ScienceCloud Hardware¶
🔧 Under Development
This page is under active development. Content is being updated as we prepare the new version of ScienceCloud for production. Some sections may be incomplete or subject to change.
ScienceCloud currently provides the compute instances and storage with the following grand totals:
Computing¶
| nodes | virtual CPUs | total RAM |
|---|---|---|
| 🔧 Coming Soon |
CPU Flavors¶
A virtual machine flavor describes the size and kind of resources that will be allocated in a virtual machine. It is basically a combination of the number of CPUs and the amount of RAM composing your virtual machine, plus an optional GPU device.
A flavor type is a group of flavors that have common features, but vary only in the size. For example the ratio between the number of virtual CPUs (henceforth: vCPU) and the RAM amount can vary between each flavor type, but is always the same for flavors of the same kind.
Common features are:
- the ratio between the number of virtual CPUs and the amount of RAM (in GB)
- the processor model
- the ratio between the number of vCPUs and physical CPUs (e.g., there could be 2 vCPUs that are shared on 1 physical core)
- the allocation of a certain number of GPU devices
- vCPU and memory cost
GPU Devices¶
These flavors have an addition of the suffix "-gpu" followed by the model of the GPU device.
- -gpuP4 for the NVIDIA T4 with 16 GB VRAM
- -gpuT4 for the NVIDIA L4 with 24 GB VRAM
| nodes | GPUs | virtual CPUs | total RAM |
|---|---|---|---|
| 🔧 Coming Soon |
Flavor Availability Report¶
The table below shows the current number of available public flavors by type.
Loading the report requires internal network or VPN.
Direct link: https://herd.science-it.uzh.ch/avail-flavors/avail-public-flavors.html
Storage¶
Block storage served by Cinder and S3-compatible Object Storage.
| type | raw capacity | usable capacity |
|---|---|---|
| Block storage | 🔧 Coming Soon | |
| Object storage | 🔧 Coming Soon |
Network¶
Every compute node has a non-blocking, redundant, 10gbps link to the internal network. This network is used to access the underlying storage infrastructure and provide network connectivity to the virtual machines.
The uplink to the University network is a redundant 20gbps link.
Hypervisors¶
A hypervisor is a virtual machine monitoring system. Flavors are statically mapped to a specific hypervisor type, so each different group of flavors will run only on a specific set of hypervisors, and the CPU type seen from within the virtual machine does not change.
| No. nodes | CPU type @ base clock | Base clock | Architecture | CPU/vCPU | RAM per node | Accelerators |
|---|---|---|---|---|---|---|
| 🔧 Coming Soon |
Both the instance root disk's data and any additionally attached volumes' data are stored on our CEPH cluster and are thus accessed via a redundant 20 or 50 gigabit network link.
To reduce fragmentation flavors are statically mapped to a specific hypervisor type. More specifically:
| Flavor type | Hypervisor CPU |
|---|---|
| 🔧 Coming Soon |
There are a few cases in which it is desirable to have instances backed by the same generation of hardware (i.e. custom compiled software optimized with specific flags).
This can be accomplished on request by enabling specific flavors that are bound to a specific CPU generation. You should keep in mind that the binding solution will be more prone to issues during hardware maintenance as the instances cannot be moved easily from one host to another.