Comment on page
The Nuvolos platform natively provides compute resources as well as configurable tools for computational research and educational use cases, as well as for combinations of both.
The pricing structure of Nuvolos consists of two parts: the Nuvolos subscription and on-demand resources. The Nuvolos subscription is based on your contract with us and runs on an annual basis. The on-demand resources are additional resources that can be made available on top of those provided for by the Nuvolos subscription.
The subscription allows usage on an annual basis of a certain number of resource units called Nuvolos Compute Units (NCUs), while the on-demand resources are separately paid for by means of Credits. These allow access to different resources in a resource pool, which can be either shared or dedicated. In this overview, we will briefly explain each of these concepts.
A subscription to Nuvolos is based on a yearly contract signed between you and the company that develops Nuvolos, Alphacruncher AG. The subscription is specific to your particular use case and requirements; there are no fixed tiers or plans for Nuvolos subscriptions. Instead, you convey us your particular use case, and we determine based on that use case the estimated annual resources you will need and provide these to you.
The cost of the Nuvolos subscription is in turn based on the quantity of such resources allocated on an annual basis. In this way, what you pay is always custom-tailored to the resource needs you actually have, preventing the problem of overpayment for underutilized plans or tiers typical of SaaS offerings.
In general terms, a subscription allocates to the subscriber a certain number of compute resources or NCUs, based on a shared resource pool. These NCUs can be spent on concurrent usage of applications (i.e., tools such as Airflow, VSCode, Spyder, etc.) up to the maximum amount that is allocated to you based on your subscription. Note that the subscription's resource allocation is always for an average usage over the long term: to prevent a rigidity in resource provision that is not tenable for scientific practice, your subscription allows for short-term fluctuations in usage in excess of the allocated NCUs.
A Nuvolos subscription offers otherwise unlimited usage of the platform's features for an unlimited number of users, as long as the average long-term NCU allocation is not exceeded. We do not charge 'per seat'.
Nuvolos Compute Units (NCUs) are the unit of resource measurement on the Nuvolos platform. Each NCU stands for 1vCPU and 4GB of RAM in a shared resource pool. These compute resources are expended on the running of applications on Nuvolos. Application resource usage is therefore expressed in NCUs as well: for example an application using 4vCPUs and 16GB of RAM uses 4 NCUs.
You can set how many NCUs an application will have available as resources by scaling applications yourself. Self-service scaling within the platform is possible between 1 and 16 NCUs.
There are two types of resource pools: shared resource pools and dedicated resource pools. The standard Nuvolos subscription provides for NCUs out of a shared resource pool.
In the shared resource pool, all applications you run will draw from the same resource allocation, i.e. from your general allocation of NCUs. A shared resource pool is essentially nothing other than a certain amount of vCPU and RAM that is configured identically and dedicated to running all applications you choose to run.
It is also possible to draw from a dedicated resource pool. This is a resource pool configured specifically to optimise for one particular application, and only one application will draw on it at a time. This is useful in case you need a particularly heavy or unusual compute usage for a particular application (say, a High Performance Computing use case).
Applications on dedicated resource pools can scale between 8 vCPUs/32GB RAM and 120 vCPUs/456GB RAM. As with shared resource pools, you can scale yourself as needed using the Nuvolos platform.
Unlike the regular NCU allocation out of a shared resource pool, usage of a dedicated resource pool requires purchasing additional Credits corresponding to the intended usage.
By default, the Nuvolos subscription includes sufficient storage for your use case. The storage provision is based on our high performance file system, backed up by Snowflake as our database provider.
Specific storage needs, such as SQL-compliant table storage and large file storage for unusually large volumes of unstructured data, can be purchased as additional services using Credits.
Besides the general Nuvolos subscription, you can also purchase Credits. These are units of measure for on-demand resources. They essentially allow access to additional NCUs, as well as special allocations such as dedicated resource pools (see above), special kinds of storage, and professional services.
Credits can be bought at the rate of 8 Euros or 9 USD per Credit. They can be purchased at any time as long as you have a Nuvolos subscription. Any unused Credits will roll over to be available for the next year provided you renew your Nuvolos subscription. Credits cannot be acquired or used independently of a Nuvolos subscription.
Credits can be applied to any resources that can be obtained using Microsoft Azure cloud services, including specific compute needs, graphic cards, storage types, and so forth.
Cost control can be exercised in various ways.
Firstly, self-service scaling combined with the ability to allocate budgets of resources or resource types gives you substantial flexibility as to how and when you use the resources available to you.
Secondly, for academic institutional use cases, NCU availability is tailored to the academic year. This means NCUs allocated to you will be available for a certain amount of hours within the academic year, excluding the summer holiday period. This in turn lowers the cost of the subscription, ensuring you do not pay for resource availability during a period when the resources are likely to go unused. For details on time constraints on NCU availability, refer to our documentation on NCU architecture.