their cloud consumers (including access that cloud consumers do not have), they have the best knowledge
of their platforms and are best placed to provide these unique insights. Security best practices can include both the
CSP’s general best practices and best practices for aligning with the ISM on the CSP’s cloud platform. CSPs are
encouraged to provide configuration guidance to its cloud consumers to assist them to develop secure systems on
their cloud services.
CSPs may also wish to provide a configuration option that proactively prevents the deployment of resources against
security best practices, enabling security by default.
Cloud consumers need to continually monitor their own systems and data hosted in the cloud. Often, CSPs have
limited visibility of the customer’s usage of their cloud platform and this will not include the context of the cloud
consumer’s systems, use case and data. Therefore, CSPs can only have limited responsibilities for monitoring a cloud
consumer’s systems and data for any cybersecurity compromises. It is therefore important for cloud consumers to
develop and maintain their own processes for monitoring of their cloud systems as part of their wider operational
security responsibilities. Some CSPs provide a cloud service that assists cloud consumers to monitor their own systems
and data. cloud consumers should review the security related features of CSP services, as well as potential alternative
solutions, to identify the capabilities the cloud consumer needs to provide effective continuous monitoring.
are self-assessed separate documents that in no way modify the contents of the independent cloud security
assessment report in order to preserve authenticity
detail any changes, deviations, corrections and clarifications required to maintain the accuracy and currency of
the report, including the Cloud controls matrix
While addendums will initially not be independently verified, they will prevent the report from becoming inaccurate
or invalid as the CSP and its cloud services evolve over time. Any information contained in the addendum is to be
independently verified the next time the CSP and its cloud service undergo an independent assessment. Additionally,
CSPs are recommended to issue addendums covering any significant changes to the Cloud security assessment report
template published by ASD.
CSPs with previous IRAP assessments issued before the Cloud security assessment report template was available, may
undertake a self-assessment addendum to the report using the structure of the new template to enable transition and
closer alignment with the updated process. Noting that cloud consumers should consider the age of the report.
publication of any addendums to the CSP security fundamentals and cloud services report
planned changes that will significantly affect the security posture of the system, including supply chain changes
the discovery that controls for the CSP, its cloud platform or any of its cloud services are not as effective as
intended
security incidents or discovered vulnerabilities that either affect the cloud consumer or cannot be ruled out as
affecting the customer – this includes the identification of misconfigurations, data spills, cybersecurity intrusions
and customer data access by the CSP
decisions to delay or not undertake a change, including those from upstream vendor updates or security
advisories
any event that significantly affects the security of the CSP or its cloud services that otherwise may trigger a
reassessment.
specific weaknesses or deficiencies in deployed controls and the source of the identified weaknesses
proposed risk mitigation approach to address the identified weaknesses or deficiencies in the control
implementations (for example, prioritisation of risk mitigation actions and allocation of risk mitigation
resources).
any related CSP security fundamentals and cloud services report addendums.
Continually disclosing information to cloud consumers will enable them to make timely, risk-informed decisions about
the operation of their systems and the protection of data, them to efficiently identify and respond to any risks
deemed unacceptable.
Contact details
If you have any questions regarding this guidance you can write to us or call us on 1300 CYBER1 (1300 292 371).
The material in this guide is of a general nature and should not be regarded as legal advice or relied on for assistance
in any particular circumstance or emergency situation. In any important matter, you should seek appropriate
independent professional advice in relation to your own circumstances.
The Commonwealth accepts no responsibility or liability for any damage, loss or expense incurred as a result of the
reliance on information contained in this guide.
Copyright
With the exception of the Coat of Arms, the Australian Signals Directorate logo and where otherwise stated, all
material presented in this publication is provided under a Creative Commons Attribution 4.0 International licence
(www.creativecommons.org/licenses).
For the avoidance of doubt, this means this licence only applies to material as set out in this document.
The details of the relevant licence conditions are available on the Creative Commons website as is the full legal code
for the CC BY 4.0 licence (www.creativecommons.org/licenses).
The terms under which the Coat of Arms can be used are detailed on the Department of the Prime Minister and
Cabinet website (www.pmc.gov.au/resources/commonwealth-coat-arms-information-and-guidelines).