Dedicated Instances¶
Last changed: 2026-08-17
Note
The Dedicated Instances service is only available in the OSL region.
This document describes the Dedicated Instances service offering in NREC. This service differs from the general compute service in one key area:
The resources assigned to an instance are “dedicated”, in the sense that there is no sharing or overcommit of resources. This applies to CPU and RAM resources, that are usually shared among instances. In the Dedicated Instances service, actual CPU cores and memory chunks are assigned exclusively to the instance.
Note that network resources (including disk I/O) are still shared among instances running on the hypervisor.
Due to the nature of the dedicated resources, live migration of instances between hypervisors is difficult or impossible. This in turn means that downtime on a hypervisor, such as scheduled downtime due to maintenance, also applies to the instances running on that hypervisor.
When to use Dedicated Instances¶
There are two usecases where the Dedicated Instances service is useful:
- Critical Instances:
Instances that are highly critical for a limited time, such as instances running exam software. The dedicated resources guarantee that no other instances on the hypervisor can interfere by causing resource starvation.
- Heavy Performance Needs:
Running high-performance computing (HPC) or complex database engines requiring predictable throughput, or instances that utilize 100% CPU and/or memory for an extended time period. Such instances may not fit in a shared environment such as Shared HPC.
There may be other usecases that we haven’t identified.
Getting Access¶
Please contact us via support@nrec.no for access to this service.
Flavors¶
We currently have the following flavors for dedicated instances:
Flavor name |
Virtual CPUs |
Memory |
|---|---|---|
|
8 |
30 GiB |
|
16 |
60 GiB |
|
32 |
120 GiB |
|
64 |
240 GiB |