FAQs about Nuvolos

Nuvolos is a unified, governed computational workspace for research and teaching, designed to operate across cloud and self-hosted infrastructures. It combines environment management, reproducibility, data governance, and HPC integration into a single system, enabling teams to run, share, and reproduce computational work without fragmenting tooling. Users work in fully versioned environments supporting Jupyter, RStudio, MATLAB, Stata, Julia, and other scientific applications, directly in the browser.

Nuvolos is used by research groups, data-intensive research institutes, research departments, and universities that need reproducible computational environments for research and teaching. Typical use cases include data science courses, econometrics, computational biology, ML/AI workloads, and cross-institution research projects.

The default option is the cloud version. In this case both the compute and storage infrastructure is managed by us. We also support fully and partially self-managed options, where you own the cloud account or the infrastructure.

For Managed Cloud, a pilot environment can usually be provisioned within the day, with first courses or projects running within a few weeks. Self-managed deployments depend on your internal change processes and cloud/HPC setup.

The Managed Cloud and self-managed version is an invitation based system - users receive e-mail based invitations to particular projects/areas of Nuvolos. We support AD/etc integration for self-managed user management.

Nuvolos is browser-based and works on any modern OS. Only the browser runs locally; compute and storage run in the cloud. We support Google Chrome, Microsoft Edge, Mozilla Firefox and Safari (latest versions). Chrome based browsers offer the most reliable performance. There is no need to have a local installation of Python/R/MATLAB.

Nuvolos uses institution-defined resource pools and quotas. Typical configurations cover anything from small teaching workloads to very large HPC-style jobs. We don’t impose a hard global limit, but each institution defines per-user, per-project, or per-department quotas. Higher limits (e.g. very large RAM/CPU/GPU nodes or PB-scale storage) can be configured on request.

Nuvolos supports FAIR practice by making data and environments structured, accessible (with permissions), interoperable through captured toolchains, and reusable via fully reproducible snapshots.

Nuvolos captures code, data, tools, dependencies, configuration, and results in a single versioned workspace. At any point in time, the project's state is saved as a snapshot. This feature enables reproducibility for collaborators, reviewers, and data editors long after the original analysis.

Traditional workflows require manual installs, environment setup, syncing via a Git repository, managing file servers, and connecting to HPC via the CLI. Nuvolos consolidates all of compute, storage, versioning, collaboration, and environment management into a single browser-based workspace. HPC and Git repositories are integrated and managed through a single interface.

Yes. Python, R, MATLAB, Stata, Julia, Bash, and many domain-specific tools run inside Nuvolos. A single workspace can combine multiple languages, and their exact versions and dependencies are automatically captured.

A snapshot is a point-in-time, immutable capture of your entire environment, code, data, dependencies, tools, configuration. It guarantees reproducibility and enables branching, sharing, debugging, and archiving for publication.

Nuvolos lets you prepare a curated snapshot of the exact environment used to produce your published results. Reviewers and replicators receive controlled access to a fully configured workspace, reducing manual packaging and preventing missing files or broken scripts.

Yes. Nuvolos provides HPC-scale compute (hundreds of vCPU, TBs of RAM + GPUs), multi-team collaboration, complete versioning for reproducibility, real-time resource monitoring, and flexible budget management through resource pools—all self-service, without DevOps expertise.

Nuvolos uses invitation-based access. Space administrators invite colleagues or external partners to specific projects. The distribution feature enables controlled sharing of code, data, and results across departments/organizations with granular access control.

Snapshots persist as long as your institution’s retention policies allow. Old snapshots can be moved to lower-cost storage or pruned according to configurable lifecycle rules. Complete environments can be exported to archival repositories (Zenodo, etc.) or downloaded locally for long-term preservation on user devices.

Yes. Data/environments are Findable (structured), Accessible (permission-based), Interoperable (captured toolchains), and Reusable (fully reproducible snapshots with complete versioning).

In the managed platform, external collaborators can be invited by email and assigned roles (view, run, edit). They only see environments and data explicitly shared with them, and their activity is logged. For self-managed environments, the setting depends on company polices.

Because environments are containerized and versioned, ecah snapshot keeps working even if upstream registers change. You can update the newer packages in a new snapshot while preserving older ones for reproducibility. You can revert if updates backfire.

Yes. Nuvolos integrates with on-prem HPC as a complementary layer. Users work through a modern UI and can dispatch jobs to HPC using your existing schedulers.

