Usage policies and legal
Triton intended use cases
Triton accounts should be used for scientific research (anything sponsored by a research PI that is aligned with Aalto’s mission), including employed researchers, masters, PhD theses, and others in a department staff database. In practice, anyone with a researcher, employee (not TAs), or visitor job title gets an account made immediately after they accept the terms of use. For others, we ask their research supervisor for confirmation before making an account.
If you have an account, exploring computing to learn new skills or test out new is OK if it doesn’t use large amounts of resources. We don’t mind other employees using it for lightweight computing tasks or experimentation, but we can help you find the best place if it gets large.
Common exceptions:
Undergraduate students doing projects which can’t be considered research should use resources for students.
Courses should generally not recommend Triton as a resource. Certain courses with a special reason (e.g. HPC courses) may use it after special agreement between the instructor and Triton staff.
Student projects which are research projects with a research PI may use Triton.
Acceptable Use Policy and Terms of Service
By using the Triton cluster resources, you shall be deemed to accept these conditions of use:
You shall only use Triton cluster to perform work, or transmit or store data consistent with the stated goals and policies of Aalto University and in compliance with these conditions of use.
You shall not use Triton cluster for any unlawful purpose and not (attempt to) breach or circumvent any administrative or security controls. You shall respect copyright and confidentiality agreements and protect your credentials (e.g. user login name, password, ssh private key), sensitive data and files.
You shall immediately report any known or suspected security breach or misuse of Triton cluster or credentials to the cluster support team.
Use of the cluster is at your own risk. There is no guarantee that the cluster will be available at any time or that it will suit any purpose.
Logged information, including information provided by you for registration purposes, shall be used for administrative, operational, accounting, monitoring and security purposes only in accordance with the policy below. This information may be disclosed to other organizations anywhere in the world for these purposes in the extent allowed by local laws. Although efforts are made to maintain confidentiality, no guarantees are given.
The cluster support team is entitled to regulate and terminate access for administrative, operational and security purposes and you shall immediately comply with their instructions.
You are liable for the consequences of any violation by you of these conditions of use.
You agree to explicitly mention and acknowledge the use of Science-IT resources in your work in any reports, workshops, papers or similar that result from such usage. Appropriate reference can be found at Acknowledgement of Triton usage.
Triton data (privacy) policy
Triton is a part of Aalto University IT systems, thus is fundamentally governed by the Aalto Privacy Policy for Employees or Privacy Policy for Students, the latest versions of which can always be found on aalto.fi.
For clarity, in this section, we describe the special cases of Triton data:
In summary:
The Triton account is not a separate account, it is part of the Aalto account. We do not control that.
Triton usage statistics and logs. Triton is used for university academic research only, so this information may be used for reporting and management in any way. Identifying information won’t be public, but note that it will be used for internal operations and contacting users as needed.
Data stored on Triton. We are not the controller of this data. Data in your personal directories is controlled by you, and data in shared directories is controlled by the manager of that group. See the section below for more information on this data.
HAKA login data (JupyterHub only). This is used to secure access to JupyterHub. Only your Aalto account name is requested, it is compared and immediately discarded (Triton is already linked to your Aalto account).
The triton-users mailing list is automatically formed from all Aalto accounts in the triton-users group (everyone with an account). This is used to send service announcement and information related to scientific computing. This subscription is intrinsically tied to the Triton account and a requirement of the cluster usage. (Email information held by Aalto IT services).
We do not consider the Triton management data to consist of a personal data file (this is covered under Aalto policies), but for full disclosure we describe our use of data.
Note about research data: This section does not cover any data which users store on the cluster: for that, the user is the controller and Science-IT is only a processor. You are responsible for any administrative privacy matters. The following subsections relate only to administrative metadata.
Controller and contact
Controller: Aalto Science-IT, Aalto University, Espoo, Finland. Contact information. Please use the support email alias for account and personal data queries.
Account information comes from Aalto ITS registers.
The purpose for processing the personal data
Data is processed and stored in accordance with our agreement to provide a HPC cluster service including accounting and reporting, in accordance with the usage agreement. The cluster may only be used to support Aalto (not personal) activities, and all thus usage metadata represents Aalto activities and is owned by Aalto University.
Types of data
Triton stores the information necessary for provision of its services, including accounting, funding, and security. This includes logs of all operations and metadata of stored data. Data is only generated when a users uses the cluster. For example (including but not limited to):
Connection logs
Job submission and statistics logs
Filesystem and storage metadata and logs
Uses of data
Data is used in the provision of the HPC cluster service. Primarily, this is through accounting, reporting, and scheduling of tasks. Historical data will automatically adjust future cluster priority.
Sources of information
Data is produced during the use of Triton for research purposes. This data is generated directly by users while using the cluster. Account information is provide by Aalto University, and in general not stored or processed here.
Data sharing
Data may be used for internal Aalto reporting and accounting (usually but not always aggregated at least at the group level), and used in non-identifiable forms in public reports and statistics. It may also be used as needed to investigate usage matters.
All users of the cluster may inspect the usage information and job statistics of the entire cluster (including all other users).
Timeframe
Data related to usage remains as long as the user has an active Triton account. Technical logging data allows accounting and reporting, and may be kept as long as needed for security and reporting purposes (indefinitely). Where possible, this may be in anonymous form.
Legal notices
Data is stored in Finland in Aalto or CSC approved facilities. Access is only via Aalto account.
You may request rectification of your data. However, most data is technical logging information which can not be removed or changed.
You may cease using the cluster, remove your research data, and request your account be closed (this does not close your Aalto account because we do not control that), but historical usage data will remain for accounting purposes. Should technical errors in data be identified, a bug should be reported.
You may access and extract your own data using the standard interfaces described in the user guide.
Identifiable administrative metadata and accounting data is not transferred outside of the EU/EEA except under proper agreement. (We have to say that, but in reality identifiable data is never transferred out of Aalto or maybe the FGCI consortium in Finland).
You may lodge a complaint with the Aalto data protection officer (see Aalto privacy notices for up to date contact information) or the Finnish supervision authority Tietosuoja.
Research and home data stored on cluster
We provide a storage service for for data stored on the cluster (scratch and home directories):
Our responsibility is limited to keeping this data secure and
providing access to the corresponding Aalto accounts. The shared
directory manager should be able to make choices about data. We do
not access this data except with an explicit request, but for
management purpose we do use the file metadata (stat $filename
).
For full information, see the Science-IT data policy.
We do not look into private files without your explicit request (if you want help with something, explicitly tell us if we can look at them).
If your files are made cluster-readable (the
chmod
“other” permissions), you give permission for others to look at contents. Note that this is not the default setting.Should you report a problem, we may run
stat
as superuser on relevant files to determine basic metadata without further checks.Should you have a problem that requires us to look at the contents of files or directories, we must first have your explicit permission (either in writing or in person)
User-owned data (home directories, work directories) may be deleted six months after an account expires. Use a group-based storage space instead.
Our data storage service is suitable for confidential data. You must ensure that permissions are such that technical access is limited.