WIP Changes are all marked in BLUE or struck-out.
Explanatory comments in GREEN are not part of text.
Start of Policy
Creation date: 20090216 Editor: iang Status: WIP m20100327.2 as of 20100404 00:00:02 UTC |
![]() |
This Security Policy sets out the policy for the secure operation of the CAcert critical computer systems. These systems include:
Board may add additional components into the Security Manual.
Critical roles are covered. These roles are defined as:
Non-critical systems are not covered by this manual, but may be guided by it, and impacted where they are found within the security context. Architecture is out of scope, see CPS#6.2.
Important principles of this Security Policy are:
Each task or asset is covered by a variety of protections deriving from the above principles.
This Security Policy is part of the Configuration-Control Specification for audit purposes (DRC-A.1). It is under the control of Policy on Policy for version purposes.
This policy document says what is done, rather than how to do it. Some sections are empty, which means "refer to the Manual."
This Policy explicitly defers detailed security practices to the Security Manual ("SM"). The SM says how things are done. As practices are things that vary from time to time, including between each event of practice, the SM is under the direct control of the Systems Administration team applicable team leaders. It is located and version-controlled on the CAcert wiki.
Section Headings are the same in both documents. Where Sections are empty in one document, they are expected to be documented in the other. "Document further in Security Manual" can be implied from any section heading in the Policy.
The team leaders may from time to time explicitly defer single, cohesive components of the security practices into separate procedures documents. Each procedure should be managed in a wiki page under their control, probably at SystemAdministration/Procedures. Each procedure must be referenced explicitly in the Security Manual.
CAcert shall host critical servers in a highly secure facility. There shall be independent verification of the physical and access security.
Computers shall be inventoried before being put into service. Inventory list shall be available to all Access Engineers and all Systems Administrators. List must be subject to change control.
Each unit shall be distinctly and uniquely identified on all visible sides. Machines shall be housed in secured facilities (cages and/or locked racks).
Acquisition of new equipment that is subject to a pre-purchase security risk must be done from a vendor that is regularly and commercially in business. Precautions must be taken to prevent equipment being prepared in advance.
Equipment that is subject to a service security risk must be retired if service is required. See also §2.2.3.3.
Storage media (disk drives, tapes, removable media) are inventoried upon acquisition and tracked in their use.
New storage media (whether disk or removable) shall be securely wiped and reformatted before use.
Removable media shall be securely stored at all times, including when not in use. Drives that are kept for reuse are wiped securely before storage. Reuse can only be within critical systems.
When there is a change to status of media, a report is made to the log specifying the new status.
Storage media that is exposed to critical data and is to be retired from service shall be destroyed or otherwise secured. The following steps are to be taken:
Records of secure erasure and method of final disposal shall be tracked in the asset inventory. Where critical data is involved, two systems administrators must sign-off on each step.
In accordance with the principle of dual control, at least two persons authorized for access must be on-site at the same time for physical access to be granted.
Access to physical equipment must be authorised.
The Security Manual must present the different access profiles. At least one Access Engineer must control access in all cases. At least one systems administrator will be present for logical access. Only the most basic and safest of accesses should be done with one systems administrator present.
There is no inherent authorisation to access the data. Systems Administrators and Application Engineers are authorised to access the raw data under the control of this policy. All others must not access the raw data. All are responsible for protecting the data from access by those not authorised.
All physical accesses are logged and reported to all.
There must not be a procedure for emergency access. If, in the judgement of the systems administrator, emergency access is required and gained, in order to avoid a greater harm, independent authorisation before the Arbitrator must be sought as soon as possible. See DRP.
All personel who are in possession of physical security codes and devices (keys) are to be authorised and documented.
Current and complete diagrams of the physical and logical CAcert network infrastructure shall be maintained by systems administration team leader. These diagrams should include cabling information, physical port configuration details, expected/allowed data flow directions, and any further pertinent information, as applicable. Diagrams should be revision controlled, and must be updated when any change is made.
Only such services as are required for normal operation should be visible externally; systems and servers which do not require access to the Internet for their normal operation must not be granted that access. If such access becomes temporarily necessary for an authorized administrative task, such access may be granted under the procedures of the SM and must be reported and logged.