No. Nuvolos modernizes how users interact with HPC and adds reproducibility, versioning, collaboration, and teaching workflows on top.

Jobs are submitted from Nuvolos to your schedulers, while artifacts and results stay captured inside the user’s workspace.

Yes. Institutions can expose multiple resource pools, cloud instances, on-prem clusters, GPUs and users select them per project or per task.

Admins define pools, attach quotas, and assign them to spaces or departments. This gives full control over usage, costs, and prioritization.

Yes. GPU resources can be configured for AI, ML, or domain-specific GPU workloads across cloud or on-prem environments.

Users avoid VPNs, SSH, and complex local installs. They access compute entirely through their browser.

Yes. All computational tools live in a unified application catalog, ensuring consistent versions for all users.

Through resource pools, quotas, usage monitoring, and idle environment management, preventing runaway compute spending.

For regular workloads, the waiting time is 1-3 minutes. For scaled workloads, such as GPUs, the waiting time is 5-15 minutes. You can schedule workloads to start by a given time to reduce this waiting time.

Long-running computations are allowed within the limits defined by your institution (e.g. job walltime and maintenance windows). Nuvolos will not arbitrarily stop active jobs; idle sessions are automatically shut down after a configurable timeout.

Nuvolos uses institution-defined resource pools and quotas. Typical configurations cover anything from small teaching workloads to very large HPC-style jobs. We don’t impose a hard global limit, but each institution defines per-user, per-project, or per-department quotas. Higher limits (e.g. very large RAM/CPU/GPU nodes or PB-scale storage) can be configured on request.

Yes, you can review each session and see the logs to debug apps.

Institutions choose where data is stored and processed, including EU-specific or country-specific regions or institution-controlled cloud tenancies.

In regions selected by the institution during deployment, aligned with compliance and data-residency policies.

Yes

Yes. Institution-controlled or region-specific cloud zones are fully supported.

Yes. Nuvolos integrates with SAML, OpenID Connect, and other IAM systems.

Yes, under appropriate institutional controls. Nuvolos provides granular permissions, encryption, and audit logs to support compliance.

Yes. Administrators can review user access, workspace activity, and sharing events.

Yes, encrypted both at rest and in transit.

YES. AI and ML workloads can run entirely in institution-chosen regions or infrastructures without sending data to unmanaged services.

Yes. We support a large variety of external storage options including AWS S3, Azure Files, Azure Blobs, Sharepoint, Dropbox, etc.

We store snapshots on a “diff” basis. We will not charge you for each snapshot, only the total amount of unique, retained data.

Yes. From Nuvolos environments you can connect to external databases and data warehouses (e.g. PostgreSQL, MySQL, SQL Server, Oracle, Snowflake, BigQuery), as well as object and file storage (S3, Azure Blob, etc.), subject to your institution’s network and security policies.

Practically, we support multi-TB and, in some setups, PB-scale datasets. Exact limits depend on storage backend and functionality. We work with you to size storage tiers (high-performance, large-file, archival) appropriately for your workloads.

We offer three basic types of storage, the high-performance file system, table storage and large file storage. The high-performance file system is best suited for I/O intensive tasks, frequently changing small files, such as source code or model outputs. Table storage is best suited for large volumes of structured data. The large file storage is ideal for large volumes of unstructured (file-based) data.

Yes. MATLAB, COMSOL, Stata, Julia, ArcGIS, and many others can be hosted centrally, with institutional control over usage.

IT sets licensing rules and assigns access to specific courses, research groups, or roles. Local installs are no longer needed.

Yes. Nuvolos offers first-class support for all major open-source scientific tools.

Yes. Each course or project defines its own environment template, which is cloned for users.

Yes. Versions of libraries and tools are captured in snapshot metadata, ensuring consistent and reproducible results.

Yes, subject to institutional policies.

Yes, including deep-learning frameworks and domain-specific GPU tools.

Yes. Nuvolos supports Bring Your Own License (BYOL). Institutions can use and centrally manage their existing commercial software licenses directly inside the platform.

Yes. You can register an SSH key from Nuvolos to your git repository provider (e.g. GitHub, GitLab) and access all private repositories in any Nuvolos application you run. If you want to retire the connection, you can just remove the SSH key registration at your repository provider.