System and server connections internal to the CAcert infrastructure should be kept to the minimum required for routine operations. Any new connectivity desired must be requested and approved by System administration team leader and then must be reflected in the appropriate infrastructure diagram(s).
All ports on which incoming traffic is expected shall be documented; traffic to other ports must be blocked. Unexpected traffic must be logged as an exception.
All outbound traffic that is initiated shall be documented; traffic to other destinations must be blocked. Unexpected traffic must be logged as an exception.
Logs should be examined regularly (by manual or automatic means) for unusual patterns and/or traffic; anomalies should be investigated as they are discovered and should be reported to appropriate personnel in near-real-time (e.g. text message, email) and investigated as soon as possible. Suspicious activity which may indicate an actual system intrusion or compromise should trigger the incident response protocol described in section 5.1.
Any operating system used for critical server machines must be available under an OSI-approved open source software license.
Any operating system used for critical server machines must support software full-disk or disk volume encryption, and this encryption option must be enabled for all relevant disks/volumes when the operating system is first installed on the machine.
Servers must enable only the operating system functions required to support the necessary services. Options and packages chosen at OS install shall be documented, and newly-installed systems must be inspected to ensure that only required services are active, and their functionality is limited through configuration options. Any required application software must follow similar techniques to ensure minimal exposure footprint.
Documentation for installing and configuring servers with the appropriate software packages and configurations will be maintained by the System Administrators.
Software used on production servers must be kept current with respect to patches affecting software security. Patch application is governed by CCS and must be approved by the systems administration team leader, fully documented in the logs and reported by email to the systems administration list on completion (see §4.2).
Application of a patch is deemed an emergency when a remote exploit for a weakness in the particular piece of software has become known (on servers allowing direct local user access, an emergent local exploit may also be deemed to be an emergency). Application of patches in this case may occur as soon as possible, bypassing the normal configuration-change process. The systems administration team leader must either approve the patch or instruct remedial action, and refer the case to dispute resolution.
Declaration of an emergency patching situation should not occur with any regularity. Emergency patch events must be documented within the regular summaries by the team leader to Board independent of filed disputes.
Systems administration is to provide a limited environment to Applications Engineers in order to install and maintain the application.
All access to critical data and services shall be controlled and logged.
General access for Members shall be provided via a dedicated application. General features are made available according to Assurance Points and similar methods controlled in the software system.
Additional or special access is granted according to the authorisations on the below access control lists (see §1.1.1):
List Name | Who | Purpose of access | Relationship | Manager |
Physical Control List | Access Engineers | control of access by personnel to hardware | exclusive of all other roles | Access team leader Board of CAcert (or designee) |
Physical Access List | Systems Administrators | hardware-level for installation and recovery | exclusive with Access Engineers and Software Assessors | systems administration team leader Board of CAcert (or designee) |
SSH Access List | Systems Administrators and Application Engineers | Unix / account / shell level | includes by default all on Physical Access List | systems administration team leader |
Repository Access List | Application Engineers | change the source code repository and install patches to application | exclusive with Access Engineers and systems administrators | software assessment team leader |
Support Access List | Support Engineer | support features in the web application | includes by default all Application Engineers systems administrators | systems administration support team leader |
All changes of personnel to the above lists are approved by the Board of CAcert.
Strong methods of authentication shall be used wherever possible. All authentication schemes must be documented.
Follow-up actions to termination must be documented. See §9.1.7.
Primary systems administration tasks shall be conducted under four eyes principle. These shall include backup performance verification, software patch application, account creation and deletion, and hardware maintenance.
Access to Accounts (root and user via SSH or console) must be strictly controlled. Passphrases and SSH private keys used for entering into the systems will be kept private to CAcert sysadmins and Application Engineers in all cases.
Only System Administrators and Application Engineers designated on the Access Lists in §3.4.2 are authorized to access accounts, unless specifically directed by the Arbitrator.
All access is secured, logged and monitored.
The procedure for changing passphrases and SSH keys shall be documented.
Response times should be documented for Disaster Recovery planning. See §6.
All changes made to system configuration must be recorded and reported in regular summaries to the Board of CAcert.
All sensitive events should be logged reliably. Logs should be deleted after an appropriate amount of time as documented in the Security Manual.
Access to logs must be restricted. The security of the logs should be documented. The records retention should be documented.
Logging should be automated, and use should be made of appropriate system-provided automated tools. Automated logs should be reviewed periodically; suspicious events should be flagged and investigated in a timely fashion.
Configuration changes, no matter how small, must be logged.
All physical visits must be logged and a report provided by the accessor and by the Access Engineer.
The procedure for all backups must be documented, according to the following sub-headings.
Backups must be taken for operational and for disaster recovery purposes. Operational backups may be online and local. Disaster recovery backups must be offline and remote.
Backups must be protected to the same level as the critical systems themselves. Disaster recovery backups may be distributed.
See §2.2.3.
Backups must be encrypted and must only be transmitted via secured channels. Off-site backups must be dual-encrypted using divergent methods.
Two CAcert System Administrators must be present for verification of a backup. Four eyes principle must be maintained when the key and backup are together. For any other purpose than verification of the success of the backup, see next.
The encryption keys must be stored securely by the CAcert systems administrators. Paper documentation must be stored with manual backups.
Conditions and procedures for examining the backups must be documented, and must be under Arbitrator control for purposes other than verification and recovery.
Termination of user data is under direction of the Arbitrator. See CCA.
See §4.2.1.
See §5.6.
The standard of monitoring, alerting and reporting must be documented.
On discovery of an incident, an initial assessment of severity and priority must be made.
An initial report should be circulated.
A communications forum should be established for direct support of high priority or high severity incidents.
A process of escalation should be established for oversight and management purposes, proportional to severity and priority. Oversight starts with four eyes and ends with the Arbitrator. Management starts with the team leader and ends with the Board.
Incidents must be investigated. The investigation must be documented. If the severity is high, evidence must be secured and escalated to Arbitration.
Incident reports shall be be published. The Incident Report is written on closing the investigation. A full copy should be appended to the documentation of the investigation. Sensitive information may be pushed out into a restricted appendix of the report. The systems administration team leader is responsible for publication and maintenance.
Incidents are not normally kept secret nor confidential, and progress information should be published as soon as possible. The knowledge of the existence of the event must not be kept secret, nor the manner and methods be kept confidential. See §9.5.
Disaster Recovery is the responsibility of the Board of CAcert Inc.
Board must develop and maintain documentation on Business Processes. From this list, Core Processes for business continuity / disaster recovery purposes must be identified.
Board should identify standard process times for all processes, and must designate Maximum Acceptable Outages and Recovery Time Objectives for the Core Processes.
Board must have a basic plan to recover.
Board must maintain a key persons List with all the contact information needed. See §10.1. The list shall be accessible even if CAcert's infrastructure is not available.
Software assessment team is responsible for the security and maintenance of the code.
The source code is under CCS. Additions to the team are approved by Board. See §3.4.2.
The primary tasks for Software Assessors are:
Software assessment is not primarily tasked to write the code. In principle, anyone can submit code changes for approval.
The primary tasks for Application Engineers are:
The application code and patches are maintained in a central repository that is run by the software assessment team.
The development code and testing patches are maintained in a central development repository that is run by the software assessment team.
The production code is maintained in a secure production repository within the critical systems that is run by the systems administation team. Access is made available to the Application Engineers.
At the minimum, patches are signed off by the team leader or his designated reviewer. Each software change should be reviewed by a person other than the author. Author and signers-off must be logged. The riskier the source is, the more reviews have to be done.
Software assessment team maintains a test system. Each patch should be built and tested. Test status of each patch must be logged.
Software assessment team maintains a bug system. Primary communications should go through this system. Management access should be granted to all Software Assessors, software developers, and systems administrators. Bug submission access should be provided to any Member that requests it.
The Application Engineer is a role within Software Assessment team that is approved to install into production the patches that are signed off. Once signed off, the Application Engineer commits the patch from the development repository to the production repository, and installs the patch from the production repository into the running code. The Application Engineer is responsible for basic testing of functionality and emergency fixes, which then must be back-installed into the repositories.
Requests to Application Engineers for ad hoc queries over the database for business or similar purposes must be approved by the Arbitrator.
See §3.3.