Yes, you can trigger and orchestrate CI/CD pipelines from Nuvolos (e.g. via git hooks, CLI tools, or API calls). We don’t currently recommend using Nuvolos itself as a general-purpose CI/CD runner; instead, we integrate with existing CI systems (GitHub Actions, GitLab CI, etc.) that execute the pipelines.

For each computational toolkit, a package management system (or in case of complex tools, a manager for each specific language) is used. We use mostly conda for system-level dependencies and python, R’s native package manager for R/RStudio workloads, Julia’s pkg for Julia. Each application comes with a description of the package manager in use.

Yes, we support desktop applications as long as they have native Linux desktop support. Applications with windows-only desktop support may work, but we cannot guarantee flawless execution.

Yes. Institutions can onboard their own applications into the Nuvolos catalog. In most cases this means defining a base image / environment, entry point, and resource profile. Our team can support IT or DevOps in packaging internal tools so they can be launched and managed like any other app in Nuvolos.

Yes, all Nuvolos workloads are containerized, this is one layer that guarantees reproducibility of computational work on the platform. When you build your environment, you start from a pre-built, working blueprint, add your dependencies and the platform offers you to create a containerized version of your environment without you having to write a single line of Dockerfile code.

We support x86-64 Linux-based software. Emulation of x86-64 Windows software is possible in certain cases.

Almost all Nuvolos applications come with terminal (shell) support.

We don’t currently support SSH into Nuvolos (no inbound SSH to sessions). Nuvolos environments can use SSH outbound, for example to connect to git providers using SSH keys.

Yes. Nuvolos lets you create networks between Nuvolos applications, and these can cross-communicate on pre-defined network ports. You do not need to specify network policies or firewall settings, the platform provides you with safe preset options that will suit most use-cases.

By default, users do not receive root access to the underlying operating system. System packages and base container images are controlled by IT or Nuvolos to maintain security and reproducibility.

In many cases, yes. We can onboard institution-provided container images into the app catalog, as long as they meet security and compatibility requirements. We need to make modifications to make the images compatible with Nuvolos.

A multi-layered model combining infrastructure security, isolated execution environments, strong authentication, granular authorization, and audit logging.

Each workspace runs in a securely isolated context with strictly controlled access to data and compute.

Nuvolos integrates with institutional IAM systems and uses roles and permissions internally.

Nuvolos can be configured to align with the technical requirements of many regulated domains when combined with appropriate institutional processes. Encryption, access control, logging and deployment options provide a strong foundation for handling sensitive data under domain-specific rules.

User sessions execute within isolated environments with controlled access to data and resources. This sandboxing limits the blast radius of misconfigurations or malicious code and supports safe experimentation.

Yes. Nuvolos supports courses from small groups to thousands of students. All students receive identical, isolated environments that work through the browser without local installs.

Instructors or IT create a master template. Nuvolos automatically provisions per-student workspaces from that template, ensuring every student works in the same environment from day one.

Any device. Compute and storage happen server-side, so even low-performance laptops work for demanding tasks.

Yes. Students can export notebooks, code, and allowed datasets, or push work to GitHub. Retention windows can be defined at the course or institutional level.

Nuvolos provides a fully managed digital lab that students access directly in their browser. Instead of installing software, configuring environments, or troubleshooting device issues, every student receives an identical, ready-to-use workspace with the exact tools and datasets defined by the instructor. This makes lab delivery consistent, scalable, and device-agnostic, even for very large courses.

Yes. Nuvolos supports Bring Your Own License (BYOL) for teaching. Institutions can use their existing commercial software licenses, including tools like MATLAB, COMSOL, ArcGIS, Stata, and others, and make them available directly inside the student workspaces. Licensing remains under institutional control while Nuvolos handles the environment distribution and execution.

Yes. Nuvolos offers an assignment feature which lets you distribute and collect assignment and homework material in a safe, transparent and auditable way. Deadline management is trivialized. The feature also supports anonymization.

Yes. Grading on Nuvolos enables the educator to review assignment material from the point of view of the student, so no more “on my computer it worked”. We also support automatic grading libraries for simple, large-scale assignments.

Nuvolos is not a replacement for an LMS like Moodle, Canvas, or Blackboard. The LMS remains the system of record for enrolments, grades, announcements, and general course management. Nuvolos focuses on delivering the computational lab environment, teaching material that runs in that environment, and assignment workflows.