The software interface gives features to Support Engineer. Access to the special features is under tight control. Additions to the team are approved by Board, and the software features are under CCS. See §3.4.2.
Support Engineers do not have any inherent authority to take any action, and they have have to get authority on a case-by-case basis. The authority required in each case must be guided by this policy or the Security Manual or other clearly applicable document. If the Member's authority is not in doubt, the Member can give that authority. If not, the Arbitrator's authority must be sought.
Support Engineers are responsible to follow the policies and practices.
Support Engineers have these responsibilities:
Support may always be contacted by email at support at cacert dot org. Other channels may be made available and documented in Security Manual.
Support Engineers refer questions requiring authority to Arbitration, and may also be called upon to act as default Case Managers. See DRP and Case Manager's Handbook. Support Engineers should be familiar with these topics, even if not listed as Arbitrators or Case Managers.
Each team should have a minimum of two members available at any time. Individuals should not be active in more than one team at any one time, but are expected to observe the other teams. See §3.4.2 for exclusivities.
One individual in each team is designated team leader and reports to Board.
New team members need:
The team supports the process of adding new team members.
The Arbitrated Background Check ("ABC") must be conducted under the direction of the Arbitrator, with a separate Case Manager to provide four eyes. ABCs are carried out with full seriousness.
An investigation within ABC should include examination of:
ABC is to be done on every individual in a critical role.
The process of the ABC should be documented as a procedure.
Documentation of each individual check should be preserved and should be reviewable under any future Arbitration. It must include:
The following privacy considerations exist:
CAcert trusted roles give up some privacy for the privacy of others.
Individuals and access (both) must be authorised by the Board. Only the Board may approve new individuals or any access to the systems. Each Individual should be proposed to the Board, with the relevant supporting information as above.
The Board should deliberate directly and in full. Board members who are also active in the area should recuse from the vote, but should support the deliberations. Deliberations and decisions should be documented. All conflicts of interest should be examined.
It is the responsibility of all individuals to observe and report on security issues. All of CAcert observes all where possible. It is the responsibility of each individual to resolve it satisfactorily, or to ensure that it is reported fully.
See §9.5.
Termination of access may be for resignation, Arbitration ruling, or decision of Board or team leader. On termination (for any reason), access and information must be secured. See §3.4.4.
The provisions on Arbitration survive any termination by persons fulfilling a critical role. That is, even after a person has left a critical role, they are still bound by the DRP (COD7), and the Arbitrator may reinstate any provision of this agreement or bind the person to a ruling.
It is the responsibility of the team leaders to coordinate technical testing and training, especially of new team members.
Root keys are generated only on instruction from the Board. They must be generated to a fully documented and reviewed procedure. The procedure must include:
Root keys must be kept on reliable removable media used for that purpose only. Private Keys must be encrypted and should be dual-encrypted. Passphrase must be strong and must be separately escrowed from media. Dual control must be maintained.
The top-level root must be escrowed under Board control. Subroots may be escrowed by either Board or Systems Administration Team.
Recovery must only be conducted under Arbitrator authority.
The Board is responsible to the Community to manage the CA at the executive level.
All external inquiries of security import are filed as disputes and placed before the Arbitrator under DRP.
Only the Arbitrator has the authority to deal with external requests and/or create a procedure. Access Engineers, systems administrators, support engineers, Board members and other key roles do not have the authority to answer legal inquiry. The Arbitrator's ruling may instruct individuals, and becomes your authority to act.
Components may be outsourced. Team leaders may outsource non-critical components on notifying the Board. Critical components must be approved by the Board.
Any outsourcing arrangements must be documented. All arrangements must be:
Contracts should be written with the above in mind.
CAcert is an open organisation and adopts a principle of open disclosure wherever possible. See Principles. This is not a statement of politics but a statement of security; if a security issue can only be sustained under some confidentiality or secrecy, then find another way.
In concrete terms, confidentiality or secrecy may be maintained only under a defined method in policy, or under the oversight of the Arbitrator (which itself is under DRP). The exception itself must not be secret or confidential. All secrets and confidentials are reviewable under Arbitration, and may be reversed. All should strive to reduce or remove any such restriction.
Contact information for all key people and teams must be documented.
All incorporated Documents must be documented.
Relevant and helpful Documents should be referenced for convenience.